本篇博文主要内容为 2026-08-28 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 (2026-08-28)
今日共更新695篇论文,其中:
- 自然语言处理共154篇(Computation and Language (cs.CL))
- 人工智能共196篇(Artificial Intelligence (cs.AI))
- 计算机视觉共114篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共156篇(Machine Learning (cs.LG))
- 多智能体系统共16篇(Multiagent Systems (cs.MA))
- 信息检索共24篇(Information Retrieval (cs.IR))
- 人机交互共26篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] One Model Many Minds: Unlocking Multi-Agent Synergy in a Single Agent via Mixture of Roles
【速读】：该论文旨在解决单智能体（single-agent）框架中模型专属性能受限与多智能体系统（Multi-Agent Systems, MAS）推理成本过高的双重挑战。现有单智能体方法依赖预定义人格或固定引导向量实现专业化，但其单一固定的专属性无法适应多样化任务需求；而多智能体系统虽可通过不同角色协作实现多视角问题求解，却需多轮交互，导致上下文长度激增和推理开销显著上升。为此，本文提出角色混合机制（Mixture of Roles, MoRe），其核心在于通过自适应组合多个角色编码的引导向量，在单轮推理中实现多视角专业化。关键创新在于构建一个多样化的引导向量代码本（diversified codebook），每个向量编码一个潜在角色，并设计查询感知路由模块（query-aware router），动态融合代码本中的多个角色生成复合引导向量。该复合向量用于驱动基础大语言模型（LLM），从而在保持单智能体架构的前提下，实现类似多智能体系统的动态多视角推理能力。MoRe采用三阶段监督微调（SFT）与GRPO后训练策略进行高效训练，且无需更新骨干模型。实验表明，相较于单智能体基线，MoRe在推理与人格基准测试中平均提升2.2%，性能接近多智能体系统，同时将令牌消耗降低20倍。
链接: https://arxiv.org/abs/2608.27338
作者: Zhichen Zeng,Huiyuan Chen,Jingru Cheng,Juan Zha,Ming Liu,Ying Chen,Xiyuan Yang,Chaosheng Dong,Haiyang Zhang,Hanghang Tong
机构: University of Illinois Urbana-Champaign(伊利诺伊大学厄本那-香槟分校); Amazon(亚马逊)
类目: Multiagent Systems (cs.MA)
备注: 19 pages, 10 figures
Abstract:Specializing Large Language Models (LLMs) toward distinct abilities underpins successes ranging from personalized assistants to multi-agent systems (MAS). Single-agent paradigms rely on pre-defined personas or steering vectors to induce specialization, yet they impose a single fixed specialization that fails to adapt to diverse queries. Conversely, MAS achieves dynamic multi-perspective problem solving by orchestrating agents with distinct text-based roles, but fusing these specializations requires multi-turn interactions that inflate context length and inference cost. To address these limitations, we propose Mixture of Roles (MoRe), which adaptively composes multiple specializations into a single steering vector for single-turn inference. Specifically, MoRe learns a diversified codeboox of steering vectors, each of which encodes a latent role. A query-aware router dynamically fuses the codebook into a steering vector that encompasses multiple roles. By steering the backbone LLM with the composed vector, MoRe enables multi-perspective specialization in a single-agent, single-turn inference process. The proposed MoRe can be efficiently trained via a three-stage SFT curriculum and GRPO post-training, while the backbone LLM remains frozen. Experiments across reasoning and personality benchmarks show that MoRe outperforms single-agent baselines by 2.2% on average, and achieves performance on par with MAS while reducing token cost by 20x.
[MA-1] A Contract-Centered Architecture for Scalable and Manageable Agent ic Runtimes
【速读】：该论文旨在解决企业在部署生成式 AI 时面临的跨组织协同难题，即在业务单元、应用与 AI 团队、测试、平台工程、基础设施、安全、运维及数据治理等多个角色之间，如何有效协调能力（capability）、模型、运行时机制、资源容量和企业数据的权属、变更、准入与可验证性问题。现有用例基准仅能衡量单个智能体完成单一任务的能力，却无法反映上述要素在动态演进过程中的系统性协同机制。其解决方案的核心在于提出四类责任对象（Responsibility Objects）作为组织级共享契约：技能（Skill，可复用、版本化的功能与工作流资产）、驾驭（Harness，运行时编译器与调控器）、脚手架（Scaffold，执行/控制边界及非功能性需求所有者）以及由首席信息官（CIO）独立治理语义与遥测的栈外数据底座。运行时核心架构定义为 A=S,H,X，其中数据底座位于该栈外部。关键贡献在于提出一个可证伪的有限假设 P1（成本感知的能力-容量可分离性）：在预设运行区域内，改变激活的能力时，容量-响应交互关系保持在预先注册的等效容差范围内；而改变兼容的脚手架容量时，能力语义保持不低于劣化阈值，且所需控制开销始终在声明的执行预算内。六个设计条件转化为可量化的义务，其覆盖度、违规情况、不确定性、成本与排除项共同决定 P1 的可判定性。为此，论文设计了一种集群周期随机交叉实验（cluster-period randomized crossover experiment），具备平衡顺序、重置/清洗期、重复种子与故障场景、集群感知不确定性等特征，并引入四状态判决机制（支持、证伪、条件性工程、不确定）。整体贡献包括契约约束的运行时架构、源保真的数据底座以及可证伪的测量协议，但未报告任何实现、实验、数据集或实测结果。
链接: https://arxiv.org/abs/2608.27086
作者: Yaxiao Liu,Pengbo Liu,Yiwen Liu,Yihua Guan,Zhenghe Hou,Jiaxing Song
机构: PwC China AI Center(普华永道中国人工智能中心); Tsinghua University (清华大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注: 48 pages, 7 figures
Abstract:Enterprise AI deployment is a coordination problem across business units, application and AI teams, testing, platform engineering, infrastructure, security, operations, and data governance. Use-case benchmarks show whether one agent completes one task, but not how changing capabilities, models, runtime mechanisms, capacity, and enterprise data should be owned, changed, admitted, or evidenced together. We present four responsibility objects as shared organizational contracts: Skill (reusable, versioned capability and workflow asset), Harness (runtime compiler and governor), Scaffold (execution/control boundary and NFR owner), and a stack-external data substrate under independent CIO-governed semantics and telemetry. The runtime core is A = S, H, X, with the data substrate outside that stack. The central contribution is one bounded, falsifiable hypothesis, P1 (cost-aware capability-capacity separability): within a declared operating region, changing activated capability preserves the capacity-response interaction within a preregistered equivalence margin, while changing compatible Scaffold capacity preserves capability semantics up to a non-inferiority margin, and the required controls stay within a declared enforcement budget. Six design conditions become measured obligations whose coverage, violations, uncertainty, cost, and exclusions determine whether P1 is decidable. We propose a cluster-period randomized crossover experiment (balanced order, reset/washout, repeated seeds and failure regimes, cluster-aware uncertainty) with a four-state verdict: supported, falsified, conditional-engineering, or inconclusive. This paper contributes a contract-bounded runtime architecture, a source-preserving data substrate, and a falsifiable measurement protocol. It reports no completed implementation, experiment, dataset, or measured result. Comments: 48 pages, 7 figures Subjects: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Software Engineering (cs.SE) Cite as: arXiv:2608.27086 [cs.AI] (or arXiv:2608.27086v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.27086 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Yaxiao Liu [view email] [v1] Thu, 27 Aug 2026 13:09:49 UTC (10,287 KB)
[MA-2] DSA: Evidence-Aware LLM -Agent Orchestration for Multi-Market Stock Research STOC
【速读】：该论文旨在解决在多市场股票研究中，如何有效整合异构证据、暴露数据与模型能力的局限性，并对生成式意见的影响进行可控管理的问题。其核心挑战在于构建一个可操作的股票研究系统，使大型语言模型（Large Language Model, LLM）在复杂金融决策场景中既能发挥认知优势，又具备可解释性、可验证性和风险控制能力。解决方案的关键在于提出DSA（Evidence-Aware Orchestration Framework for Multi-Market Stock Research），该框架通过分阶段的工作流设计——包括证据获取、结构化上下文构建、模型路由分析、可选的角色与策略技能推理、以及基于选定上下文和诊断信息的报告生成——实现对整个研究过程的精细化控制。其中，关键创新点在于引入双配置模式：默认报告配置与可选的代理式（agentic）配置，二者共享证据与模型路由服务，但采用不同的输出验证机制与风险防护策略；在代理模式下，核心角色输出经由专用解析器处理，而策略技能意见则经过信号有效性划分后才进行合成，且分歧意见被显式传递至决策代理，再由保守的风险覆盖机制进行干预。这一架构确保了系统在保持灵活性的同时，具备可审计性与稳健性，其参考实现包含六大区域市场路径、十五项捆绑策略技能、本地与托管模型路由及多种执行与交付界面，通过1,457个便携式离线后端合约测试的验证，表明系统在既定软件契约上的实现一致性，但未证明其报告质量、预测准确性或投资回报的优越性。
链接: https://arxiv.org/abs/2608.26990
作者: Linsen Zhu,Yi Shi
机构: 未知
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: 6 pages, 2 figures, 3 tables. Code available at this https URL
Abstract:Large language models can summarize financial information, but an operational stock-research system must first assemble heterogeneous evidence, expose unavailable data and model capabilities, and control how generated opinions affect a final report. We present DSA, an evidence-aware orchestration framework for multi-market stock research with large language model (LLM) agents. DSA organizes the workflow into evidence acquisition, structured context construction, model-routed analysis, optional role and Strategy Skill reasoning, and report generation with selected context and diagnostics. A default report profile and an optional agentic profile share evidence and model-routing services but use profile-specific output validation and risk safeguards. In the agentic profile, core role outputs are processed by role-specific parsers, whereas Strategy Skill opinions undergo an additional signal-eligibility partition before synthesis; disagreement is supplied explicitly to the decision agent, followed by a conservative risk override. The reference implementation includes six regional market paths, fifteen bundled Strategy Skills, hosted and local model routes, and multiple execution and delivery surfaces. At a frozen software snapshot, a selected manifest of 1,457 portable offline backend contract tests passed; 596 cases were retrospectively mapped to six contract families central to the reported LLM-agent architecture. This evidence establishes implementation conformance for the tested software contracts, not superior report quality, forecasting accuracy, or investment returns.
[MA-3] Dynamic Haven Selection for Multi-Agent Pickup and Delivery in Constrained Warehouses ECAI2026 IJCAI
【速读】：该论文旨在解决在空间受限的仓库布局中，多机器人协同完成拾取-配送任务（Multi-Agent Pickup and Delivery, MAPD）时，如何有效保障机器人在执行任务过程中具备安全等待位置（称为Haven）的问题。此类布局通常包含仅容单个机器人通行的狭窄通道和死端工作站，导致机器人在任务间隙难以寻找不阻塞他人的等待位置。现有方法Safe HAven Retreat Planner (SHARP) 通过为每个任务路径预留至专属初始Haven的回撤路径来保证安全性，但其固定Haven分配策略可能导致机器人在完成任务后被引导至距离较远的Haven，降低效率。本文提出A-sharp（Adaptive SHARP），其核心创新在于在任务分配阶段动态调整机器人的回撤目标（retreat target），以优化Haven利用效率。为防止因动态切换引发冲突，A-sharp引入了候选Haven可用性检测机制与待释放规则（pending-release rule），确保前一个Haven在代理离开前仍被保护，避免两个代理依赖同一等待位或已占用/预留位置被穿越。在明确的Haven结构及安全时间路径规划（Safe Interval Path Planning, SIPP）假设下，本文证明了不变量保持性与有限释放完备性：任意有限释放序列中的所有任务均能在有限时间内完成。在4张地图、14,400组地图-机器人数量-任务率-种子组合下共72,000次实验中，SHARP与A-sharp均成功完成全部运行。在任务完成时间（makespan）方面，经预设配对比较并采用Holm校正的138种“Haven数量大于机器人数量”的配置中，A-sharp在107种配置中显著优于SHARP，且从未显著劣于后者；在测试的树状地图上，中位数完成时间减少达16.7%。因此，解决方案的关键在于通过动态自适应的退避目标选择与严格的资源保护机制，在保障安全的前提下显著提升系统整体效率。
链接: https://arxiv.org/abs/2608.26939
作者: Taisei Hirayama,Kohei Yoshida,Hiroki Sakaji,Itsuki Noda
机构: 未知
类目: Multiagent Systems (cs.MA); Robotics (cs.RO)
备注: 19 pages, 4 figures, and 3 tables. Accepted at the Joint Workshop on Planning for Complex Real-World Applications (CAIPI) and Bridging the Gap Between AI Planning and (Reinforcement) Learning (PRL), co-located with IJCAI-ECAI 2026
Abstract:Space-efficient warehouse layouts often contain single-agent-width aisles and dead-end workstations where robots have few places to wait without blocking others. In Multi-Agent Pickup and Delivery (MAPD) on such constrained layouts, robots must accept online pickup-delivery tasks while preserving protected waiting locations called Havens. The Safe HAven Retreat Planner (SHARP) introduced a mechanism that extends each committed task path with a validated retreat to the agent’s dedicated initial Haven, but fixed-Haven commitments can send agents toward distant Havens after deliveries. We present A-sharp (Adaptive SHARP), which changes an agent’s retreat target at task assignment time. A naive switch can cause two agents to rely on the same waiting location or let another committed path pass through a location that is still occupied or reserved. A-sharp prevents these failures with an availability test for candidate Havens and a pending-release rule that keeps the previous Haven protected until the agent departs. Under explicit Haven-structure and Safe Interval Path Planning (SIPP) assumptions, we prove invariant preservation and finite-release completeness: every task in any finite release sequence is delivered in finite time. Across 72,000 runs on 14,400 paired map-agent-count-rate-seed cases over four maps, both SHARP and A-sharp complete their respective 14,400 runs. For makespan (final delivery time), a prespecified paired comparison with Holm correction over all 138 configurations with more Havens than agents finds A-sharp significantly better in 107 configurations and never significantly worse than SHARP; on the tested tree map, the median reduction is 16.7%.
[MA-4] AI agents in Algorithmic Electricity Markets: On the Emergence of Tacit Collusion
【速读】：该论文旨在解决电力市场中基于学习的自主代理（autonomous learning-based agents）在独立决策过程中可能引发隐性共谋（tacit collusion）的问题。随着生成式AI（Generative AI）与强化学习技术在投标策略中的广泛应用，电力市场正逐步演变为算法驱动型市场，而此类市场在结构上具有寡头垄断特征且参与者间存在频繁重复互动，极易催生非竞争性行为。论文的关键解决方案在于：将战略投标建模为具有不完全公共监控的重复博弈，并采用多智能体强化学习（multi-agent reinforcement learning, MARL）模拟市场主体的涌现行为；同时提出一套多维度评估标准（超越传统的利润对比纳什均衡），以识别是否存在隐性共谋的证据。实验结果表明，在未被明确指令协同的情况下，学习代理仍能自发演化出维持超竞争性收益的行为模式，验证了隐性共谋在电力市场中真实存在的可能性。
链接: https://arxiv.org/abs/2608.26896
作者: Jakub Seredyński,Georgios Tsaousoglou
机构: Technical University of Denmark (丹麦技术大学)
类目: Artificial Intelligence (cs.AI); Computer Science and Game Theory (cs.GT); Multiagent Systems (cs.MA); Systems and Control (eess.SY)
备注: 10 pages
Abstract:As electricity market participants increasingly adopt learning-based agents for their bidding strategies, electricity markets are becoming algorithmic. Evidence from algorithmic markets in other domains shows that tacit collusion can arise purely through independent learning. Moreover, electricity markets are typically oligopolistic and feature repeated interaction among a small number of participants, making them structurally susceptible to non-competitive behavior. In the face of these observations, this paper investigates the hypothesis that tacit collusion may emerge in electricity markets where participants’ actions are controlled by autonomous learning-based algorithms. We model strategic bidding as a repeated game with imperfect public monitoring, and model the participants’ emergent behavior using multi-agent reinforcement learning. We propose a multi-dimensional set of criteria (going beyond profit comparisons against Nash equilibria) to assess whether the resulting behavior constitutes tacit collusion. Our experimental results showcase that such a danger is realistic for electricity markets: there are cases where agents do learn to sustain supra-competitive outcomes that are supportive of tacit collusion indicators, even though the agents were never instructed to collude.
[MA-5] LiveSim: Simulating Environment-Shaped Users in Multi-Agent Live-Stream Ecosystems
【速读】：该论文旨在解决现有基于大语言模型（LLM）的用户行为模拟系统在社交密集型场景（如直播生态）中因依赖静态用户画像而导致的行为建模失效问题。传统方法通过历史观测推断用户特征，难以捕捉动态交互下用户行为的持续演化。为此，论文提出LiveSim框架，其核心在于将用户建模为可编辑的行为假设，并通过轨迹驱动的交互逐步优化这些假设；当模拟轨迹与真实观测轨迹出现偏差时，系统自动识别出被忽略的环境塑造效应，并将其提取为可迁移的“环境-行为模式”，累积至集体行为记忆库中，从而提升个体层面的行为保真度，并支持平台级风险演化分析与干预效果评估。该方案的关键在于利用轨迹差异信号实现环境影响的动态识别与知识沉淀，构建具备自适应能力的直播生态系统仿真机制。
链接: https://arxiv.org/abs/2608.26849
作者: Jiaqi Xu,Yiran Qiao,Jing Chen,Qiwei Zhong,Xiang Ao,Xueqi Cheng
机构: University of Chinese Academy of Sciences (中国科学院大学); ByteDance(字节跳动)
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Multiagent Systems (cs.MA)
备注: 20 pages, 8 figures, 7 tables
Abstract:User behavior simulation with large language models~(LLMs) is increasingly used to support multi-agent ecosystem simulation. Existing simulators typically rely on static user profiles inferred from historical observations, which become inadequate in socially intensive environments such as live streaming where interaction dynamics continuously reshape user behavior. We propose \textbfLiveSim, an LLM-based framework for live-stream ecosystem simulation. It represents users as editable behavioral hypotheses and progressively refines them through trajectory-grounded interactions, where discrepancies between simulated and observed trajectories reveal missing environmental shaping effects. These signals are further extracted as transferable environment-behavior patterns and accumulated in a collective behavioral memory to improve user-level behavioral fidelity and support ecosystem-level simulation. Experiments on real-world live-stream risk-control data validate the effectiveness of LiveSim in improving user-level behavioral fidelity and enabling ecosystem-level analysis of risk evolution and platform intervention effects.
[MA-6] Decoupling Planning and Control for Instructable Agents
【速读】：该论文旨在解决当前视觉-语言模型（VLM）在复杂陌生环境中难以将高层规划可靠、低延迟地转化为具体动作序列的问题，同时克服世界模型控制器（world-model controller）缺乏开放式任务引导能力的局限。其核心解决方案在于提出一种名为“Instruct-to-Act”的解耦系统架构：由一个预训练的、指令微调的视觉-语言模型（VLM）作为高层规划器，生成稀疏、高阶且延迟较高的文本指令；这些指令被用于条件化一个世界模型控制器，使其以高频自主执行动作。为使控制器具备语言可指导性，研究通过合成指令对控制器策略轨迹进行重标注，并联合优化行为克隆（behavior cloning）、奖励最大化及世界建模目标。实验在七个具身环境（包括三个多智能体环境）中验证，结果表明该方法在保持快速控制性能的同时，显著优于仅依赖控制器或直接由VLM生成动作的基线方案，且支持无需微调即可更换不同预训练的VLM规划器，整体表现与强视觉-语言-动作及多智能体强化学习基线相当，在七项任务中的六项上实现竞争力。
链接: https://arxiv.org/abs/2608.26788
作者: Zineng Tang,Kelsey R. Allen,Sjoerd van Steenkiste,Ishita Dasgupta,Alane Suhr
机构: UC Berkeley(加州大学伯克利分校); Google DeepMind(谷歌深度思维); University of British Columbia (不列颠哥伦比亚大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Multiagent Systems (cs.MA); Robotics (cs.RO)
备注: Published as a conference paper at COLM 2026. Project page: this https URL
Abstract:Recent work shows that pre-trained, instruction-tuned vision-language models (VLMs) perform well at mapping from instructions and observations to high-level plans, but struggle to realize such plans as reliable low-latency action sequences in unfamiliar environments. At the same time, world-model controllers excel at fast observation-to-action control, but lack open-ended task guidance. In this work, we combine these strengths into a single system, Instruct-to-Act, where we train a world-model controller to act autonomously at high frequency when conditioned on sparse, higher-latency, and high-level text instructions generated by a VLM planner. To train controllers to be language-instructable, we relabel segments of controller policy rollouts with synthetic instructions and jointly optimize a behavior-cloning objective along with existing reward-maximizing and world-modeling objectives. We evaluate our proposed approach across seven embodied environments, including three multi-agent environments where VLM planners coordinate through language while trained controllers serve as their actuators. Under matched observation and action spaces, our decoupled approach consistently outperforms controller-only and direct VLM action-generation variants, preserves fast control, and lets us swap in different pretrained VLM planners without fine-tuning, while remaining competitive with strong vision-language-action and multi-agent RL baselines on six of seven tasks.
[MA-7] Fixed-Haven Reservation for Online Multi-Agent Pickup and Delivery in Dense Warehouses ICAPS2026
【速读】：该论文旨在解决在密集仓库环境中，由于单车道通道、死胡同及树状引导路径导致的多智能体搬运与配送（MAPD）任务执行难题。此类布局下，智能体难以找到空闲等待位置而不阻塞其他路径，而现有MAPD方法通常依赖额外的等待端点或双连通拓扑结构来保证任务完成，这些假设在实际复杂布局中往往不成立。为此，论文提出基于固定安全港（Fixed-Haven Reservation）的在线MAPD解决方案，其中每个智能体拥有一个专属的安全港（Haven），仅其所有者可占据，其余智能体将其视为障碍。通过证明在“安全港可达性”和显式规划/进展假设下，该机制可确保所有有限释放任务均被完成，作者实现了名为SHARP（Safe-Haven Retreat Planner）的算法框架，确保所有忙碌或撤退中的智能体始终沿无碰撞的预留路径返回其安全港。实验表明，在具有树状结构的复杂布局中，SHARP是唯一在所有测试配置下均达到100%成功率的方法，尽管其集中式规划开销较高；此外，对比实验揭示了固定返回机制对系统鲁棒性的关键作用，且禁用中途撤退过程中的重新分配会显著恶化服务时间（提升1.89倍）和总工期（提升1.53倍）。因此，该方案的关键在于利用固定安全港作为强制性回撤锚点，从而实现高鲁棒性的路径规划与冲突避免。
链接: https://arxiv.org/abs/2608.26759
作者: Taisei Hirayama,Kohei Yoshida,Hiroki Sakaji,Itsuki Noda
机构: 未知
类目: Multiagent Systems (cs.MA); Robotics (cs.RO)
备注: 11 pages, 9 figures. Accepted at the 14th Workshop on Planning and Robotics (PlanRob), co-located with ICAPS 2026
Abstract:Dense warehouses often contain single-lane aisles, dead ends, and tree-like guidepaths that leave little room for idle agents to wait without blocking others. Existing Multi-Agent Pickup and Delivery (MAPD) guarantees for completing all finitely released tasks typically rely on extra waiting endpoints that planned paths can avoid, or on biconnected topology; these assumptions may fail in such layouts. We study fixed-Haven reservation for online MAPD, where pickup-delivery tasks are released over time. Each agent owns a fixed Safe Haven (Haven for short), usually its start cell, that only the owner may occupy and that other agents treat as blocked. For finite task releases, we prove that this fixed-Haven contract completes all released tasks under Haven-Reachability and explicit planning/progress assumptions. We implement the contract in SHARP, a Safe-Haven Retreat Planner that keeps every busy or retreating agent on a collision-free reserved route ending at its Haven. We compare SHARP with representative TP and PIBT-family MAPD baselines: Token Passing (TP), Priority Inheritance with Backtracking (PIBT), and PIBT with Temporary Priority and Temporary Avoidance (PIBTTP-TA) for biconnected main areas with attached trees. In the robustness sweep, SHARP is the only method with 100% success on all tested configurations, at substantially higher centralized planning cost on tree-like layouts. A TP-style fixed-home-return counterfactual with full-route validation also recovers robustness on tested tree-like layouts, suggesting that fixed return is a central robustness mechanism there. A no-overwrite variant shows that disabling mid-retreat reassignment worsens service time (release-to-delivery latency) by 1.89 times and makespan by 1.53 times in the tested high-load tree condition.
[MA-8] SIGMA: Structured Noise-Effect-Aware Grouped Multi-Agent Aggregation
【速读】：该论文旨在解决合作式多智能体强化学习（Cooperative Multi-Agent Reinforcement Learning, MARL）在观测噪声环境下难以保持鲁棒协作的问题。现有方法通常假设观测扰动在各智能体间独立分布，但实际中这些噪声通过底层协作结构产生具有局部相关性的下游影响，形成“结构性噪声效应”（structured noise effects），即在任务依赖性强的局部智能体间表现出相关性，而在全局范围内则呈现异质性。传统鲁棒MARL方法未显式建模或利用此类结构依赖的噪声特性，导致性能受限。为此，本文提出SIGMA（Structured Noise-aware Hierarchical Collaboration Framework），其核心在于构建一种分层协作机制：首先基于密度聚类将智能体自适应划分为局部协作结构，并通过组内共识聚合保留任务相关的共享信息，同时抑制个体表示偏差；随后通过组间注意力机制动态融合不同局部结构的信息，在保持全局协调性的同时适应各组的异质贡献。实验在《星际争霸II》（StarCraft II）的带噪观测任务上验证了结构性噪声效应的存在，并表明SIGMA在噪声环境下显著提升鲁棒性，同时在无噪声场景下仍保持竞争力。
链接: https://arxiv.org/abs/2608.26683
作者: Li Mingqian
机构: 未知
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注:
Abstract:Cooperative multi-agent reinforcement learning (MARL) faces significant challenges in maintaining robust coordination under noisy observations. Although observation disturbances are often introduced independently across agents, their downstream effects on cooperative decision-making can become structured through underlying cooperation structures. We characterize this phenomenon as structured noise effects, where noise-induced decision effects exhibit local correlation among agents with stronger task-related dependencies while remaining globally heterogeneous across different agents and local structures. Existing robust MARL methods, however, rarely explicitly characterize or exploit such structure-dependent noise effects. To address this limitation, we propose SIGMA, a hierarchical collaboration framework that exploits cooperation structures to learn robust representations under noisy observations. SIGMA first organizes agents into adaptive local structures through density-based grouping and performs intra-group consensus aggregation to preserve shared task-relevant information while smoothing agent-specific representation deviations. Inter-group attention then adaptively integrates information across different groups to preserve global coordination while accommodating their heterogeneous contributions. Experiments on noisy-observation tasks in StarCraft II empirically validate the structured noise effects and demonstrate that SIGMA consistently improves robustness under observation noise while maintaining competitive performance in noise-free environments.
[MA-9] Risks and Controls for Multi-Agent Systems: an analytical framework for deployment of AI agents across organisational boundaries WWW
【速读】：该论文旨在解决多智能体系统中因人工智能代理（AI agents）相互交互而产生的新型风险问题，特别是当这些交互跨越组织边界时，所引发的治理盲区与失控风险。随着组织内部及外部合作伙伴、客户、供应商乃至开放互联网上的智能体之间交互日益频繁，单一组织难以全面感知、控制或管理跨边界的交互行为，从而导致潜在的系统性失败。其解决方案的关键在于提出一个基于治理层级的三阶框架：（1）单一治理（singular governance），即由单一组织统一管控所有代理；（2）联邦治理（federated governance），多个组织在共享环境中依据共同规则协作部署；（3）开放环境（open environments），代理在无中央权威下运行，仅依赖自愿采纳的共享标准。该框架在每一层级中系统分析风险因素、故障模式及可实施的控制措施，并明确责任主体；对于治理空白区域，则识别出所需的集体行动机制，以推动跨组织协同治理，从而构建适应复杂交互场景的韧性治理体系。
链接: https://arxiv.org/abs/2608.26626
作者: Alistair Reid,Simon O’Callaghan,Dustin Venini,Liam Carroll,Tiberio Caetano
机构: 未知
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI)
备注: This paper has been published by the Australian AI Safety Institute within the Department of Industry, Science and Resources under a CC BY 4.0 licence: this https URL
Abstract:This report presents a framework to help organisations, policymakers and researchers reason about the risks that emerge when AI agents interact with each other, how those risks change as interactions cross organisational boundaries, and the controls that may help address them. As organisations deploy AI agents, those agents will increasingly interact with each other: inside the organisation, with the agents of partners, customers and suppliers, and with unknown counterparties on the open internet. Failures can emerge from the interactions themselves, and once those interactions cross an organisation’s perimeter, no single organisation can fully see, control or govern them. The report introduces three deployment tiers, defined by the minimum common governance binding any two interacting agents: singular governance, where one organisation governs every agent; federated governance, where multiple organisations deploy into a shared environment under agreed rules; and open environments, where agents operate with no central authority and shared standards are adopted voluntarily if at all. Within each tier, the report examines risk factors, failure modes and available controls. It identifies who is positioned to apply the controls, and where no actor is positioned to act, it characterises the gap and the collective action required to close it. Comments: This paper has been published by the Australian AI Safety Institute within the Department of Industry, Science and Resources under a CC BY 4.0 licence: this https URL Subjects: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI) Cite as: arXiv:2608.26626 [cs.MA] (or arXiv:2608.26626v1 [cs.MA] for this version) https://doi.org/10.48550/arXiv.2608.26626 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Alistair Reid [view email] [v1] Thu, 27 Aug 2026 05:28:06 UTC (2,123 KB)
[MA-10] Zero-Shot Self-Orchestration with Ledger-Based Control for Improved LLM Coding Performance
【速读】：该论文旨在解决多智能体大语言模型系统在实际性能评估中存在混淆变量（如提示词、工具调用和令牌预算的同步变化）导致结果不可靠的问题，试图明确“管理者-工作者”架构（manager-worker scaffold）在不进行训练或特定基准微调的前提下，对代码生成任务的真实贡献。其核心解决方案在于引入基于共享文件系统的管理者-工作者框架，通过结构化协作提升推理效率与准确性。关键创新点在于：通过管理者协调多个工作者并利用共享状态记录（如短调用与共享笔记）实现上下文管理，有效缓解序列截断问题，并促进问题分解。实证结果显示，该架构在部分模型上带来显著且统计显著的性能提升（如Qwen3.8-27B在五次配对推理中提升23.4%），尤其在关闭推理模式时对中小模型效果更佳；同时，在保持较低成本的情况下，其精度增益远超直接升级模型规模——例如，使用管理者后GPT-5.6-Terra仅需原价五分之一即可接近Fable 5的单次调用表现，验证了该架构在资源-性能权衡上的高效性。
链接: https://arxiv.org/abs/2608.26480
作者: Victor Gao,Vida Khosrowshahi,Ali Khosrowshahi,Xihao Sun,Juhyun Lee,Simon(Sang Won)Lee
机构: Persis Capital Inc.(Persis资本公司); Xihao Sun; Juhyun Lee; Simon (Sang Won) Lee
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Software Engineering (cs.SE)
备注:
Abstract:Multi-agent large language model systems are widely reported to beat single-model baselines, but the evidence is mixed, and comparisons are usually confounded: pipelines change token budgets, tool calls, and prompts simultaneously, so an aggregate gain rarely reveals what actually helped. We investigate the effect of introducing the manager-worker scaffold over a shared filesystem workspace, with no training and no per-benchmark tuning, measured against the same model answering in a single pass. Across nine models – five open-weight, spanning 9B to ~2.8T parameters, and four frontier closed models – on the 100 latest hard LiveCodeBench problems, the scaffold’s benefit is real but conditional: large and statistically significant for some (Qwen3.8-27B +23.4, GPT-5.6-Luna +10.6 and GPT-5.6-Terra +8.0, each over five paired passes; Kimi-K3 +30.4 and Minimax-M3 +11.0 over five paired passes with reasoning off, both at p 10^-4 , and +42 and +12 in a single pass at a 128k cap) and null or negative for others (Qwen3.6-35B -1 to -9 with reasoning off). With the manager, Opus-5 achieves the highest score in the study at 91% in one pass. Running a manager roughly triples the token bill, but it buys accuracy more cheaply than moving to a larger model does: GPT-5.6-Terra with a manager nearly matches Fable 5’s single-call accuracy (85.0 against 87.4, p = 0.59 ) at a fifth of the price (\ 11.71 against \ 61.11 per 100-problem pass, p 10^-4 ), and the Qwen-27B arm does it for \ 51.75 on weights anyone can self-host. Our transcript analysis finds several mechanisms behind the gains, of which two recur: context management, in which short worker calls and shared notes organize state and reduce truncation, and problem decomposition. Improvements are modest for large models with reasoning enabled, but larger for some models with reasoning disabled and for smaller models with reasoning enabled.
[MA-11] MemToC: Benchmarking Memory-Tool Conflict Resolution in Large Language Models
【速读】：该论文旨在解决工具增强型大语言模型（Tool-augmented LLMs）在工具返回结果与模型参数记忆存在冲突时，如何进行有效仲裁的问题。现有评估方法仅衡量模型对不同信息源的偏好，却未明确判断哪一方来源正确，导致评估结果缺乏可靠性。为此，作者提出MemToC——一个基于可执行工具的受控基准测试框架，通过6,504个由542个高质量事实问题构建的评估场景，独立获取各模型的闭卷回答，并设定已知正确性的工具返回结果，从而精确区分四种来源正确性情况（包括工具错误和无工具使用作为对照）。实验表明，在五个7-9B规模的开源模型中，工具返回结果显著主导模型决策；四款指令微调模型在面对正确工具与错误参数记忆时，仅6.5%-17.1%的案例能保留正确答案，而86.0%-93.1%的情况会采纳正确工具，当两者均错误时则有78.4%-86.0%重复工具输出。此外，模型表现对提示词形式高度敏感，跨模型排序在固定问题内容下不稳定。研究进一步采用链级交叉拟合方法（chain-level cross-fitting）对比监督微调（SFT）与直接偏好优化（DPO），确保共享底层事实的问题不跨越训练与评估集。采用非对称成功标准：正确答案保留率提升的同时，不得出现正确工具采纳率下降。结果显示，仅有两种指令微调骨干网络在该标准下取得改进。值得注意的是，多数方法虽提升正确响应，但往往伴随拒绝回答（abstention）增加，且迁移效果有限，依赖具体模型与呈现方式。结论表明，基于正确性条件的仲裁能力可通过微调改善，但性能提升必须综合考量正确工具使用、拒绝行为及对提示形式的鲁棒性。
链接: https://arxiv.org/abs/2608.26295
作者: Arseniy Varlamov,Rishat Zinnatullin,Elisei Rykov,Alexander Panchenko,Ilseyar Alimova
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注: 26 pages, 2 figures
Abstract:Tool-augmented LLMs must arbitrate between two fallible sources when a tool return conflicts with their parametric memory, yet existing evaluations measure source preference without establishing source correctness. We introduce MemToC, a controlled benchmark for post-tool-return arbitration with executable tools. MemToC comprises 6,504 evaluation episodes constructed from 542 quality-controlled factual questions, independently elicited model-specific closed-book answers, and controlled tool returns of known correctness. These components instantiate four source-correctness cases; tool-error and no-tool conditions are separate controls. Across five open-weight 7-9B models, tool returns strongly dominate elicited closed-book answers. The four instruction-tuned models retain a verified-correct answer against an incorrect tool in only 6.5-17.1% of eligible cases, follow a correct tool in 86.0-93.1%, and repeat the tool return in 78.4-86.0% of cases where both sources are wrong. No cross-model ordering remains stable across three instruction-wording variants with the question and episode content held fixed. We compare prompting with SFT and DPO using chain-level cross-fitting over ToolHop, so questions sharing an underlying fact never straddle training and evaluation. We apply an asymmetric success criterion: correct-answer retention must improve without a detected reduction in correct-tool following. SFT and DPO meet this criterion on the same two of four instruction-tuned backbones. Improvements rarely come cleanly: 19 of 20 tested method-model combinations reduce abstention after tool errors or on unanswerable inputs. Transfer beyond MemToC is positive but partial and depends on the model and presentation frame. Correctness-conditioned arbitration can be improved through fine-tuning, but gains must be evaluated jointly with correct tool use, abstention, and robustness to formulation.
[MA-12] SKILL.state: Scalable Long-Horizon Agent Skills EMNLP
【速读】：该论文旨在解决大语言模型（Large Language Models, LLMs）作为自主代理执行复杂、长时序过程性任务时，因持续累积对话历史（包括观测、动作和中间推理轨迹）而导致的延迟恶化与上下文污染（context-poisoning）问题。现有运行时架构依赖于追加式（append-only）对话历史，导致提示词长度随执行时间线性增长，严重影响模型性能与可扩展性。其解决方案的关键在于引入一种显式的、可变的执行状态（explicit, mutable execution state）机制，替代传统的追加式历史记录。在每个执行步骤中，模型仅接收不可变的技能规范、当前结构化的执行状态以及最新的观测结果，而中间推理过程在生成有效状态更新后立即丢弃，从而实现提示词长度的恒定，避免上下文膨胀。该方法在多种数据集、模型及执行环境中均显著提升了任务准确率，并大幅降低了累计的令牌消耗。研究结果表明，显式执行状态是一种高效且与架构无关的抽象方式，能够支持可扩展的长时序代理技能执行。
链接: https://arxiv.org/abs/2608.26263
作者: Sanket Badhe,Priyanka Tiwari,Jonghyun Chung
机构: Google LLC(谷歌); Purdue University(普渡大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: accepted at EMNLP
Abstract:Large Language Models (LLMs) increasingly act as autonomous agents executing complex, long-running procedural skills. Existing agent runtimes maintain execution by continually appending observations, actions, and intermediate reasoning traces to an ever-growing conversation history, causing latency degradation and context-poisoning failures over long horizons. We present this http URL, a runtime architecture that replaces append-only conversational history with an explicit, mutable execution state. At each execution step, the model receives only the immutable skill specification, the current structured execution state, and the latest observation. Intermediate reasoning is discarded immediately after producing a validated state update, preventing prompt growth with execution history. Across diverse datasets, models, and execution environments, this http URL improves task accuracy while substantially reducing cumulative token consumption. Our results demonstrate that explicit execution state is an effective and architecture-agnostic abstraction for scalable long-horizon agent skills.
[MA-13] Agent Mesh: Reliability Primitives for Non-Idempotent Agent Delegation - Identity Adequacy and Evidence Adequacy
【速读】：该论文旨在解决生成式 AI（Generative AI）在生产环境中作为自治代理（autonomous agents）执行软件交付任务时，因依赖服务网格（service mesh）中的重试、超时与错误率熔断等机制而导致的可靠性失效问题。研究通过对147个编号故障事件的实证分析发现，这些基础机制所依赖的三个核心假设——即错误可被正确识别、失败信号具有可变性且可归因、以及相同条件下行为具有确定性——在实际运行中均被违反。关键问题在于：身份识别不足导致系统无法区分真实故障与正常波动，证据充分性缺失使得可靠性决策缺乏可追溯性和确定性。因此，论文提出以“委托”（delegation）为基本单元而非消息进行可靠性控制的七种新原语，并强调应基于受控评估来验证其有效性，从而构建更鲁棒的自治代理系统。
链接: https://arxiv.org/abs/2608.26225
作者: Mazhar Shaikh,Anurag Rajkumar Bombarde,Harshal Pathak
机构: 未知
类目: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注: 14 pages, 9 figures, 6 tables; 5-page ancillary supplement
Abstract:Autonomous agents increasingly perform bounded software tasks under an orchestrator that retries, resumes, and budgets them. The machinery such orchestrators reach for is the service mesh’s: retry, timeout, and error-rate circuit breaking. We report a failure study of a production agentic software-delivery platform over 147 numbered incidents spanning 81 runs, each with a measured cost and, in most cases, a mutation proof reproducing the failure. All three assumptions those primitives rest on are violated in practice, and we quantify the consequences: a loop of fifty-four consecutive successful tool calls no error-rate breaker could see; a progress signal constant by construction, guaranteeing a false trip on the third repair round and driving one run from six of six components to three; twenty-one events accumulated across six invocations of one delegation, making a correct, idempotent component unwinnable; a misrouted failure that woke five components for a two-component fault, leaving three bystanders regressing working code; and twelve incidents in which the enforcement layer blocked correct work, the most expensive costing 107 agent turns and zero accepted writes. We find one cross-cutting cause and its dual. Identity adequacy: in five separate subsystems an identity that failed to discriminate produced a confident wrong answer, and two of them derived the corrective rule independently. Evidence adequacy: a reliability decision may be taken only on evidence capable of moving, attributable to what it measures, and deterministic under identical conditions. From the findings we derive seven reliability primitives whose enforcement unit is the delegation rather than the message, and specify the controlled evaluation the study motivates but does not constitute.
[MA-14] From Sound to Symptom: Real-Time Respiratory Signal Understanding for Conversational Healthcare Agents SIGDIAL2026
【速读】：该论文旨在解决现有对话系统将咳嗽等呼吸性副语言事件视为声学噪声并直接丢弃的问题，而这些事件实际上蕴含着重要的临床呼吸信号。其核心挑战在于如何在不中断实时对话流的前提下，实现对咳嗽等呼吸行为的高精度、细粒度分析与实时监测。解决方案的关键在于提出HealthCUES（Clinical Understanding from Embodied Sounds），一个面向具身声音的流式处理管道，通过与对话轮次边界对齐的滚动音频缓冲机制，实现亚秒级事件检测。系统不仅支持二元咳嗽检测，还进一步实现了咳嗽与清嗓的区分、咳嗽亚型（干咳、湿咳、犬吠样、哮鸣样）分类及起止时间边界估计，并引入对话感知的触发门控机制以缓解警报疲劳。其核心技术依托于多模态大语言模型（Multimodal Large Language Model, MLLM）Qwen3Omni，采用受限结构化输出策略，将咳嗽分析分解为并行预测任务，支持独立提示优化。实验结果表明，该系统在自建数据集上达到93%的F1值，湿咳/干咳分类加权F1为0.75，端到端延迟仅340ms；外部验证在AMI会议语料库中表现出色，宏平均F1达0.91，证实了其在语音混合环境下的鲁棒性。用户研究进一步验证了亚型信息的临床价值及其在远程医疗工作流中的实用性。
链接: https://arxiv.org/abs/2608.26163
作者: Tanmay Laud,Herprit Mahal,Subhabrata Mukherjee
机构: Hippocratic AI
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Multiagent Systems (cs.MA); Sound (cs.SD)
备注: Accepted for publication at SIGDIAL 2026
Abstract:Cough events during live spoken conversations carry clinically valuable respiratory signals, yet existing dialogue systems treat them as acoustic noise to be discarded. We present HealthCUES (Clinical Understanding from Embodied Sounds), a streaming pipeline for paralinguistic respiratory monitoring in real-time conversational agents, a capability that, to the best of our knowledge, is absent from all prior systems. HealthCUES processes audio through a rolling buffer aligned with dialogue turn boundaries, enabling sub-second event detection without interrupting conversational flow. Beyond binary cough detection, the system provides fine-grained analytics: (i) differentiation between coughing and throat clearing, (ii) cough subtype classification (dry, wet, barking, whooping) with confidence scores, and (iii) temporal duration estimation with start-end boundaries. To prevent alert fatigue, HealthCUES introduces dialogue-aware gating mechanisms that modulate triggering based on conversational context. The system leverages Qwen3Omni, a multimodal large language model (MLLM), with constrained structured outputs, decomposing cough analysis into parallel prediction tasks for independent prompt optimization. Evaluation on 847 in-house conversational audio segments demonstrates 93% F1 for cough detection, 0.75 weighted-F1 for wet/dry subtype classification, and average end-to-end latency of 340ms; external validation on the AMI meeting corpus confirms robust cough, throat-clearing, and speech separation in the presence of speech (0.91 macro-F1). A user study with licensed healthcare professionals confirms the clinical relevance of subtype information and the system’s utility in telehealth workflows.
[MA-15] Prompt Sensitivity of Generative Agents : Evidence from an Epidemic Model
【速读】：该论文旨在解决生成式智能体（Generative Agents）在模拟人类行为时对提示（prompt）敏感性及人格设定（persona name）影响的可预测性问题。其核心挑战在于：当生成式智能体被用于构建复杂社会行为模型（如流行病传播模拟）时，其行为是否对提示文本的细微变化或人格标签的调整具有显著响应。解决方案的关键在于通过构建一个基于生成式智能体的流行病传播模型，系统性地测试不同提示变体与人格名称对整体行为模式的影响。研究发现，使用语义相近的提示词仅带来可忽略的行为差异，而微小的提示变化或上下文调整则会实质性改变模型结果；此外，赋予不同人格名称（尤其是具有特定人格特征的名称）并未显著影响流行病传播的最终结果。这一发现表明，生成式智能体的行为主要受提示内容的动态语境驱动，而非静态的人格标签，为未来在社会仿真中使用生成式智能体提供了重要的设计依据与验证框架。
链接: https://arxiv.org/abs/2608.26221
作者: Ross Williams,Niyousha Hosseinichimeh
机构: 未知
类目: Physics and Society (physics.soc-ph); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注: 14 pages, 3 figures. Code and data: this https URL
Abstract:As generative AI gains traction, researchers are investigating its potential to serve as proxies for humans. From undergoing cognitive psychology experiments to experiencing an epidemic, generative agents, agents powered by generative AI models, produce realistic human behavior when prompted. This study explores the sensitivity of these generative agents’ behavior to prompt modifications and varied persona names of the agents. To assess this sensitivity, we use a generative agent epidemic model, wherein each agent is prompted daily on whether it wants to isolate or commingle with other agents. We found that using synonymous prompts results in negligible changes to the model’s outcomes. However, minor variations in prompts, as well as contextual changes, do influence the model’s results. Lastly, our data indicates that different persona names assigned to generative agents, specifically those imbued with personas, do not significantly impact epidemic outcomes.
自然语言处理
[NLP-0] CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在推理阶段依赖重复生成或外部验证以提升推理性能所带来的高计算开销与低效问题。现有方法虽能通过测试时扩展（test-time scaling）改善表现，但往往伴随显著的生成次数增加和更高的令牌消耗。为此，论文提出了一种名为CritICL的新颖推理时框架，其核心创新在于将模型在不同规模下的失败模式（failure modes）视为可利用的结构化信息而非单纯负面输出。关键思想是：利用较弱模型产生的失败模式作为批判性提示（critique-based in-context examples），引导更强模型进行更准确的推理。该方法设计了两种变体——CritICL-dynamic通过动态预测输入相关的失败模式并检索相应批判内容，实现个性化指导；CritICL-static则基于全局失败模式概要提供稳定一致的引导。实验表明，CritICL在保持高效的同时，显著优于标准上下文学习，并在多数任务上达到或超越测试时扩展方法的性能，且生成次数与令牌成本大幅降低。
链接: https://arxiv.org/abs/2608.27455
作者: Yufan Wu,Yinghui He,Zhengyi Hu,Lang Wei,Ruichen Li,Qifan Yang,Ting Zhu
机构: The Ohio State University(俄亥俄州立大学); Princeton University(普林斯顿大学)
类目: Computation and Language (cs.CL)
备注: this https URL
Abstract:Recent advances in inference-time scaling have significantly improved the reasoning performance of large language models (LLMs). However, these methods typically rely on repeated generation or external verification. To address this limitation, we introduce CritICL, a novel inference-time framework that improves reasoning while maintaining high efficiency. Our key insight is that LLM failure modes exhibit structured patterns across model scales within the same family. Instead of treating failures as undesirable outputs, CritICL leverages them as a source of guidance. Specifically, we utilize failure modes derived from weaker models and incorporate them into inference through critique-based in-context examples. We propose two variants: CritICL-dynamic, which adaptively predicts input-specific failure modes and retrieves critiques, and CritICL-static, which uses a global failure mode profile to provide stable guidance. Experimental results show that CritICL consistently outperforms standard in-context learning and achieves performance competitive with or superior to test-time scaling methods, while requiring significantly fewer generations and lower token cost. Code available at: this https URL
[NLP-1] WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution
【速读】： 该论文旨在解决当前生成式AI代理（Agent）在技能演化过程中知识碎片化、难以系统复用的问题。现有方法虽能从代理的交互经验中自动发现技能，但关键的优化洞察往往分散于历史记录中，导致技能发展缺乏连贯性与可重用性。其解决方案的核心是提出WikiSkill框架，通过构建一个持续演化的持久化知识库（wiki），将原始执行经验、累积知识与可执行技能进行分离，并持续将经验提炼并整合至wiki中，供后续技能迭代使用。该设计实现了对代理经验的系统性积累与精炼，显著提升了技能的可复用性与迁移能力。实验表明，该框架在多种基准和模型上均优于现有最优技能演化方法，且技能演化与模型规模具有互补效应：大模型更受益于演化技能，而小模型借助技能可超越未使用技能的大模型。此外，演化技能具备跨模型与模型家族的良好迁移性，甚至外部模型生成的技能表现优于自演化技能。消融实验证实，知识库的持续积累是有效技能演化的关键。这一成果证明了系统性积累与精炼代理经验对于构建可复用、可迁移技能的重要价值。
链接: https://arxiv.org/abs/2608.27454
作者: Liyan Tang,Cyrus Rashtchian,Chun-Sung Ferng,Andrew Tomkins,Da-Cheng Juan,Tu Vu
机构: Google Research(谷歌研究院); Virginia Tech(弗吉尼亚理工学院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Agent skills package specialized knowledge and workflows into reusable resources that extend AI agent capabilities. Recent work automatically discovers such skills from agent experience, which enables agents to progressively adapt through interaction. However, the insights that guide skill development typically remain scattered across optimization histories, limiting their systematic reuse across iterations. We introduce WikiSkill, a framework that co-evolves agent skills with a persistent knowledge base (wiki). At a high level, WikiSkill separates raw execution experience, accumulated knowledge, and executable skills, while continuously consolidating experience into the wiki, which subsequent skill updates can build on. Across diverse benchmarks and models, WikiSkill consistently outperforms state-of-the-art skill-evolution methods and improves over no-skill baselines in most model-benchmark settings. We find that skill evolution complements model scaling: larger models generally benefit more from evolved skills, while smaller models with skills can outperform substantially larger models without them. We also find that evolved skills transfer effectively across models and model families, and skills evolved by other models can outperform self-evolved skills. Finally, our ablation studies confirm that persistent knowledge accumulation in the wiki is critical for effective skill evolution. These results demonstrate the benefits of systematically accumulating and refining agent experience for developing reusable and transferable skills.
[NLP-2] SWE-Prime: Fewer Trajectories Better Performance
【速读】： 该论文旨在解决大语言模型在处理真实软件问题时，因依赖低质量监督信号而导致的性能瓶颈问题。现有方法通过构建大规模代理轨迹数据集并基于成功轨迹进行有监督微调（SFT），但成功轨迹中仍可能包含无效、冗余或存在风险的操作步骤，直接使用此类数据会引入噪声监督，促使模型学习到不良的问题求解行为。为应对这一挑战，本文提出SWE-Prime，一种多粒度、两阶段的数据筛选方法：第一阶段在轨迹层面依据过程质量、结果质量和数据代表性进行筛选，提取高质量且具有代表性的成功轨迹子集；第二阶段在段落层面将连续步骤聚类为语义段，并基于其对最终解决方案的贡献度、可学习性及潜在风险进行评估，仅保留关键段落参与损失计算，同时保持完整上下文序列。实验结果表明，在SWE-Bench Pro和SWE-Bench Verified基准上，使用SWE-Prime筛选出的10%轨迹子集进行微调，性能分别提升12.2%和24.2%，显著优于全量成功数据集的训练效果，验证了其在提升监督质量与模型表现方面的有效性。
链接: https://arxiv.org/abs/2608.27449
作者: Dewu Zheng,Ruizhe Ye,Yanlin Wang,Yang Ye,Hongyu Zhang,Ensheng Shi,Xilin Liu,Yuchi Ma,Jianxing Yu,Zibin Zheng
机构: 未知
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 9 pages, 5 figures
Abstract:To improve large language models’ ability to resolve real-world software issues, prior work has focused on constructing large-scale agent trajectory datasets and performing supervised fine-tuning (SFT) on successful trajectories. However, task success does not guarantee high-quality supervision: successful trajectories may still contain ineffective, redundant, or risky steps. Directly using such trajectories for SFT can introduce noisy supervision and encourage models to imitate undesirable problem-solving behaviors. Therefore, we propose SWE-Prime, a multi-granularity, two-stage SFT data selection method that progressively filters training data at the trajectory and segment levels. Specifically, the first stage performs trajectory-level screening based on process quality, result quality, and data representativeness, selecting a high-quality and representative subset of successful trajectories. The second stage performs segment-level selection by grouping consecutive steps into semantic segments and assessing each segment based on its contribution to the final solution, learnability, and potential risks. During SFT, all segments remain in the sequence to preserve context, while only selected segments contribute to the loss computation. Experiments on SWE-Bench Pro and SWE-Bench Verified show that training on the 10% trajectory subset selected by SWE-Prime outperforms training on the full resolved dataset, yielding relative performance gains of up to 12.2% and 24.2%, respectively.
[NLP-3] PO: Test-Time Policy Optimization
【速读】： 该论文旨在解决大语言模型在测试时训练（Test-Time Training, TTT）中缺乏真实标签导致的可训练性问题，尤其针对现有后训练方法（如强化学习RL与基于策略的自蒸馏OPSD）依赖于高质量真值标签而无法在无标签环境下进行有效优化的局限。其核心挑战在于：使用多数投票伪标签虽为替代方案，但一旦出现错误投票，将导致教师模型被污染，并引发整个推理过程的系统性偏差。为此，作者提出测试时策略优化（Test-Time Policy Optimization, TTPO），其关键创新在于构建一种非对称目标函数——通过基于策略的自蒸馏（OPSD）对与伪标签一致的推演路径进行知识提取，同时利用分组强化学习（Grouped RL）对不一致路径施加惩罚。该设计充分利用了“不一致推演通常错误”的不对称特性，使模型即使在频繁伪标签错误的情况下仍能保持更新的稳健性。进一步地，引入逐标记选择机制，在蒸馏分支中降低已收敛位置的权重，在强化学习分支中仅惩罚高置信度错误，从而实现更精细的梯度引导。实验表明，TTPO在无需任何真实标签的情况下，性能达到监督式OPSD水平，在多个竞赛级数学推理基准上显著提升模型表现（如Qwen3-1.7B从38.0%提升至45.2%），且展现出强大的跨任务泛化能力。
链接: https://arxiv.org/abs/2608.27448
作者: Aozhe Wang,Zhengxi Lu,Jianze Wang,Shangke Lv,Ying Liu,Weiming Lu,Jun Xiao,Yueting Zhuang,Hua Yang,Qianglong Chen,Yongliang Shen
机构: Zhejiang University (浙江大学); Alibaba Group (阿里巴巴集团)
类目: Computation and Language (cs.CL)
备注: Project Page: this https URL Code: this https URL
Abstract:Recent prominent post-training methods, such as Reinforcement Learning (RL) and On-Policy Self-Distillation (OPSD), have driven rapid progress in mathematical reasoning for large language models, yet their reliance on ground-truth labels precludes test-time training (TTT). Replacing ground truth with majority-vote pseudo-labels is a natural alternative, yet it is fragile: an incorrect vote corrupts the teacher and misleads every token. We observe that this failure mode is asymmetric: rollouts that disagree with the pseudo-label are typically wrong regardless of whether the vote itself is correct. Building on this observation, we propose Test-Time Policy Optimization (TTPO), an asymmetric objective that distills agreeing rollouts via OPSD and penalizes disagreeing rollouts with Grouped RL. Token-level selection further refines both branches: distillation down-weights already-converged positions, while RL penalizes only confident errors. Both updates remain well-grounded even under frequent pseudo-label errors, and majority-vote routing yields tighter self-supervision as the model improves. Without any labels, TTPO matches label-supervised OPSD on five competition-level benchmarks, raises Qwen3-1.7B from 38.0% to 45.2% in TTT, yields +25.2% to +36.4% without thinking, and shows strong cross-task generalization.
[NLP-4] From Static to Dynamic: Benchmarking Real-World Code Review with MCR-Bench ISSTA2026
【速读】： 该论文旨在解决现有大语言模型（LLM）在自动化代码审查中普遍存在的“单轮静态决策”局限性问题，即当前方法未能真实反映现实场景中代码审查所具有的多轮交互性和复杂问题求解过程。其核心挑战在于：现有模型难以有效捕捉缺陷在多轮交互中的演化状态，导致缺陷检测与生命周期追踪能力不足。解决方案的关键在于提出首个面向真实多轮代码审查的缺陷状态感知基准——MCR-Bench，该基准包含2,269个来自五种主流编程语言的真实多轮代码审查任务，并通过细粒度缺陷元数据（如缺陷描述、类型、严重程度）及跨轮次状态标签，全面刻画缺陷在整个审查流程中的动态演变轨迹。基于该基准的实验揭示了主流LLM在缺陷检测与状态追踪方面的三大短板：整体能力有限、对不同类型和严重程度缺陷表现不一，以及存在跨轮次时间错位与长程记忆缺失等根本性缺陷机制。因此，该研究的核心创新在于构建了一个高保真、可量化评估的多轮交互式代码审查基准体系，为提升生成式AI在复杂协作式开发场景中的可靠性提供了关键基础。
链接: https://arxiv.org/abs/2608.27442
作者: Dewu Zheng,Yanlin Wang,Xiwen Wang,Kefeng Duan,Hongyu Zhang,Xilin Liu,Yuchi Ma,Zibin Zheng
机构: Sun Yat-sen University (中山大学); Zhuhai Key Laboratory of Trusted Large Language Models (珠海市可信大语言模型重点实验室); Chongqing University (重庆大学); Huawei Cloud Computing Technologies Co., Ltd. (华为云计算技术有限公司)
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted at ISSTA 2026
Abstract:In real-world software development, code review typically involves iterative interactions between developers and reviewers to improve software quality, making the process costly and time-consuming. Although recent work explores large language models (LLMs) for automated code review, most approaches oversimplify code review into a single-round, static decision task, which fails to capture the multi-round interactive nature and the complex problem-solving processes inherent in realistic review scenarios. To bridge this gap, we introduce MCR-Bench, the first defect state-aware benchmark designed for realistic multi-round code review. MCR-Bench covers five commonly-used programming languages and consists of 2,269 real-world multi-round code review tasks, each of which is annotated with fine-grained defect information and cross-round state labels. Each task in MCR-Bench is equipped with fine-grained defect metadata (e.g., description, type, severity) alongside dynamic state annotations, capturing the complete evolutionary trajectory of a defect throughout the multi-round process. We obtain several findings through extensive experiments on MCR-Bench with mainstream LLMs. (1) Limited overall capability: experiments reveal that mainstream LLMs exhibit limited overall performance in defect detection and defect lifecycle state tracking, with performance degrading significantly as the number of interaction rounds increases; (2) Defect-sensitive performance: LLMs’ performance varies substantially across different defect types and severity levels, with semantically complex or low-salience defects being significantly more likely to be missed; (3) Underlying Failure Mechanisms: our in-depth error analysis dissects the distinct drivers of false positives and false negatives, revealing critical weaknesses such as cross-round temporal misalignment and inadequate long-range memory.
[NLP-5] Stochastic Estimation of Transduced Language Models
【速读】： 该论文旨在解决生成式语言模型在跨语言转换（如从源语言到目标语言）过程中，对目标前缀概率进行高效且无偏估计的问题。传统方法依赖于基于源前缀概率的计算捷径，并通过阈值剪枝的束搜索求和近似目标概率，但这种方法仅提供下界估计且误差未知。其核心解决方案是采用无放回重采样策略，对每个选中的源前缀按其包含概率的倒数进行加权校正，并递归应用此修正以构建无偏估计器。该方法不仅能够准确估计因阈值剪枝丢失的概率质量，还通过动态扩展并选择保留的源前缀，在保证运行终止概率为1的前提下显著降低计算开销。实验表明，在百科文本与DNA序列上的对比中，该方法相较于基于有放回重采样的顺序蒙特卡洛基线，在计算资源与方差之间取得更优平衡；在长目标串的前缀概率估计任务中，相较阈值剪枝束搜索实现数量级的加速，使原本不可行的任务变为可解。此外，在一项已发表的阅读时间分析中，用该无偏采样替代原阈值剪枝后，虽显著降低了语料库意外度（surprisal）估计值，但未改变原有结论的可靠性。
链接: https://arxiv.org/abs/2608.27428
作者: Vésteinn Snæbjarnarson,Samuel Kiegeland,Manuel de Prada Corral,Ryan Cotterell,Tim Vieira
机构: ETH Zürich(苏黎世联邦理工学院); University of Copenhagen(哥本哈根大学); CHI-FRO
类目: Computation and Language (cs.CL)
备注:
Abstract:Transduced language models (TLMs) compose a pretrained \emphsource language model with a functional finite-state transducer to induce a language model over \emphtarget strings. Computing the probability of a target prefix under a TLM amounts to summing the source-model probabilities of all source strings that the transducer maps to target strings beginning with that prefix. This set can be exponentially large or infinite. Prior work uses a computational shortcut based on source prefix probabilities, then approximates the resulting sum with threshold-pruned beam summing. This produces a lower bound with unknown error. Instead, we resample source prefixes without replacement and reweight each selected prefix by the inverse of its inclusion probability. We show that applying this correction recursively gives an unbiased estimator of the target prefix probability and lets us estimate the mass lost by threshold pruning. Our beam-summing algorithm extends the retained source prefixes and samples which prefixes to keep, reducing their number as more probability mass is added to the running estimate. This can save computation and guarantees that the run halts with probability one. We evaluate the method on encyclopedic text and DNA against sequential Monte Carlo baselines that resample with replacement. It achieves a better compute–variance tradeoff on text and lower error at the same maximum number of particles on DNA. On a DNA-to-amino-acid transduction, it reduces runtime by several orders of magnitude relative to threshold-pruned beam summing and makes estimating prefix probabilities for long target strings feasible. Replacing threshold pruning with unbiased sampling in a published reading-time analysis substantially lowers the estimated corpus surprisal but leaves the published conclusions unchanged.
[NLP-6] Boosting LLM Exploration via Weak-Model Guidance in RLVR
【速读】： 该论文旨在解决生成式强化学习（Reinforcement Learning with Verifiable Rewards, RLVR）在提升大语言模型（LLM）推理能力过程中导致的策略熵下降问题，该问题会引发推理覆盖范围缩小及高k值下的pass@k性能退化。其解决方案的关键在于引入外部前缀（outer prefixes）机制：通过让目标模型基于由更小、更弱的语言模型生成的部分推理轨迹进行回答，利用这些不熟悉且分布不同的前缀来打破模型的过度自信，从而激发对多样化推理路径的探索。实验表明，该方法能有效缓解熵崩溃现象，显著扩展推理覆盖范围，尤其在k值增大时性能优势更为突出，且无需额外的监督微调（SFT）、复杂奖励设计或提示工程，具有高效性与通用性。
链接: https://arxiv.org/abs/2608.27420
作者: Xingyu Shen,Huishuai Zhang,Peng Li,Yinchun Wang,Dongyan Zhao
机构: Wangxuan Institute of Computer Technology, Peking University(北京大学王选计算机技术研究所); National Engineering Research Center of New Electronic Publishing Technologies(国家电子出版新技术工程研究中心)
类目: Computation and Language (cs.CL)
备注: 13 pages, 4 figures
Abstract:Reinforcement Learning with Verifiable Rewards (RLVR) significantly improves LLM reasoning but often causes a drop in policy entropy, leading to narrowed reasoning coverage and degraded pass@ k for large k . While existing methods mitigate this entropy collapse through algorithmic regularizations, cross-model non-parametric perturbation is also neglected. In this work, we propose a simple yet effective approach to preserve the generative diversity of LLMs during RLVR. Instead of relying solely on internal exploration, we force the target model to generate answers based on partial reasoning trajectories generated by a smaller, weaker language models. These unfamiliar prefixes effectively disrupt over-confidence and encourage the exploration of distinct reasoning paths. We empirically study the potential of outer prefixes, revealing the mechanism of the impact of distributional discrepancy to the exploration dynamics in RLVR training. Experiments across multiple mathematical benchmarks show that our method consistently outperforms vanilla RLVR. Notably, the performance gain becomes increasingly pronounced as k scales up, demonstrating a substantial expansion of reasoning coverage. Furthermore, our approach efficiently mitigates entropy collapse without requiring additional SFT, intricate reward designs, or complex prompting.
[NLP-7] Consolidating RLVR Capabilities Across Domains: A Deep Dive into Fusion Paradigms
【速读】： 该论文旨在解决大语言模型在强化学习中引入可验证奖励（RLVR）时，为覆盖多种能力而需训练多个独立领域专家，进而导致系统复杂性和整合成本增加的问题。其核心挑战在于如何高效融合多个专家模型的能力，同时保持性能稳定与跨域泛化能力。解决方案的关键在于系统性比较三种基于复用资源不同类型的融合范式：**合并（Merge）**通过整合专家任务向量实现知识融合；**混合强化学习（Mix RL）**通过池化各专家数据集构建统一训练分布；**多教师在线策略蒸馏（MOPD）**则同时利用任务向量与数据集进行联合优化。研究发现，尽管三者在平均性能上差异较小（最大仅1.4分），但在特定基准上差距可达8.6分，且领域级表现差异与任务向量几何结构中的跨域关系高度相关。训练动态分析揭示了各自固有约束：Mix RL对领域混合比例敏感，MOPD受限于教师模型表现上限，而Merge将所有专家更新压缩至单一向量。三者均提升单样本准确率，但未显著改善解空间覆盖率或牺牲保留能力。据此提出实用指导原则：当已有专家且追求低成本融合时采用Merge；训练统一模型且无需预先专家时选择Mix RL，可调节数据分布以促进跨域迁移；若更重视保留领域特异性优势而非超越教师或降低端到端开销，则推荐使用MOPD。
链接: https://arxiv.org/abs/2608.27409
作者: Siye Wu,Kai Yang,Yuchen Cai,Xin Xu,Peng-Yuan Wang,Jiaxuan Wang,Jiashun Liu,Jiafei Lyu,Yangkun Chen,Saiyong Yang,Yanghua Xiao
机构: Tencent(腾讯); Fudan University (复旦大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Reinforcement learning with verifiable rewards (RLVR) improves specific capabilities of large language models, but covering multiple capabilities often involves training separate domain experts and subsequently consolidating them. We organize three fusion paradigms by the artefacts they reuse: Merge combines expert task vectors, Mix RL pools their datasets, and multi-teacher on-policy distillation (MOPD) uses both. Because they have largely been studied in isolation, how they compare and how to choose among them remain unclear. We compare all three using shared experts and data across model scales and a multi-domain benchmark suite. Although their average performance differs by at most 1.4 points, the gap reaches 8.6 points on a single benchmark, with domain-level variation tracking cross-domain relations visible in task-vector geometry. Training dynamics expose distinct constraints: Mix RL depends on domain mixture proportions, MOPD remains bounded by its teachers, and Merge compresses all expert updates into one. All three improve single-sample accuracy without measurable gains in solution coverage or losses in held-out capabilities. These results yield a practical guideline: use Merge when experts already exist and cheap fusion is paramount; Mix RL when training a unified model without experts, with domain proportions adjusted for cross-domain transfer; and MOPD when preserving domain-specific gains matters more than surpassing teachers or minimizing end-to-end cost.
[NLP-8] How Language Models Organize and Structure Moral Knowledge
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）如何组织道德知识这一核心问题，具体关注模型是否能够超越简单的道德内容检测，实现对道德基础（Moral Foundations Theory, MFT）之间的区分及其内在几何关系的建模。其解决方案的关键在于通过在开放权重的语言模型上训练六个独立的线性探测器（linear probes），分别对应MFT中的六种道德基础（关怀/伤害、公平/欺骗、自由/压迫、忠诚/背叛、权威/颠覆、神圣/亵渎），并分析这些探测器在表征空间中的方向关系。研究发现，各道德基础的方向并非完全融合或彼此孤立，而是占据近最大数量的独立维度，同时共享一个正向的公共成分——这一共享成分被识别为道德整合的标志，且在与非道德概念对比时表现出显著的道德特异性（平均成对余弦相似度0.26 vs. 0.013）。该几何结构在不同架构和规模下保持一致，并在预训练早期即已形成，早于探测器性能饱和。此外，模型在处理道德困境时，其表征方向由各基础成分部分组合而成（相比不匹配对基线提升2.7倍），但大部分方差仍编码了冲突特有的结构性信息，表明模型捕捉的是道德张力本身，而非预先解决的判断。
链接: https://arxiv.org/abs/2608.27402
作者: Orion Reblitz-Richardson
机构: Distiller Labs(迪斯蒂勒实验室)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 31 pages, 16 figures. Code and outputs at this https URL
Abstract:How do large language models (LLMs) organize moral knowledge? Models detect moral content broadly, but detection is a low bar. We ask whether they go further, distinguishing moral foundations from one another and organizing the relationships between them geometrically. We train six independent linear probes on open-weight language models, one per Moral Foundations Theory (MFT) category (care/harm, fair/cheat, lib/oppress, loy/betray, auth/subv, sanc/degrade), and examine how the resulting directions relate to each other in representation space. We find the directions neither collapse into a single moral detector nor isolate from one another. Rather, they span a near-maximal number of independent dimensions while sharing a positive common component. The shared component is the signature of integration, and it is moral-specific relative to a matched non-moral concept battery built identically (mean pairwise cosine 0.26 vs. 0.013). The geometry is consistent across architectures and scale and reaches its integration regime early in pre-training, well before probe accuracy saturates. The structure the model discovers shows no evidence of the individualizing/binding distinction predicted by Moral Foundations Theory (an underpowered test: only 20 candidate partitions exist) but rather reflects corpus statistics. Extending to moral dilemmas, each dilemma direction partially composes from its component foundations, at 2.7x a mismatched-pair baseline, while the majority of its variance encodes conflict-specific structure. The model represents moral tension itself, not a pre-resolved judgment. Comments: 31 pages, 16 figures. Code and outputs at this https URL Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) ACMclasses: I.2.6; I.2.7 Cite as: arXiv:2608.27402 [cs.CL] (or arXiv:2608.27402v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2608.27402 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-9] Making Clinical Language Models Auditable: Concept-Guided Fine-Tuning for Robust Prediction
【速读】： 该论文旨在解决临床语言模型在实际部署中性能下降的问题，其根源在于模型过度依赖病历文本中的特定模板、分隔符和固定语句等“文本特异性伪特征”（note-specific artifacts），而这些特征并不反映患者的真实临床状态。为此，论文提出了一种基于稀疏自编码器（Sparse Autoencoders, SAE）的可审计临床文本分类框架CAST（Concept-guided Artifact Suppression Tuning）。其解决方案的关键在于：利用SAE从Transformer中间层激活中提取稀疏且可人工解读的特征表示，通过大语言模型（LLM）辅助的语义解释与ICD-10诊断代码检索双重约束对这些特征进行概念标注；在微调阶段，通过残差减法抑制经验证的伪特征相关隐变量（latent），从而降低模型对非临床相关模式的依赖；同时，提供基于概念的后验归因分析，生成每个预测结果所依据的临床概念及其训练过程中被抑制的伪概念的完整可审计记录。该方法在MIMIC-IV出院小结死亡率预测任务上显著优于基线微调模型，并保持与强基准大模型相当的性能，同时实现了细粒度的模型决策可解释性与审计能力。
链接: https://arxiv.org/abs/2608.27397
作者: Jin Mu,Guanhua Chen
机构: University of Wisconsin–Madison (威斯康星大学麦迪逊分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Clinical language models can achieve strong in-hospital accuracy yet fail under deployment shifts because they exploit note-specific artifacts (e.g., templates, separators, boilerplate) that do not reflect patient state. We propose CAST (Concept-guided Artifact Suppression Tuning), an SAE-based framework for auditable clinical text classification. CAST uses Sparse Autoencoders to expose sparse, human-auditable features from intermediate Transformer activations, labels SAE latents with an LLM-assisted interpretation pipeline and ICD-10 retrieval constraints, suppresses verified artifact latents via residual subtraction during fine-tuning, and provides post-hoc per-concept attributions for auditing model decisions. On MIMIC-IV discharge-note mortality prediction, CAST improves over its corresponding fine-tuned encoder baselines and remains competitive with strong LLM baselines, while producing a feature-level audit trail of the clinical concepts that support each prediction and the artifact concepts suppressed during training.
[NLP-10] D2C-Routing: Dimension-to-Composition Evidence Routing for Mixed-Origin AI-Generated Text Detection EMNLP2026
【速读】： 该论文旨在解决生成式文本检测在混合来源写作（mixed-origin writing）场景下的有效性问题，即当文本的内容来源（content origin）与表达方式来源（expression origin）不一致时，传统二元判断框架（仅区分人工或机器生成）失效。其核心挑战在于准确识别内容与表达的双重来源，并据此推断出四种协作类型（如人类内容+人工智能表达等）。解决方案的关键是提出“维度到组合路由”（Dimension-to-Composition Routing, D2C-Routing）机制：该方法分别对内容侧和表达侧的证据进行独立建模，通过监督的维度头提取特征，再经由可学习的门控组合层融合信息，最终输出四分类结果。实验表明，在基于HART基准重构的MixD2C数据集上，所提系统达到0.8603的四类平均真正例率（Avg TPR@1%FPR），显著优于基线模型。消融实验证明了路由设计的有效性，而错误分析揭示了区分“AI内容+人类表达”与“完全AI生成”文本仍是当前最困难的边界。
链接: https://arxiv.org/abs/2608.27380
作者: Xin Chen,Fuwei Zhang,Yiqi Tong,Wei Guo,Yutian Xiao,Fuzhen Zhuang
机构: Institute of Artificial Intelligence, Beihang University(北京航空航天大学人工智能研究院), Beijing, China
类目: Computation and Language (cs.CL)
备注: 17 pages, 4 figures. To appear in EMNLP 2026
Abstract:AI-generated text detection is commonly framed as a binary document-level judgment about whether a text is human-written or machine-generated. This framing breaks down for mixed-origin writing, where content origin and expression origin may differ. We cast mixed-origin detection as dimension-to-composition source attribution, inferring content origin and expression origin before composing them into four collaboration types. We propose Dimension-to-Composition Routing (D2C-Routing), which routes content-side and expression-side evidence to supervised dimension heads before a learned gated composition layer predicts the final label. On MixD2C, a reconstructed split derived from the HART mixed-origin benchmark, our disclosed D2C-Routing-based detector system reaches 0.8603 four-way Avg TPR@1%FPR, 6.5 points above the same-split RACE-local rerun. Core ablations support the routing design, while error analysis shows that distinguishing AI-content/human-expression from fully AI-generated text remains the hardest boundary. Code is available at this https URL.
[NLP-11] Puro-2B: Poor Labs Qwen 2-1.5B Trained on RTX 5090 within 5090
【速读】： 该论文旨在解决大语言模型（Language Model）预训练成本过高导致学术界与开源社区难以参与的问题。当前即使在小规模场景下，复现如Llama-3.2-3B或SmolLM3-3B等模型的训练成本也分别超过150万美元和70万美元，严重限制了可及性。其核心解决方案在于提出一套开源、低成本且硬件友好的预训练方法——Puro-2B预训练方案，该方案通过多项关键技术实现显著成本压缩：选用消费级RTX 5090 GPU进行训练、采用FP8低精度计算、结合超球优化（hyperball optimization）、课程学习型模型平均（curriculum model averaging）以及精心设计的数据处理流程。在此基础上，研究者从零开始训练出一系列Puro-2B模型，最高达到1.4万亿词元的训练量，最优模型仅需不到6.9千美元的算力成本，即可逼近Qwen2.5-1.5B的性能表现。此外，研究进一步推导出“Puro成本缩放定律”（Puro Cost Scaling Law），表明约4.4千美元即可达到Qwen2-1.5B的性能水平。本工作还通过端到端实验揭示了预训练数据课程对下游微调后性能的影响机制，凸显完整预训练流水线开放的重要性。所有代码、数据与模型权重均已按Apache 2.0许可开源。
链接: https://arxiv.org/abs/2608.27370
作者: Kairong Luo,Jiarui Cui,Yaorui Yin,Shengqi Chen,Yiming Yang,Linxiang Gao,Yanmohan Wang,Mingzhe Zhang,Kaiyue Wen,Kaifeng Lyu,Wenguang Chen
机构: Tsinghua University (清华大学); Pengcheng Laboratory
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 62 pages, 20 figures, 24 tables
Abstract:Language model pretraining has become almost synonymous with prohibitive cost, placing it out of reach for much of the academic and open-source communities. Although strong open-source efforts already exist, including open-weight models and open-source training recipes, a cost-efficient, hardware-accessible, and open-source pretraining recipe has long been missing. Even at a small scale, training Llama-3.2-3B costs over \ 1.5M, and reproducing SmolLM3-3B needs over \ 700K. In this report, we present an open pretraining recipe designed to lower this barrier. Using this recipe, we train a collection of Puro-2B models from scratch on up to 1.4 trillion tokens with FP8 precision on consumer-grade RTX 5090 GPUs. The models in the collection differ in token budgets and selected recipe variants. Our best model is trained at a compute cost of less than \ 6.9K and approaches Qwen2.5-1.5B performance under our evaluation protocol. This cost efficiency is enabled by a combination of approaches, including hardware selection, low-precision training, hyperball optimization, curriculum model averaging, and the data recipe. Beyond the recipe itself, we provide two additional results. First, across the Puro-2B collection, we derive a Puro Cost Scaling Law that relates training cost to average model performance; the fitted law suggests that about \ 4.4K, less than \ 5,090, is sufficient to reach the performance of Qwen2-1.5B. Second, as an end-to-end case study, we examine how pretraining data curricula shape downstream performance after post-training. Such controlled studies are enabled by having access to the full pretraining pipeline rather than model weights alone. We release the full training recipe for Puro-2B, including data, code, and model weights under Apache 2.0 at this https URL.
[NLP-12] Your Voice Cloning System is Secretly a Voice Anonymizer
【速读】： 该论文旨在解决语音中说话人身份信息泄露的隐私问题，即在保留语音语言内容与听觉质量的前提下，有效抑制语音中的说话人特异性特征。其核心挑战在于如何在不牺牲语音可懂性与自然度的情况下实现强隐私保护。解决方案的关键在于复用已训练好的多语言语音克隆模型XTTSv2（27,000小时语音数据训练），通过引入伪说话人（pseudo-speaker）作为条件输入，利用该模型在语音克隆过程中对韵律结构的独立建模能力，实现无需重新训练即可完成说话人匿名化。进一步地，提出一种迭代优化策略，通过最大化说话人差异性与语音可懂性之间的调和平均（harmonic mean），在隐私保护与语音实用性之间取得平衡。实验在七种欧洲语言的CommonVoice与Multilingual LibriSpeech数据集上验证，系统实现了接近最优的隐私水平（等错误率EER ≈ 0.49），在语音可懂性方面表现竞争力，并显著优于现有专用匿名化基线方法的语音质量，且无需针对每种语言进行单独训练。
链接: https://arxiv.org/abs/2608.27360
作者: Romolo Muletta,Felix Matthias Saaro,Mark Cieliebak,Jan Deriu
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Speaker anonymization suppresses speaker-identifying attributes from speech while preserving linguistic content and quality. We propose repurposing XTTSv2, a multilingual voice cloning model trained on 27k hours of speech, for speaker anonymization without retraining. Our key insight is that XTTSv2’s voice cloning capabilities preserve prosodic structure independently of speaker identity, enabling voice conversion by conditioning on a pseudo-speaker. We introduce an iterative refinement strategy that balances privacy and utility by maximizing a harmonic mean of speaker dissimilarity and intelligibility. Evaluated on seven European languages across CommonVoice and Multilingual LibriSpeech, our system achieves near-optimal privacy (EER \approx 0.49), competitive intelligibility, and substantially better speech quality than dedicated anonymization baselines, while requiring no language-specific training. We release the code here: this https URL.
[NLP-13] RCMN: Understanding Misleadingness in Influential Public Discourse
【速读】： 该论文旨在解决公共话语中误导性信息的生成机制及其对读者理解影响的系统性认知不足问题，尤其关注信息框架、省略、语境化及传播方式如何隐性塑造受众解读。其解决方案的关键在于提出一种以读者为中心的误导性理解框架（Reader-Centric Misleadingness Understanding, RCMN），通过五个可操作维度——误导机制、潜在读者解读、证据支持的合理解读、情绪唤起程度以及传播意图——实现对误导性的多维建模。基于该框架，研究构建了一个基于证据的高质量公共话语数据集，实证发现误导性形式多样，远超于虚假陈述，常见机制包括无依据推断、夸大和信息缺失，且常伴随显著的情绪激发与扭曲的传播意图。进一步研究表明，尽管轻量级的声明-上下文表示在缺乏丰富上下文、证据或跨模态信息的情况下仍能有效恢复读者层面的解读，但识别误导机制本身仍需更深层次的语境与证据支撑。这一发现揭示了轻量级表示在可扩展误导性分析中的潜力，同时强调可靠解析误导机制仍依赖于更全面的信息基础。
链接: https://arxiv.org/abs/2608.27358
作者: Peiling Yi
机构: Kingston University London(金斯顿大学), United Kingdom(英国)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Influential public discourse shapes public beliefs and can also mislead, not only through what is stated, but also through how information is framed, omitted, contextualised, and communicated. Yet less research has focused on how such misleadingness arises and shapes the interpretations formed by readers. To address this gap, we introduce Reader-Centric Misleadingness Understanding (RCMN), a framework that operationalises misleadingness through five dimensions: misleading mechanism, likely reader interpretation, evidence-warranted interpretation, emotional arousal, and communicative intent. Based on this framework, we construct an evidence-grounded dataset of influential public discourse. Empirical findings show that misleadingness is diverse and extends well beyond fabrication, with unsupported inference, exaggeration, and omission among the prevalent mechanisms, and is frequently associated with heightened emotional arousal and distortive communicative intent. Moreover, we investigate whether lightweight claim-and-context representations retain sufficient cues for understanding reader-centric misleadingness without access to richer contextual, evidential, and multimodal information. Evaluation across five recent generative foundation models shows that reader-level interpretations can often be recovered from such limited representations, whereas identifying how misleadingness is produced remains considerably more challenging. These findings highlight the potential of lightweight representations for scalable misleadingness analysis, while reliable understanding of misleading mechanisms continues to require richer contextual and evidential grounding.
[NLP-14] INTENT-AS-A-TOOL Makes it Easy to Track Agent ic Misalignment
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）作为自主代理（autonomous agents）在实际部署中因目标冲突与外部压力导致的安全失效问题，即“代理错位”（agentic misalignment）。其核心挑战在于：当模型在推理过程中产生有害行为时，传统事后分析方法（如链式思维，Chain-of-Thought, CoT）提供的标签过于粗粒度，难以捕捉意图在生成过程中的动态演变。为此，论文提出INTENT-AS-A-TOOL方法，通过引入专门用于表达特定行为承诺的意图工具（intent-targeted tools），为模型提供一个明确的、专用的表达通道以表征其对某一目标行为的倾向性。该方法的关键在于：利用调用意图工具的概率，构建一种无需人工评判、细粒度的模型行为偏好信号，从而实现对推理过程中意图演化的连续追踪。实验表明，该方法不仅可补充传统CoT监控，还能将离散的后验标签扩展为密集的意图轨迹，并识别出关键干预节点，揭示了行为偏好在实时检测代理错位中的有效性。
链接: https://arxiv.org/abs/2608.27348
作者: Yutong Zhang,Jianshuo Dong,Peng Xu,Long Wang,Jie Zhang,Tianwei Zhang,Xiaoping Zhang,Han Qiu
机构: Tsinghua University(清华大学); MatrixOrigin; Nanyang Technological University(南洋理工大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:As large language models (LLMs) are deployed as autonomous agents, safety failures increasingly involve consequential actions. We study agentic misalignment, where agents take harmful actions under goal conflicts and pressures. Using chain-of-thought (CoT) monitoring, we find that harmful execution is often preceded by intent signals in reasoning. However, post-hoc CoT labels are too coarse to show how intent changes during generation. We introduce INTENT-AS-A-TOOL, an approach that adds intent-targeted tools to give the model a dedicated channel for expressing commitment to a target behavior. The probability of calling an intent tool provides a judge-free, fine-grained signal of the model’s tendency to pursue that behavior. Our results show that INTENT-AS-A-TOOL complements CoT monitoring, expands post-hoc CoT labels into dense trajectories, and identifies critical steps for online intervention. These findings suggest that action preferences are useful for tracking agentic misalignment during reasoning. Our code and data are accessible: this https URL.
[NLP-15] Pair-Level Essay-Scale Republication and Reuse from Frag mented Historical Text Reuse: A Workflow Study on Eighteenth-Century Books and Newspapers
【速读】： 该论文旨在解决从碎片化文本复用证据中恢复长篇作品级再版与再利用（essay-scale republication and reuse）的问题，其核心挑战在于对成对的零散证据进行整合以构建可信的传播关系，而不仅仅是片段检索。由于输入为非结构化的碎片化复用命中结果，且正例覆盖存在固有不完整性，研究将任务定义为在不完整标注条件下实现成对证据的可审计式整合。解决方案的关键在于提出一种分阶段的规则化工作流，并通过自动化规则适应优化，以实现对成对证据的有效聚合。实验表明，在仅使用成对特征聚合的情况下，已有0.948的F1值表现；最终工作流在多个数据集上实现了最优的精确率-召回率权衡。相比直接调用大语言模型（LLM）作为基线的方法，后者虽能极大扩展候选对数量（如在ECCO–ECCO中达14,886对），但缺乏精度控制，更适合作为高召回的候选扩展器而非分类器。在ECCO–报纸数据集上，经人工审核确认全部176个预测正例均为真实再版或再利用案例，同时揭示出来源多重性与期刊重复等深层出处结构。因此，在地面实况不完整的前提下，可审计的成对证据整合提供了一种高效生成紧凑候选空间、支持历史文献深度检视的实用路径。
链接: https://arxiv.org/abs/2608.27343
作者: Ke Shu,Kira Hinderks,Eetu Mäkelä,Mikko Tolonen
机构: University of Helsinki(赫尔辛基大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:This paper addresses the recovery of essay-scale republication and reuse from fragmented text-reuse evidence, a setting whose central challenge is pair-level evidence consolidation and not fragment retrieval alone. The study focuses on a candidate set centered on essays by eighteenth-century Scottish philosopher David Hume, spanning books from ECCO (Eighteenth Century Collections Online) and historical newspapers. Because the input consists of fragmented reuse hits instead of clean document pairs, and positive coverage is inherently incomplete, we formulate the task as pair-level evidence consolidation into plausible transmission relations and compare three methodological families: a staged rule-based workflow, baselines (a decision tree and two direct LLM settings), and automated rule adaptation. On labeled ECCO–ECCO slices, pair-level feature aggregation alone already reaches 0.948 F1 on the main labeled slice, while the final workflow gives the strongest overall precision-recall trade-off among the tested rule stages. On the full ECCO–ECCO candidate universe, direct LLM baselines flag up to 14,886 pairs as reprints compared to 771 for the final workflow, behaving in this direct-prompt setup as high-recall candidate expanders rather than precision-controlled deployment classifiers. On ECCO–Newspaper, manual audit confirms all 176 predicted positives as genuine cases of republication or reuse, while issue duplication and source-side multiplicity reveal additional provenance structure. Under incomplete ground truth, auditable pair-level evidence consolidation provides a practical way to produce compact candidate spaces for historical inspection.
[NLP-16] Beyond Parallel Blindness: Information Floors and Model Gaps in Block Drafting
【速读】： 该论文旨在解决生成式 AI（Generative AI）中块草稿（block drafters）在并行生成过程中因信息不完整而导致的效率与准确性问题。其核心挑战在于：块草稿在单次前向传播中提出多个候选词，但尚未生成的目标词无法提供完整的上下文条件，导致模型在拒绝低质量候选时同时受到两方面损失的影响——即块内路径信息缺失与可观测信息建模不充分，而现有方法难以区分这两者。为此，作者提出引入“信息下限”（information floor）的概念，定义为在特定条件顺序下预期的最小拒绝率，从而将总拒绝率分解为信息下限与模型差距（model gap）两部分。关键创新在于通过目标回溯（target rollout）在四个领域、四种开源权重模型及前沿API模型上对信息下限和模型差距进行估计，揭示出三个重要发现：首先，即使最优提案也受限于0.286的信息下限，使得最终槽位的接受率上限仅为71%；其次，仅需一个已实现的词即可消除86%-100%的信息下限，验证了局部性（locality）的重要性；第三，当前草稿器仍远高于其理论下限，最终槽位的模型差距贡献了DFlash拒绝率的43%-64%以及DSpark理想条件拒绝率的85%-92%。这些结果明确区分了短程条件依赖与提案质量之间的差异，为优化生成效率提供了可量化的理论基准。
链接: https://arxiv.org/abs/2608.27339
作者: Xinwei Qiang,Xiang Fang,Chang Chen,Yue Guan,Yufei Ding
机构: University of California San Diego(加州大学圣地亚哥分校)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Information Theory (cs.IT)
备注:
Abstract:Block drafters propose several tokens in one forward pass, before earlier target tokens are realised. Their rejection mixes two losses: missing within-block path information and imperfect modelling of observable information. Accepted length cannot distinguish them. We separate the two with an information floor, the minimum expected rejection at a specified conditioning order; rejection above this floor is the model gap. Estimating both from target rollouts across four domains, four open-weight targets, and a frontier API target yields three findings. First, the all-parallel floor reaches 0.286 at the final slot on Qwen3-4B, limiting even the best proposal to 71% per-slot acceptance. Second, one realised token removes 86 – 100% of this floor, a locality also recovered by an independent mutual-information analysis. Third, current drafters remain far above their floors: the final-slot model gap accounts for 43 – 64% of DFlash rejection and 85 – 92% of DSpark’s oracle-conditioned rejection. These findings separate the value of short-range conditioning from proposal quality.
[NLP-17] BTS-Agent Bench: A Deterministic Replayable Pipeline from Read-Only Telemetry Logs to Agent Benchmarks
【速读】： 该论文旨在解决工业场景中海量只读遥测数据（telemetry）缺乏标准化基准，难以有效转化为可执行的多轮智能体任务（multi-turn agent tasks）的问题。其核心挑战在于如何将原始遥测记录与元数据结构化为具备明确目标、证据支持和可验证性的任务实例。解决方案的关键在于提出一种名为BTS-AgentBench的遥测到任务剧集（telemetry-to-episode）构建方法：通过规范化遥测元数据与原始历史记录，构建只读工具库（read-only tool store），基于工具推导出的正确答案（gold answers）与证据生成静态任务，并将保留的任务升级为类型化、边界约束的面向操作员的剧集（typed, bounded operator-facing episodes）。该方法实现了任务的可复现性与质量可控性，通过代码契约预检（coded contract preflight）零发现验证，且构建排除控制器（construction-exclusion controller）在532条记录中未触发任何排除，表明构造过程稳健可靠。两次独立的原始数据到剧集构建均精确复现11个逻辑工具库导出项及训练/验证/测试集（356/87/89）的发布结果，验证了流程的一致性；进一步应用于XAI4HEAT数据集时，成功生成204个剧集，且在41条保留测试集上，仅操作员任务执行完成全部样本，凸显其对高阶推理模型的有效适配性。
链接: https://arxiv.org/abs/2608.27334
作者: Jeong-Yoon Kim
机构: 未知
类目: Computation and Language (cs.CL); Software Engineering (cs.SE)
备注:
Abstract:Industrial sites contain large volumes of read-only telemetry, but few benchmarks specify how to compile these records into executable multi-turn agent tasks. We present a telemetry-to-episode construction method instantiated as BTS-AgentBench. The pipeline normalizes BTS metadata and raw histories into a read-only tool store, compiles static tasks with tool-derived gold answers and evidence, and lifts retained tasks into typed, bounded operator-facing episodes. The 532-row release adds clarification, goal revision, timestamp policy, quality-gated reporting, and evidence attribution while preserving the source computation and split. Coded contract preflight reports zero findings, and the construction-exclusion controller completes 0/532 rows. Two independent raw-to-episode builds match all 11 logical tool-store exports and reproduce the released 356/87/89 train/dev/test artifact exactly. Applying the shared construction path to XAI4HEAT produces 204 episodes; on its 41-row held-out test split, the controller completes 0 rows and the retained GPT-5.5 execution completes all 41. Code, artifacts, and replay reports are available at this https URL.
[NLP-18] Difference-in-Differences on a Censored Rating Scale Can Manufacture an Effect: Evidence from a Pre-Registered LLM -Judge Audit
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）评判者在评估教育干预效果时可能存在的系统性偏差问题，特别是其评估结果中隐含的“双差分”设计所导致的识别失效问题。其核心挑战在于：现有审计方法通过双重差分（double difference）结构来检测评判者对特定属性（如学习者画像）的偏好差异，但该方法的终点指标在量表上无法被唯一确定。原因在于，双重差分中的每一项均受其自身观测值分布的截断（censoring）影响，导致观测到的统计量同时混杂了“偏好差异”与“衰减差异”——即当两个候选响应在量表边界附近具有不等距离时，即使无真实偏好差异，仅因共同的严重性偏移（severity shift）也会产生虚假的交互效应。论文通过一个预先注册的、冻结于首次调用前的“教学法评判者”审计案例，验证了这一机制：其主效应（学习者画像对支架支持偏好的影响）为零（+0.085点，95% BCa 置信区间 [-0.167, +0.353]，p = 0.684），而名义上显著的交互效应（+0.378，p = 0.002）可由量表下限与严重性偏移单独解释，且不包含任何差异偏好成分，其贡献度高达观测交互效应的79%–85%。研究进一步以闭式表达推导出该机制，并证明其影响可从审计本身生成的评分数据中直接测量。因此，解决方案的关键在于揭示并量化这种由量表截断与严重性偏移引发的伪交互效应，从而重新审视当前基于双差分的评测范式在生成式 AI 评判中的有效性与可识别性。
链接: https://arxiv.org/abs/2608.27309
作者: Shuyi Fan,Boyuan Deng,Mengyu Xu,Xinhong Xie,Chenyang Li,Hongyang Zhang
机构: Columbia University (哥伦比亚大学); Johns Hopkins University (约翰霍普金斯大学); The University of Chicago (芝加哥大学); The Pennsylvania State University (宾夕法尼亚州立大学); The Hong Kong Polytechnic University (香港理工大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 15 pages, 3 figures, 3 tables
Abstract:Audits of LLM judges certify a bias by contrasting matched conditions, and the strongest designs difference twice: a within-item contrast between two candidate responses, differenced again across a manipulated attribute, read off a bounded rating scale. We show that this endpoint is not identified on the scale that reports it. Each term of the double difference is censored by its own share, so the observed statistic confounds differential preference with differential attenuation: a severity shift common to both responses manufactures an interaction whenever the two censor it unequally, as unequal distances from the bounds make them, exactly where good stimuli place them. We exhibit the failure inside a pre-registered audit of a frozen pedagogy judge, sealed before the first of its 990 calls. The registered primary endpoint, the effect of a stated learner profile on the judge’s scaffolding preference, is null: +0.085 points (95% BCa [-0.167, +0.353] , p = 0.684 ). The audit’s one nominally significant interaction, +0.378 ( p = 0.002 ), is not identified as preference: a construction containing zero differential preference reproduces 79 to 85% of it from the observed severity shift and the scale floor alone. We derive the mechanism in closed form and show that its contribution is measurable from an audit’s own ratings.
[NLP-19] Naive Prompt Optimization: Rethinking the Need for Complex Prompt Search
【速读】： 该论文旨在解决如何在多样化任务中高效提升自主智能体（autonomous agents）性能，以加速代理型人工智能（agentic AI）的递归自进化（recursive self-improvement, RSI）。其核心挑战在于如何在保持高性能的同时降低计算成本，尤其是在提示优化（prompt optimization）过程中避免过度复杂的搜索机制。论文提出的解决方案之关键在于引入轻量级、单线性（single-lineage）的朴素提示优化（Naive Prompt Optimization, NPO）方法：通过一个教师模型（teacher model）利用回溯反馈（rollout feedback）迭代地修正提示，无需复杂的多路径搜索或高维空间探索。实验表明，NPO在较少回溯次数下即可达到甚至超越几何增强提示优化（GEPA）的性能，且随着教师模型推理能力的增强，其优势愈发明显，表明高质量的教师推理可部分替代优化器端的复杂搜索。此外，经NPO优化的提示在不同学生模型间具有良好的迁移性，尤其在同一模型家族内表现一致。研究结果表明，简单的线性提示优化策略能够与复杂度显著更高的搜索方法相媲美，为高效、低成本的智能体自我改进提供了新范式。
链接: https://arxiv.org/abs/2608.27266
作者: Yuan Chang,Xiaoqi Chen
机构: Purdue University(普渡大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 22 pages, 12 figures
Abstract:Efficiently improving autonomous agents across diverse tasks is central to accelerating recursive self-improvement (RSI) in agentic AI, with prompt optimization emerging as a promising approach capable of delivering performance gains comparable to those achieved by fine-tuning model weights, while reducing computational costs in both optimization and serving. However, recent developments increasingly favor unnecessarily complex prompt optimizers. We introduce Naive Prompt Optimization (NPO), a lightweight single-lineage method that iteratively revises prompts using a teacher model with rollout feedback. NPO achieves comparable or better performance than GEPA with fewer rollouts, and its advantage increases with stronger teacher models, suggesting that stronger teacher reasoning can partially substitute for optimizer-side search complexity. In interactive games, NPO remains broadly competitive with GEPA, while GRPO performs better on some tasks less amenable to prompt optimization. We also show that NPO-optimized prompts elicit similar performance improvements when applied verbatim to other student models, especially across models within the same family. Overall, our preliminary results show that simple, linear prompt optimization can rival substantially more sophisticated and complex search procedures.
[NLP-20] SCIT: Testing Causal Cache Carriers in Latent Chain-of-Thought Models EMNLP2026
【速读】： 该论文旨在解决生成式模型中潜在推理链（latent chain-of-thought）的可解释性问题，即中间推理过程被编码于连续状态而非显式文本，导致因果对象难以追踪。其核心挑战在于识别推理过程中真正承载关键计算的模型组件。解决方案的关键是提出一种名为后缀缓存交换测试（SCIT, Suffix Cache Interchange Test）的因果诊断协议，通过结合充分性测试与键值（K/V）组件拆分、隐藏状态控制、语义源控制、解码验证及匹配破坏等多重机制，构建精确的源-接收方反事实情境，从而定位执行反事实计算的具体变压器组件。实验表明，在CODI-GPT2及模拟CoT风格的GPT-2复现模型中，反事实算术推理主要通过值缓存后缀轨迹实现，而非隐藏状态、键、可重用答案槽或单标记触发器。对于主版本的CODI-GPT2检查点，晚阶段值缓存后缀机制具备充分且必要的证据支持；而模拟版本虽显示相同充分性模式，但缺乏足够的匹配破坏证据以确立必要性。此外，SCIT揭示了载体模式的转移现象：算术类小规模（如GPT-2/1B）单元维持潜在尾部值/KV传递，而高能力8B及修复后的非算术单元则转向提示前缀或完整缓存的K/V路径，边界单元未接收到任何机制调用。因此，该研究不仅提供了缓存层级的诊断工具，还明确了特定检查点下GPT-2的算术机制，并构建了基于能力的载体映射图谱，而非主张普适性的潜在尾部机制。
链接: https://arxiv.org/abs/2608.27265
作者: Yi Ding,Lijun Huang,Menglin Yang
机构: The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州)); Guangzhou, China
类目: Computation and Language (cs.CL)
备注: accept by emnlp2026
Abstract:Latent chain-of-thought models move intermediate reasoning from emitted text into continuous states, improving compactness but hiding the causal object. We introduce SCIT, the Suffix Cache Interchange Test, a causal protocol that constructs exact source-recipient counterfactuals, patches declared cache segments, and identifies which transformer object carries the counterfactual computation. SCIT combines sufficiency tests with K/V component splits, hidden-state controls, semantic source controls, decoded validation, and matched corruption. On CODI-GPT2 and a Sim-CoT-style GPT-2 reproduction, counterfactual arithmetic transfers primarily through value-cache suffix trajectories rather than hidden states, keys, reusable answer slots, or single-token triggers. Complete sufficiency-and-necessity evidence for the late-value-suffix mechanism holds for the main CODI-GPT2 checkpoint; the Sim-CoT-style checkpoint shows the same sufficiency and decoded-control pattern but insufficient matched-corruption evidence for a necessity call. Beyond these local arithmetic cells, SCIT reveals carrier-regime shifts: arithmetic-like GPT-2/1B cells preserve latent-tail value/KV transfer, whereas competent 8B and repaired non-arithmetic cells route through prompt-prefix or full-cache K/V; boundary cells receive no mechanism call. SCIT therefore contributes a cache-level diagnostic, a checkpoint-specific GPT-2 arithmetic mechanism, and a competence-gated carrier map rather than a universal latent-tail claim.
[NLP-21] What Makes Good Agent ic Data? An ACE Lens on Data Generation for LLM Agents
【速读】： 该论文旨在解决生成式智能体（LLM agents）在与外部环境交互时，如何高效生成既具一致性又富有信息量的交互数据这一核心问题。现有方法普遍存在领域割裂、评估标准不统一的问题，导致候选经验的构建与验证过程混杂，难以识别通用的数据生成机制。为此，本文提出一个两层框架：首先，将智能体交互数据形式化为统一的因子化对象（E, q, τ, v），包含环境定义、任务信号、交互实现及可选验证器，从而实现对生成范式的结构化组织；其次，引入基于准确率-复杂度-多样性（ACE）的约束分布设计视角，其中准确率确保数据在语义与逻辑上的一致性，复杂度根据学习者能力与执行配置动态调整学习分布，多样性则控制数据覆盖范围并抑制冗余。研究表明，当前研究正从依赖表面特征或数据规模的多样性转向以执行结果为基础的准确性、面向学习者能力的复杂度以及深层次的行为覆盖。总体而言，该框架的核心在于突破“数量驱动”的局限，强调在智能体与环境持续演化的背景下，持续分配有效、高信息量且非冗余的经验。
链接: https://arxiv.org/abs/2608.27260
作者: Xingshan Zeng,Zishan Xu,Boju Zhang,Yuzhou Wu,Lingzhi Wang,Jianghao Lin,Liangyou Li,Yasheng Wang,Lifeng Shang,Xin Jiang,Weinan Zhang,Yong Yu,Qun Liu,Weiwen Liu
机构: 华为(华为); 上海交通大学(Shanghai Jiao Tong University)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:LLM agents increasingly rely on generated interaction data to learn how to interact with external environments. Agentic data generation must maintain consistency among environments, tasks, interactions, and success signals while producing experience that is useful rather than merely abundant. Existing work spans many agent domains, but domain-centered organization and heterogeneous evaluation often obscure common generation mechanisms and conflate candidate construction with verification and selection. This work develops a two-level framework for the field. First, we represent agentic data as a common factorized object (E,q,\tau,v) , comprising an environment specification, task signal, interaction realization, and optional verifier. We organize generation paradigms by their primary anchor and dependency structure. Second, we formulate generation as constrained distribution design through the Accuracy-Complexity-divErsity (ACE) lens. Accuracy establishes the feasible support of grounded and internally consistent data. Within this support, Complexity places learning mass relative to the capability of a declared learner and execution configuration, while divErsity controls coverage and redundancy of data. Using this framework, we explore how prior work verifies generated experience, constructs and calibrates difficulty, and expands behavioral coverage. The literature reveals a shift toward execution-grounded accuracy, learner-relative complexity, and diversity beyond surface variation or dataset size. We further discuss broader directions and emerging trends in agentic data generation through the ACE lens, including their implications for scaling, data sources, training regimes and adaptive learning. Overall, the central challenge is not simply to generate more data, but to continually allocate valid, informative, and non-redundant experience as agents and environments evolve.
[NLP-22] BALMS: Benchmarking Agent ic LLM s for Longitudinal Mental Health Sensing EMNLP2026
【速读】： 该论文旨在解决现有生成式AI（Generative AI）在长期心理健康感知中缺乏对时序生理与行为信号进行推理和预测能力的问题。当前基于大语言模型（LLM）的健康代理主要局限于短期、检索式的查询（如一周内最高步数），无法有效处理跨时间维度的连续数据，以准确预测心理状态评分并生成基于证据的可解释性推理过程。为此，研究提出BALMS——首个系统性的基于LLM的智能体（agentic system）在长期心理健康感知领域的基准测试框架。其关键创新在于构建了涵盖三个真实世界纵向数据集、两类任务（封闭形式的心理健康评分预测与由LLM作为裁判自动评估的推理理由生成）、三种智能体范式及五种开源与闭源LLM底座的综合性评估体系。实验结果表明，零样本智能体通常难以超越简单均值基线，仅在具备更强模型能力或使用语义紧凑特征时表现较好；思维链（Chain-of-thought）提示虽提升部分推理型模型性能，但无法确保时间上下文的正确锚定或数值准确性。综合效率与时间尺度扩展分析进一步揭示：未来长期心理健康智能体亟需具备选择性历史检索、时间证据精准定位以及基于可解释行为特征进行推理的能力。
链接: https://arxiv.org/abs/2608.27219
作者: Yu Yvonne Wu,Arvind Pillai,Yuliang Chen,Yuwei Zhang,Sudarshan Regmi,Tess Z. Griffin,Michael V. Heinz,Lisa A. Marsch,Nicholas C. Jacobson,Andrew Campbell
机构: Dartmouth College; University of Cambridge
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 Main Conference
Abstract:Mental health assessment relies on episodic self-report scales, which convert subjective states such as stress into numerical scores but provide only sparse snapshots of wellbeing. Wearable devices offer longitudinal behavioral and physiological signals for continuous, low-burden monitoring. Recent LLM-driven personal-health agents enable natural language queries over wearable signals, but mainly handle short-term, retrieval-based lookups (e.g., highest step count over a week). They do not evaluate whether agents can reason over long-term signals to predict wellbeing scores paired with evidence-grounded rationales. To address this gap, we introduce BALMS, the first systematic benchmark of LLM-based agentic systems for longitudinal mental health sensing. BALMS spans 3 real-world longitudinal datasets, 2 task families (closed-form wellbeing-score prediction and rationale generation auto-graded by an LLM-as-Judge), 3 agentic paradigms evaluated across 5 open- and closed-source LLM backbones. We find that zero-shot agents rarely outperform a simple mean baseline, except with stronger backbones or compact, semantically meaningful features. Chain-of-thought prompting improves reasoning-oriented backbones, but does not guarantee temporal grounding or numerical correctness. Together with more analysis on efficiency and temporal scaling, BALMS highlights the need for longitudinal mental health agents that selectively retrieve history, ground temporal evidence, and reason over interpretable behavioral features.
[NLP-23] When Text Misleads: Inconsistent-Aware Reasoning for Audio-Grounded Dialogue
【速读】： 该论文旨在解决当前语音对话理解模型中存在的“文本依赖陷阱”问题，即模型过度依赖文本转录（transcript）而忽视声学信号（如语调、说话风格等）中的关键语用信息，导致其预测缺乏对语音的真实语境依赖。其核心解决方案是提出一种可扩展的框架，通过识别文本与声学信号之间的跨模态不一致（cross-modal disagreement），将存在矛盾的区域转化为冲突型问答（conflict QA）样本，并构建包含501个问题的可控基准数据集ContraTalk，涵盖互动行为、情感状态、对话行为、社会立场和对话意图五个语用维度。进一步提出“音频孪生”（Audio Twin）机制，将局部声学特征转化为可读文本表示，显式暴露声学证据以增强大语言模型的推理能力。实验表明，仅依赖文本的大型语言模型在一致场景下准确率超过90%，但在冲突场景中骤降至33%-48%；直接使用音频-语言模型（AudioLLM）虽有一定声学感知能力，仍约30%-40%情况下陷入文本偏见陷阱；而引入音频孪生框架显著提升了冲突场景下的准确性并减少了错误选择，但其表现仍受底层模型架构影响。研究揭示了文本捷径是语音对话理解中的重要失败模式，并证明显式聚合声学证据能提供更可控、可诊断的接口以改进语音驱动的推理能力。
链接: https://arxiv.org/abs/2608.27176
作者: Yen-Ju Lu,Yuzhe Wang,Yaohan Guan,Xiluo He,Jiarui Hai,Mingrui Liang,Kaavya Chaparala,Thomas Thebaud,Laureano Moro-Velazquez,Najim Dehak,Jesus Villalba
机构: Johns Hopkins University (约翰霍普金斯大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
备注: 24 pages, 4 figures
Abstract:Understanding spoken dialogue requires joint reasoning over lexical content and paralinguistic acoustic signals such as emotion and conversational intent. However, existing evaluations often allow shortcuts based on transcripts or single-modality solutions, obscuring whether models genuinely ground predictions in speech. We formalize this failure mode as cross-modal disagreement, where transcripts suggest plausible but incorrect surface interpretations while acoustic cues such as prosody or speaking style support different answers. We develop a scalable framework that identifies text-biased surface interpretations and converts disagreement regions into conflict QA examples. We also include consistent cases where transcript-based and speech-grounded interpretations agree, enabling evaluation beyond adversarial audio dependence. This results in ContraTalk, a controlled benchmark containing 501 questions across five discourse dimensions: interaction behavior, emotion state, dialogue act, social stance, and conversational intent. We further develop an agentic-style reasoning framework that converts speech into an Audio Twin, a text-readable representation of localized acoustic cues that exposes acoustic evidence to the reasoning model. Experiments show that strong text-only LLMs exceed 90% accuracy in consistent cases but drop to 33-48% in conflict cases. Direct AudioLLMs provide only partial grounding, still selecting the transcript-biased trap in roughly 30-40% of conflict cases. Our Audio Twin framework improves conflict-case accuracy while reducing trap selection, but its consistent-case behavior remains backbone-dependent. These results identify transcript-based shortcuts as an important failure mode in spoken dialogue understanding and show that explicit acoustic evidence aggregation provides a more controllable interface for diagnosing and improving speech-grounded reasoning.
[NLP-24] Calibrated Enough to Know Not Calibrated to Act: Fabricated Evidence Makes LLM Agents Commit to the Unknowable
【速读】： 该论文旨在解决大语言模型（LLM）在面对不可预测问题时，因虚假或虚构的可视化数据包装而产生过度自信决策的问题。其核心问题是：模型并非因信息内容本身作出判断，而是受“呈现形式权威性”误导，导致在无实际依据的情况下仍做出高置信度的方向性预测。解决方案的关键在于识别并修复模型中“行动/不行动”的决策门控机制（act/don’t-act gate），该机制在面对伪造数据时表现出显著失灵，但对可回答问题仍保持高度准确。研究发现，这一门控机制具有可训练性——通过对3B参数量模型在540个合成案例（如骰子、硬币、罐子、计时器等）上进行监督微调，可使模型在原始不可预测问题上的承诺率降至0%，且在三个未见过的领域中实现泛化。然而，该门控机制具有上下文脆弱性：当响应格式允许推理空间时，门控有效；而强制固定格式则剥夺了推理能力，使模型在本可正确回答的问题上仍盲目自信。因此，解决方案的核心是构建一个既可训练又依赖于适当交互设计的“智能决策门”，确保模型在缺乏真实证据时拒绝行动，从而实现安全可靠的生成式人工智能部署。
链接: https://arxiv.org/abs/2608.27167
作者: Pranav Aggarwal
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 28 pages, 6 figures. Code, data, pre-registration and all cached model outputs: this https URL . Also archived at Zenodo, DOI https://doi.org/10.5281/zenodo.22043517
Abstract:An LLM agent shown a professional-looking market panel commits to a directional call on a provably unpredictable question far more often than one asked the bare question: across 12 frontier models, commitment rises from 6.5% to 54.0% as evidence is escalated. It commits just as readily when every number on the panel is invented: fabricating the entire display, so nothing the model can see is true except the question itself, still lifts commitment from 24.5% to 36.8%, statistically indistinguishable from the 37.6% produced by genuine market data. What unlocks confident action is not information but the authority of its packaging. The failure is narrow and locatable. Incapacity is not the answer: on matched answerable questions attached to the same panels, the same models answer essentially always, at near-perfect accuracy. Nor is it belief - stated probabilities barely move across the gradient that swings action by 48 points, and score worse than a climatological baseline. Missing judgment isn’t it either: asked to classify a question’s knowability before acting, models call it irreducible 90% of the time and then commit on just 0.4% of those. The act/don’t-act gate is what fails, and the effect is concentrated in a few models rather than universal. Because the gate is separable, it can be trained. Supervised fine-tuning of a 3B model on 540 synthetic cases, predominantly dice, coins, jars and timers, drives commitment to 0.0% on the original cases and transfers to three unseen domains. It does not survive everything: the gate holds exactly when the response format leaves room to reason, and rigid formats that remove that room leave the model confident and wrong on questions it otherwise answers correctly. The gate is trainable and context-fragile, and deployment needs both halves of that sentence.
[NLP-25] Prediction of Prediction (PoP): Inter-Layer Activation Fusion for Single-Pass Hallucination Detection in Large Language Models
【速读】： 该论文旨在解决自回归大语言模型（LLM）在生成过程中产生事实性错误但表现出高置信度的问题，这一现象严重制约了其在高风险场景中的应用。现有方法要么在模型对错误陈述过度自信时失效，要么依赖多样本验证流程带来显著的内存与延迟开销。本文提出的核心解决方案是“预测的预测”（Prediction of Prediction, PoP），其关键在于利用单次前向传播中模型内部隐藏状态的层间动态变化，捕捉生成过程中的不确定性。PoP通过融合深度方向上的中间隐藏表示，无需额外解码调用即可实现对事实正确性的判断。在TruthfulQA基准测试上，该方法实现了75.5%的受试者工作特征曲线下面积（AUROC），且仅引入小于1.2%的运行时延迟，同时不增加任何生成轮次，具备高效、低开销的部署优势。
链接: https://arxiv.org/abs/2608.27165
作者: Himal Badu
机构: 未知
类目: Computation and Language (cs.CL)
备注: 7 pages, 3 figures, 9 tables. Single-author preprint on white-box, single-pass hallucination detection using inter-layer activation divergence, cross-layer fusion, temporal drift, and calibrated risk scoring. Evaluated on TruthfulQA, HaluEval 2.0, and FaithDial with Llama-3, Qwen2.5, and Mistral backbones
Abstract:Autoregressive large language models (LLMs) routinely generate factually incorrect outputs with high decoding confidence, limiting their deployment in high-stakes workflows. Existing output-stage uncertainty metrics can fail when models are overconfident on false assertions, while multi-sample verification pipelines introduce substantial memory and latency overhead. This work evaluates whether internal hidden-state transition dynamics during generation can signal factual errors without auxiliary decoding calls. We introduce Prediction of Prediction (PoP), a mechanism that captures layer-transition uncertainty by fusing intermediate hidden representations across depth during a single forward pass. Evaluated on the TruthfulQA benchmark using autoregressive transformer backbones, PoP achieves an area under the receiver operating characteristic curve (AUROC) of 75.5% for factual-correctness classification. The mechanism operates within the base forward pass, adding less than 1.2% runtime latency and requiring zero additional generation passes. The numerical results are reported from the author-verified experimental implementation and are bounded by the evaluation scope described below.
[NLP-26] STAR : Sentence Translation Alignment Rate for Document-to-Document Machine Translation EMNLP2026
【速读】： 该论文旨在解决生成式 AI 在单次遍历的文档到文档（Doc2Doc）机器翻译中常见的结构错位问题，即译文中出现句子遗漏或幻觉，违背源文与目标文之间的逐句对应关系。其核心解决方案是提出一种新的辅助评估指标——句级翻译对齐率（Sentence Translation Alignment Rate, STAR），用于显式量化句级结构保真度；在此基础上，设计了基于动态对齐掩码的偏好优化框架 StarPO，通过文档级假设的结构质量排序与聚焦于错位片段的动态优化策略，显著提升翻译的结构完整性与整体质量。实验表明，StarPO不仅在新闻和文学领域均有效提升翻译性能，更使小型模型超越 GPT-4o 等大型专有系统，同时保持更高的令牌效率。
链接: https://arxiv.org/abs/2608.27161
作者: Yichen Dong,Hao Wang,Junhui Li,Linlong Xu,Longyue Wang,Weihua Luo
机构: Soochow University (苏州大学); Alibaba Group (阿里巴巴集团)
类目: Computation and Language (cs.CL)
备注: EMNLP 2026
Abstract:Large Language Models (LLMs) have enabled a shift from sentence-level to document-to-document (Doc2Doc) machine translation, promising improved global coherence. However, document-to-document generation in a single pass frequently suffers from structural misalignment, manifesting as sentence omissions or hallucinations that violate the core requirement of source-target correspondence. To address this, we introduce Sentence Translation Alignment Rate (STAR), an auxiliary metric that explicitly quantifies sentence-level structural fidelity. Building on this, we propose STAR-masked Preference Optimization (StarPO), a framework that ranks document-level hypotheses by structural quality and utilizes a dynamic alignment mask to focus optimization on misaligned segments. Experimental results across news and literary domains demonstrate that StarPO significantly enhances translation quality and structural integrity. Notably, StarPO allows compact models to surpass the performance of massive proprietary systems like GPT-4o while maintaining superior token efficiency.
[NLP-27] Said Aloud Read Different: Cross-Modal Instability in Multimodal Models
【速读】： 该论文旨在解决多模态基础模型在语音优先的智能助手应用中，面对语义等价但模态（文本与语音）和语言（英语与阿拉伯语）差异时，判断一致性不足的问题。其核心挑战在于，现有评估方法难以捕捉由模态或语言转换引发的局部推理碎片化现象，即模型在特定三元组内无法一致地分辨支持性陈述与非支持性替代项。解决方案的关键是提出一种基于视觉锚定的对比三元组基准（visually grounded contrastive triplet benchmark），涵盖来自18个中东与北非（MENA）国家的10,150张具有文化背景的图像，每张图像配有一个支持性陈述及两个合理但不成立的替代项，并引入“对比不稳定性”（contrastive instability）作为度量指标，量化模型在三元组层面因模态或语言变化导致的判断不一致率。实验表明，语音输入显著放大了部分失败现象，而传统平均准确率指标未能充分反映此类问题，该研究通过公开基准数据集，推动对多模态模型鲁棒性与跨模态一致性更精细的评估。
链接: https://arxiv.org/abs/2608.27135
作者: Basel Mousi,Fahim Dalvi,Shammur Chowdhury,Firoj Alam,Nadir Durrani
机构: 未知
类目: Computation and Language (cs.CL)
备注: Interpeech 2026
Abstract:Multimodal foundation models are increasingly used in speech-first assistants that must interpret spoken queries and produce visually grounded decisions. Yet it remains unclear whether semantically equivalent queries yield consistent judgments across modality (text vs. speech) and language (English vs. Arabic). We introduce a speech-augmented visually grounded contrastive triplet benchmark spanning 10,150 culturally grounded images from 18 MENA countries, where each image is paired with one supported statement and two plausible but unsupported alternatives. We define contrastive instability as the conditional rate at which a model fails to resolve all statements within a triplet, isolating fragmented reasoning from complete failure. Evaluating recent multimodal models under text and speech in English and Arabic, we find that modality and language shifts introduce substantial triplet-level inconsistencies that are not fully captured by aggregate accuracy, with speech amplifying partial failures. We make the benchmark publicly available to the community.
[NLP-28] winKV: A Composable Repair Pass for KV Cache Eviction via Pairwise Key Redundancy
【速读】： 该论文旨在解决小模型在资源受限条件下，长上下文推理中因键值（Key-Value, KV）缓存内存开销过大而导致的性能瓶颈问题。现有KV缓存淘汰策略通常基于注意力分布或键向量与全局参考点的距离来评分，但研究通过控制变量的留一探针实验发现，注意力强度与令牌对答案的因果贡献无关（斯皮尔曼等级相关系数ρ = -0.004），这质疑了主流淘汰方法的基本假设。为此，论文提出TwinKV——一种无需训练、不依赖注意力机制的冗余信号检测方法，其核心是识别上下文中是否存在某令牌键的近似重复项。TwinKV并非替代现有策略，而是作为可组合的修复步骤：在给定淘汰策略保留集合的基础上，识别被错误淘汰且无副本留存的“孤儿”令牌，以及其信息已被其他保留令牌重复的“冗余捐赠者”，并在保持原始预算和评分规则的前提下进行交换。在LongBench、LooGLE、RULER及短上下文MMLU-Pro基准上的多组实验表明，TwinKV在多数配置下提升了性能，尤其在非性能饱和的策略中效果显著；值得注意的是，在Llama-3.2-1B上，尽管平均增益较小，但提升的样本比例更高，并在RULER任务中实现全面超越。此外，研究还发现，在少样本分类示例这类任务结构中，TwinKV对两个模型均无效，揭示了其适用边界。
链接: https://arxiv.org/abs/2608.27128
作者: Hong Chen,Yudong Zeng,Yongwei Huang,Zuhao Ouyang,Junyan Zhang,Xuming Hu
机构: The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州)); Bosum Institute of Management Science(博森管理科学研究院)
类目: Computation and Language (cs.CL)
备注:
Abstract:Long-context inference is bottlenecked by the memory footprint of the key-value (KV) cache, especially for small models under tight resource budgets. Existing KV cache eviction methods score tokens using the model’s attention distribution or, in attention-free variants, each key’s distance from a global reference point. Using a controlled leave-one-out probe, we find that attention magnitude is unrelated to a token’s causal contribution to the answer (Spearman \rho=-0.004 ), challenging the premise behind dominant eviction methods. We introduce TwinKV, a training-free, attention-free redundancy signal that detects whether a token’s key has a near-duplicate elsewhere in context. Rather than replacing existing policies, TwinKV acts as a composable repair pass: given a policy’s fixed retained set, it identifies evicted tokens with no surviving duplicate (\emphorphans) and retained tokens whose information is duplicated elsewhere (\emphredundant donors), then swaps them while preserving the original budget and scoring rule. We compose TwinKV with four recent eviction policies across LongBench, LooGLE, RULER, and a short-context MMLU-Pro no-harm control at compression ratios 0.3,0.5,0.7 . On Qwen3-4B, TwinKV improves a majority of configurations for two policies, is near-even for a third, and helps only a minority for a fourth adaptive baseline already near a performance ceiling; gains across the three non-ceiling policies are smallest at the loosest ratio. On RULER with Llama-3.2-1B, however, that fourth policy improves in every evaluated cell because its Alone score leaves substantial room to improve. More broadly, Llama-3.2-1B shows a smaller average LongBench gain but a higher fraction of improved cells on LongBench and LooGLE than Qwen3-4B, plus a clean RULER win. We also identify few-shot classification exemplars as a task structure where TwinKV does not help on either model.
[NLP-29] Cross-Lingual Alignment Without Joint Training: Do Monolingual Language Models Converge on Universal Representations?
【速读】： 该论文试图解决的问题是：在缺乏联合训练（如共享参数、混合语言批次或显式对齐目标）的情况下，仅在非平行语料上训练的单语语言模型是否仍能学习到可对齐的跨语言表示。其解决方案的关键在于揭示了跨语言对齐的潜在机制——即语言本身的结构与所承载的信息本身足以驱动模型在表征空间中自发形成可对齐的几何结构。研究发现，即使在严格单语训练条件下，不同语言模型间的隐藏状态仍表现出显著的可对齐性，且这种对齐随数据规模、模型规模或语言相似性的增加而增强；通过单一的Procrustes旋转即可实现跨模型的表征映射，并验证该旋转可有效转移功能性内容（如将英语模型的残差向量经旋转后嵌入德语模型，可正确预测对应国家首都）。这表明跨语言对齐并非必须依赖联合训练，而是源于语言内在规律，为未来基于单语组件的模型拼接、融合及模块化多语言系统提供了重要理论基础与实践路径。
链接: https://arxiv.org/abs/2608.27115
作者: Ej Zhou,Suchir Salhan,Catherine Arnett,Anna Korhonen
机构: University of Cambridge (剑桥大学); EleutherAI
类目: Computation and Language (cs.CL)
备注:
Abstract:Cross-lingual alignment in multilingual language models is typically attributed to joint training: shared parameters, mixed-language batches, or explicit alignment objectives. We ask whether monolingual models trained on non-parallel data learn alignable representations without joint training. By testing on strictly monolingual language models, such as the Goldfish model families and independently developed models from different research labs, we find three results. Correlation: these models develop alignable representational geometry across layers, with alignment strengthening as data scale, model scale, or linguistic proximity increases. Construction: a single Procrustes rotation fit on parallel sentences maps hidden states between models. Causation: the same rotation transfers functional content; patching a rotated English residual into a German model on a factual cloze flips the prediction to the donor’s capital in most cases. We confirm that cross-lingual alignment can emerge from the structure of language and the information it carries rather than from joint training, and this points to practical future directions including model stitching, merging, and modular multilingual systems built from monolingual components.
[NLP-30] DocTalkBN: A Novel Dataset of Expert Telemedicine Conversations in Bengali
【速读】： 该论文旨在解决低资源语言（如孟加拉语）中可靠医疗对话人工智能（Medical Conversational AI）缺乏真实专家-患者互动数据的问题。现有医疗对话数据集多源自医学论坛、书面健康内容或合成数据，难以反映真实临床场景中的口语化特征与上下文复杂性。为此，本文提出DocTalkBN，一个大规模多模态真实世界医疗远程会诊数据集，涵盖557.63小时的音视频与文本配对数据、1,515次多轮患者来电及10,274个医患问答对，覆盖26个医学专科，总计170万词元。其核心创新在于采集自国家级广播远程医疗项目的真实医生-患者对话，保留了口语表达、情境依赖性和自然交互特性，显著提升了数据的真实性与实用性。为推动基准研究，作者基于该数据集构建了三项下游任务：医疗分诊分类、建议安全性评估和医学命名实体识别，并对多种大语言模型与编码器基线模型进行了评测。实验表明，DocTalkBN在临床推理类任务中具有显著应用价值。研究团队已公开发布数据集与源代码，以促进面向低资源语言的可信医疗自然语言处理及更安全、文化适配的医疗系统发展。
链接: https://arxiv.org/abs/2608.27110
作者: Anik Saha,Fahmida Sultana Naznin,Sadatul Islam Sadi,Ananya Shahrin Promi,Wahid Al Azad Navid,Rifat Shahriyar
机构: Bangladesh University of Engineering and Technology (孟加拉国工程技术大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Reliable medical conversational AI requires authentic expert–patient interaction data, yet such datasets remain scarce, especially for low-resource languages such as Bengali. We present DocTalkBN, a large-scale multimodal dataset of real-world expert telemedicine conversations in Bengali, collected from nationally broadcast telemedicine programs featuring board-certified physicians. DocTalkBN contains 557.63 hours of paired audio and text, 1,515 multi-turn patient calls, 10,274 host–doctor question–answer exchanges, totaling 1.7M tokens, spanning 26 medical specialties. Unlike prior resources derived from medical forums, written health content, or synthetic data, our dataset preserves the spontaneity, contextual richness, and spoken characteristics of authentic medical interactions in a low-resource setting. To support benchmark-driven research, we further construct three downstream tasks from the corpus, medical triage classification, advice safety evaluation, and medical named entity recognition, and benchmark a diverse set of large language models and encoder-based baselines. Our results show that DocTalkBN is a practically useful resource, particularly for clinically grounded reasoning tasks. We release this resource to facilitate future research on reliable medical NLP and safer, more culturally grounded healthcare systems for low-resource languages. Our source codes and dataset are publicly available at this https URL.
[NLP-31] Unifying Detection and Adaptation in Task-Free Continual Learning EMNLP2026
【速读】： 该论文旨在解决大语言模型（LLM）在下游持续学习（Continual Learning, CL）中面临的灾难性遗忘问题，尤其针对现实场景中缺乏显式任务边界（task boundary）的无任务（task-free）情形。现有方法通常依赖于明确的任务划分来约束参数更新或引入特定任务的适配模块，但在真实应用中难以满足。为此，本文提出一种基于费雪信息矩阵（Fisher Information Matrix, FIM）引导的统一框架（Fisher-guided Unified, FiUni），其核心创新在于利用预训练模型的克罗内克积近似曲率（Kronecker-Factored Approximate Curvature, K-FAC）对FIM主子空间的正交性特征，通过少量下游任务样本即可有效反映不同任务间的相似性。在此基础上，FiUni构建由FIM导出的冻结子空间，指导低秩适应（Low-Rank Adaptation, LoRA），并动态匹配当前批次窗口与历史子空间的费雪主子空间，从而自适应地决定是复用已有知识、扩展相关子空间，还是创建新子空间，实现知识共享与任务隔离之间的动态平衡。该方案无需显式任务标签，具备良好的参数效率和对潜在批次级任务归属的推断能力，在性能上可媲美先进任务感知方法，同时显著减少可训练参数量。
链接: https://arxiv.org/abs/2608.27070
作者: Dezheng Han,Anbang Zhang,Zhihao Zhu,Shuaishuai Guo
机构: Shandong University (山东大学); The Hong Kong University of Science and Technology (香港科技大学); National Medicine-Engineering Interdisciplinary Industry-Education Integration Innovation Platform, Shandong University (山东省医工交叉产教融合创新平台，山东大学); Shandong Key Laboratory: Magnetic Field-free Medicine Functional Imaging, Qilu Hospital of Shandong University (山东省磁场无创医学功能成像重点实验室，山东大学齐鲁医院)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Accepted to Findings of EMNLP 2026
Abstract:To mitigate catastrophic forgetting in downstream continual learning (CL) for large language models (LLMs), existing methods typically constrain parameter updates or introduce task-specific adaptation modules. However, these methods often rely on explicit task boundaries during training, limiting their applicability to realistic task-free scenarios. In this paper, we propose a \textbfFisher-guided \textbfunified (\textbfFiUni) framework for batch-level task detection and parameter-efficient continual adaptation. FiUni is motivated by a key observation about the Fisher information matrix (FIM) of pre-trained models: the orthogonality among the principal subspaces of its Kronecker-Factored Approximate Curvature (K-FAC) approximation, estimated from a small number of downstream task samples, can reflect the similarity between different tasks. Based on this observation, FiUni constructs FIM-derived frozen subspaces to guide low-rank adaptation (LoRA), while matching the Fisher principal subspace of each incoming batch window with historical subspaces. This enables FiUni to adaptively determine whether to reuse existing knowledge, expand a related subspace, or create a new subspace, dynamically balancing knowledge sharing and task isolation. Experiments show that FiUni can effectively infer latent batch-level task affiliations and achieve competitive performance against advanced task-aware CL methods with fewer trainable parameters.
[NLP-32] Research Design Tracking and Assessment for the Social Sciences EMNLP2026
【速读】： 该论文旨在解决社会科学中因果研究设计评估的可靠性问题，传统上此类评估完全依赖人工专家分析，存在效率低、可扩展性差等局限。为此，论文提出自动化研究设计追踪与评估（ARDTrA）任务，核心目标是自动识别学术论文所采用的研究设计类型，并评估其应用质量。解决方案的关键在于构建一个由专家标注的涵盖六类反事实研究设计的语料库，并基于多轮检索增强生成（RAG）对话流程进行评估。研究发现，段落长度是影响模型性能的主要因素，解释了52%至66%的性能方差；此外，人类与机器在任务难度上的不一致性表明，存在两种独立的困难来源：一是专家间分歧较大的设计，二是机器难以处理的设计，这为后续模型优化提供了重要方向。
链接: https://arxiv.org/abs/2608.27049
作者: Marco Rovera,Sergiu Burlacu,Dominique Cappelletti,Alessio Tomelleri,Sonia Marzadro,Martina Bazzoli,Annalisa Tassi,Jessica Gagete-Miranda
机构: Fondazione Bruno Kessler(布鲁诺·凯斯勒基金会); Trento, Italy(特伦托, 意大利); IRVAPP(意大利研究与创新中心)
类目: Computation and Language (cs.CL); Computers and Society (cs.CY)
备注: Paper accepted at EMNLP 2026 - Main Conference
Abstract:Reliable assessment of causal research designs in the social sciences is critical for evidence-based policy-making, yet has so far relied entirely on manual expert analysis. We introduce Automated Research Design Tracking and Assessment (ARDTrA), a task that involves detecting the research design used in a paper and assessing the quality of its application. We create an expert-annotated dataset of papers covering six families of counterfactual research designs and evaluate the task using a multi-turn RAG-based conversational pipeline. Across four retrieval strategies, four LLMs and six embedding models, we find that passage length is the main driver of performance, explaining 52-66% of the variance. A per-research-design analysis also shows that human and machine difficulty do not align: the designs that prove hardest for the system are not those on which expert annotators disagree most, pointing to two independent sources of task difficulty.
[NLP-33] Cascaded Batch Prompting EMNLP2026
【速读】： 该论文旨在解决传统批量提示（batch prompting）在大语言模型推理中因同时处理多个实例而导致下游任务性能不可预测的问题。其核心解决方案是提出级联批量提示（cascaded batch prompting），一种两阶段方法，通过将复杂的推理过程与符号锚定（symbol grounding）解耦，提升了推理的稳定性和可预测性。实验结果表明，该方法在多项选择题问答和自然语言蕴含任务上均优于标准单次提示基线，同时实现了与批量大小成比例的加速，从而在帕累托前沿（Pareto frontier）上建立了新的性能-效率标杆。
链接: https://arxiv.org/abs/2608.27038
作者: Sho Hoshino,Peinan Zhang
机构: CyberAgent( CyberAgent)
类目: Computation and Language (cs.CL)
备注: EMNLP 2026 Findings
Abstract:Although batch prompting makes large language model inference more efficient by processing multiple instances simultaneously, it suffers from unpredictable downstream task performance. We propose cascaded batch prompting, a two-stage approach designed to resolve the unpredictability of conventional batch prompting by disentangling complex reasoning from symbol grounding. Experiments on multiple-choice question answering and natural language inference demonstrate that the proposed method outperforms the standard single prompting baseline while achieving a speedup proportional to batch size, establishing a new state of the art on the Pareto frontier.
[NLP-34] Reasoning about In-Context Samples for Machine-Translation
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在机器翻译（Machine Translation, MT）任务中因缺乏显式推理过程而导致的翻译可靠性不足问题。现有方法如标准k-shot提示或简单草稿生成，难以有效利用上下文示例中的语义结构信息，从而限制了翻译质量的提升。其解决方案的关键在于提出一种基于片段（fragment-based）的推理框架：模型首先从检索到的相似示例中提取平行的源-目标语言片段作为中间推理痕迹，再基于这些结构化片段生成最终翻译结果。为训练该模型，研究通过知识蒸馏从大型教师模型中提取“银质”片段与草稿数据。实验基于Qwen3模型族，在涵盖6种语言、每种语言最多5个领域的多场景设置下验证，结果表明该方法显著优于标准k-shot及基础草稿法，证明了显式片段化推理在提升翻译准确性和鲁棒性方面的有效性。
链接: https://arxiv.org/abs/2608.27036
作者: Maxime Bouthors,Josep Crego,François Yvon
机构: SYSTRAN by ChapsVision(SYSTRAN by ChapsVision); Sorbonne Université (索邦大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large Language Models (LLMs) can be trained to perform chain-of-thoughts reasoning in order to improve the reliability of their responses. In this work, we investigate how explicit reasoning can be leveraged for LLM-Based Machine Translation (MT) with in-context samples. We introduce a novel fragment-based reasoning framework in which the model first extracts parallel source-target fragments from retrieved similar exemplars, and uses these fragments as intermediate reasoning traces to produce the final translation. To train our model, we distill silver fragments and drafts from a large teacher model. Our experiments with the Qwen3 model family, over 6 languages, including up to 5 domains per language, demonstrate that fragment-based MT significantly outperforms alternative methods like standard k-shot or basic drafting.
[NLP-35] Representing and Parsing Korean Constituency Structure at Different Levels of Granularity
【速读】： 该论文旨在解决韩语成分句法分析中因句法树终端单元与表面词不直接对应而带来的表征难题。韩语的基本单位“词素块”（eojeol）具有复杂的形态结构，且现有句法资源在如何表示词素块内部形态及非显性成分方面存在差异。为应对这一挑战，本文基于宾夕法尼亚韩语语料库（Penn Korean Treebank），构建了三种不同的成分句法表示：词素+词性标注（Morpheme+XPOS）、词素块+词性标注（Eojeol+XPOS）和词素块+统一词性标注（Eojeol+UPOS）。其关键解决方案在于通过移除空成分、对齐宾夕法尼亚句法结构与显性词素块标记、尽可能保留原句法标签，并在终端层与非终端层上进行系统性调整，从而实现不同粒度的表征对比。在此统一建模与评估框架下，所有实验均使用金标准终端分段与预终端标签，仅评估在给定金标准形态句法标注条件下的句法分析性能。结果表明，以词素块为终端单元可缩短转换序列长度，但仅依赖词素块+统一词性标注的分析表现显著落后；而引入更丰富的词性标注信息（Eojeol+XPOS）可有效缩小差距，其中词素+词性标注（Morpheme+XPOS）在所有条件下表现最优，即使其预测结果被投影至词素块终端域后仍保持优势。该实证发现表明，细粒度的形态与词性标注信息对句法解析器具有重要价值。尽管此结果反映的是解析过程中可用信息的有效性，而非语言学上的最优终端域选择，但结合语言学与资源设计考量，词素块作为句法标注的稳定且可解释的表层域更具合理性，同时应将词素级与词性标注信息作为对齐的形态句法证据予以保留。
链接: https://arxiv.org/abs/2608.27035
作者: Jungyeul Park,KyungTae Lim,Zihao Huang,Eunkyul Leah Jo,Yige Chen,Chulwoo Park
机构: Korea Advanced Institute of Science and Technology, South Korea(韩国科学技术院, 韩国); The University of British Columbia, Canada(不列颠哥伦比亚大学, 加拿大); The Chinese University of Hong Kong, Hong Kong(香港中文大学, 香港); Anyang University, South Korea(安阳大学, 韩国)
类目: Computation and Language (cs.CL)
备注:
Abstract:Korean constituency parsing raises a representational challenge because the terminal units of a phrase-structure tree do not straightforwardly correspond to simple surface words. Korean eojeols are morphologically complex spacing units, and existing constituency resources differ in how they represent eojeol-internal morphology and non-overt elements. This paper compares three constituency parsing representations derived from the Penn Korean Treebank: Morpheme+XPOS, Eojeol+XPOS, and Eojeol+UPOS. We construct these representations by removing null elements, aligning Penn Korean phrase structure with overt eojeol tokens, preserving Penn Korean phrase labels where possible, and varying the terminal and preterminal layers. We then evaluate canonical non-binary transition-based constituency parsers in top-down, in-order, and bottom-up orders under a shared modeling and evaluation setup. All experiments use gold terminal segmentation and gold preterminal labels and therefore evaluate constituency parsing conditioned on gold morphosyntactic annotation. Eojeol terminals yield shorter transition sequences, but Eojeol+UPOS parsing substantially underperforms the morphologically richer conditions. Eojeol+XPOS narrows this gap, while Morpheme+XPOS gives the strongest results even after its predictions are projected to the eojeol terminal domain. Under these gold-annotation conditions, the results show that fine-grained morphological and XPOS representations provide valuable evidence for the evaluated parsers. This empirical finding concerns the information available for parsing and does not by itself determine the linguistically preferable terminal domain. Independently, linguistic and resource-design considerations motivate eojeol as a stable and interpretable surface domain for phrase-structure annotation, with morpheme-level and XPOS information retained as aligned morphosyntactic evidence.
[NLP-36] ITL: Interpretable Document Alignment with Structured Reference Frameworks
【速读】： 该论文旨在解决文档与结构化参考框架之间对齐度量的难题，核心挑战在于如何从文本中识别分散的语义证据，并以可量化、可解释且可追溯的方式报告其支持依据。传统检索与分类方法通常仅输出成对相似性分数或类别标签，难以提供与术语证据直接关联的概念级评分。为此，论文提出了一种领域无关、语言可迁移的方法——智能定位器（Intelligent Target Locator, ITL），用于估计目标文档中的文本单元与结构化参考文档（Structured Reference Document, SRD）中定义概念之间的亲和性。ITL 从 SRD 中构建基于独立词项、双词组、三词组及共现关系的概念特异性术语特征谱，通过融合概念归属性、词项类型特异性以及跨概念区分能力，为每个术语分配重要性权重。最终输出一个文本单元-概念亲和性矩阵，支持多粒度聚合分析。通过对 17 个可持续发展目标（SDGs）进行内部一致性评估，结果表明每个目标陈述与其对应概念的亲和性最高，而与其他概念的平均亲和性显著偏低，验证了 ITL 能够有效区分框架内各概念的语义特征。因此，ITL 的关键创新在于实现了对文档与结构化框架对齐度的定量评估，同时确保每项结果均可追溯至支撑性的术语证据。
链接: https://arxiv.org/abs/2608.27031
作者: Raúl Giráldez,Dayrelis Mena,Jesús S. Aguilar–Ruiz
机构: Pablo de Olavide University (帕勃罗·德·奥瓦拉德大学); School of Engineering (工程学院)
类目: Computation and Language (cs.CL)
备注: 21 pages, 2 figures, 7 tables, 2 Appendices
Abstract:Measuring alignment between documents and structured reference frameworks requires identifying conceptual evidence distributed throughout the text and reporting it through measures that are quantitative, interpretable, and traceable. Many commonly used retrieval and classification approaches return either pairwise similarity scores or one or more class labels, whereas fewer methods provide concept-level scores that are directly traceable to the terminological evidence supporting them. We present \emphIntelligent Target Locator (ITL), a domain-agnostic and language-portable methodology that estimates the affinity between the textual units of a target document and the concepts defined in a \emphStructured Reference Document ( SRD ). From the SRD , ITL induces concept-specific terminological profiles built from independent terms, bigrams, trigrams, and co-occurrences. Each term is assigned an importance weight that combines concept membership, term-type specificity and inter-concept discriminability. The output is a textual-unit–concept affinity matrix that can be aggregated at different levels of granularity. We conduct an internal consistency assessment using the 17 Sustainable Development Goals (SDGs), evaluating each official goal statement against the SRD induced from the same set of descriptors. Every statement reached its highest affinity with the corresponding concept, and the mean affinity across the remaining concepts stayed marginal relative to the mean reference affinity. This separation indicates that ITL distinguishes the conceptual profiles of the framework. ITL thus offers a general basis for quantifying document alignment with structured frameworks while keeping each result traceable to the terminological evidence that supports it.
[NLP-37] JudgeStealer: Extracting LLM Judging Capabilities across Evaluation Protocols
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）作为评判者（LLM judge）在黑盒访问场景下面临的模型提取攻击问题，尤其针对现有方法在多评估协议（包括点对评分、成对比较和列表排序）下查询预算受限时效率不足、缺乏针对性的问题。其解决方案的关键在于提出JUDGESTEALER——首个面向多种评估协议的高效模型提取框架，通过利用不同评估协议间的强一致性（cross-protocol agreement），仅依赖点对评分输出即可无额外查询地生成成对比较与列表排序的监督信号，从而显著提升查询效率；同时，该框架动态选择具有语义多样性、预测不确定性和潜在判断偏差的输入样本，结合分数平滑与多协议审查机制，在保留评分序关系的同时有效缓解代理模型训练中的灾难性遗忘，最终在多种主流LLM-as-a-judge及奖励模型上实现高达73.3%、87.0%和71.6%的点对评分、成对比较与列表排序任务准确率，并展现出对不同代理模型规模、适配策略及推理设置的鲁棒性，以及对典型防御机制的强抗性。
链接: https://arxiv.org/abs/2608.26982
作者: Chen Chen,Yaolin Chen,Xuehan Sun,Juan Lin,Xueluan Gong,Yuhang Zheng,Qian Wang,Kwok-Yan Lam
机构: Nanyang Technological University (南洋理工大学); Wuhan University (武汉大学)
类目: Computation and Language (cs.CL)
备注: 20 pages, 8 figures
Abstract:Large language model (LLM) judges are increasingly used across various evaluation scenarios, making their judgment capabilities valuable intellectual property. However, black-box access exposes these capabilities to model extraction attacks. Existing extraction methods do not specifically target LLM judges and provide limited support for multiple evaluation protocols under restricted query budgets. In this study, we propose JUDGESTEALER, the first query-efficient model extraction framework for replicating judging capabilities across pointwise scoring, pairwise comparison, and listwise ranking protocols. JUDGESTEALER exploits the strong cross-protocol agreement to acquire pointwise scores and transform them into pairwise and listwise supervisions without additional victim queries. To capture informative judge patterns and improve query efficiency, JUDGESTEALER dynamically selects pointwise inputs based on semantic diversity, predictive uncertainty, and potential judge biases. It further applies score smoothing and multi-protocol review to preserve the ordinal structure of scores and mitigate catastrophic forgetting during surrogate adaptation. Extensive experiments on state-of-the-art LLM-as-a-judge and reward models show that JUDGESTEALER consistently outperforms existing extraction baselines, achieving up to 73.3%, 87.0%, and 71.6% accuracy for pointwise, pairwise, and listwise evaluation, respectively. JUDGESTEALER also remains effective across different sur- rogate model scales, adaptation strategies, and reasoning settings. Moreover, JUDGESTEALER demonstrates robustness against representative extraction defenses.
[NLP-38] rrain signatures in Welsh settlement names
【速读】： 该论文旨在解决地名是否能够保留超出宏观地理结构的可测量环境信息这一科学问题，尤其关注威尔士定居点名称中地形相关词汇与实际地形高程之间的关联性。其解决方案的关键在于采用预先注册的、针对特定结果的建模方法，结合经过审计的24个词素构成的词汇框架和地理结构化验证，对3,757个威尔士定居点进行系统分析。研究通过对比包含高地形词素（如 bryn 或 mynydd）与低地形词素（如 cwm 或 pant）的定居点，发现前者所处位置平均比周围2公里范围高出24.4米（95%置信区间：10.8–38.1米；霍姆校正后 p = 0.00137），且该关联在不同邻域尺度（1、2、5公里）下均保持稳定，并在独立高程数据源中得到复现。进一步地，将地形命名极性纳入非词汇性空间与聚落基线模型后，显著降低了地理隔离区域的均方误差（分别在10、25、50公里空间分块下降低4.63%、6.22%、7.30%），表明命名系统中蕴含可量化的地形信息。尽管河流相关名称提供较弱但方向一致的证据，林地相关模型因数据限制无法估计，且残余空间结构、命名语言未解构及缺乏外部独立验证仍限制解释力，但研究证实部分威尔士定居点名称类别确实保留了与当前地形相关的可测量信息，而无需假设个体词源、因果命名机制、历史环境记忆或跨命名系统的可迁移性。
链接: https://arxiv.org/abs/2608.26978
作者: Oktay Karakuş,Can Eyupoglu
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Landscapes are named, but whether names retain measurable environmental information beyond broad geographic structure is rarely tested. We analysed 3,757 Welsh settlements using a frozen, source-audited 24-element lexical framework, preregistered outcome-specific models and geographically structured validation. The central comparison contrasted 101 settlements carrying high-terrain elements (\textitbryn or \textitmynydd) with 139 carrying low-terrain elements (\textitcwm or \textitpant). High-terrain names occupied locations 24.4 m higher relative to their 2-km surroundings (95% CI, 10.8–38.1 m; Holm-adjusted p = 0.00137). The association remained positive across prespecified 1-, 2- and 5-km neighbourhood definitions and was reproduced using an independently produced elevation source (24.1 m; 95% CI, 10.6–37.6 m). Adding terrain-name polarity to a non-lexical spatial and settlement baseline reduced geographically held-out mean squared error by 4.63%, 6.22% and 7.30% under 10-, 25- and 50-km spatial blocking, respectively, although improvement varied among held-out regions. River-related names provided weaker, directionally consistent evidence, while the preregistered woodland model was non-estimable. Residual spatial structure, unresolved name language and the absence of independent external replication limit interpretation. Selected Welsh settlement-name categories therefore retain measurable information about present-day terrain within Wales, without establishing individual etymology, causal naming, historical environmental memory or transferability to other naming systems.
[NLP-39] Squeezing More from Limited Data with Recursive Transformers
【速读】： 该论文旨在解决在数据量有限（10M–100M词）的预训练场景下，传统Transformer模型因参数规模与计算资源不匹配而导致的过拟合及泛化性能下降问题。其核心挑战在于：在固定数据预算但计算资源充足的情况下，单纯增加参数量会超过最优规模，导致模型过拟合，而标准Transformer架构由于嵌入层占据大量参数且每标记计算复杂度与表征能力强耦合，难以有效扩展。解决方案的关键在于解耦计算与表征容量——提出递归式Transformer（recursive Transformers），通过在深度方向复用共享模块以高效扩展计算量，并结合因子化嵌入（factorized embeddings）降低词汇映射参数占比。实验表明，该方法在10M和100M词级预训练任务中均优于标准Transformer，且在性能上可媲美BabyLM Challenge 2025的优胜模型，验证了其在低数据场景下的有效性与可扩展性。
链接: https://arxiv.org/abs/2608.26973
作者: Serdar Gülbahar,Lukas Edman,Alexander Fraser
机构: Technical University of Munich(慕尼黑工业大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Pre-training under limited data requires a different view of scaling than web-scale language modeling. With a fixed data budget but relatively abundant compute, increasing parameter count helps only up to an optimal scale; beyond that point, models overfit and generalization worsens. We study this behavior across 10M-100M word pre-training budgets, two corpora, and multiple downstream evaluations, and find that optimal size depends strongly on both the data budget and the downstream target. We argue that standard Transformers scale down poorly to this setting, because embeddings consume a large fraction of the parameter budget and per-token computation is tied to representational capacity. To address this coupling, we study recursive Transformers, reusing a shared block across depth to scale compute, together with factorized embeddings to reduce vocabulary-map parameters. We train three recursive models and find that they outperform standard Transformers at 10M and 100M words, while remaining competitive with BabyLM Challenge 2025 winners.
[NLP-40] EMPLAR Wales: A georeferenced environmental and toponymic dataset of Welsh settlements
【速读】： 该论文旨在解决地名数据在定量研究中因缺乏对地理实体、词汇标注与环境测量三者之间明确分离而导致的可复现性与准确性问题。其核心解决方案在于构建一个基于稳定标识符的地理参考型地名-环境数据集（TEMPLAR Wales），通过将3,757个威尔士聚落记录与确定性的词汇筛选机制及聚落层级的环境属性相链接，实现数据的可复现性与结构化管理。关键创新点在于：采用冻结的24个威尔士地名元素词库进行词汇检测，保留精确匹配与前缀匹配及其溯源信息；同时整合多尺度的河流/海岸距离、高程、地形背景、土地覆盖及邻域林木覆盖等环境属性，并通过独立高程产品验证地形数据的一致性。最终以四个关系型表格、字段级数据字典、源数据溯源记录和授权元数据形式发布，确保数据关系完整性与可追溯性，为地名学、语言地理学、历史与环境景观研究、地理信息系统（GIS）及空间数据分析提供可重复使用的基准资源，同时明确区分计算生成的词汇检测结果与真实语源或历史环境重建之间的界限。
链接: https://arxiv.org/abs/2608.26970
作者: Oktay Karakuş,Can Eyupoglu
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Place names provide persistent records of how landscapes have been described and organised, but their quantitative reuse requires explicit separation between mapped places, lexical annotations and environmental measurements. TEMPLAR Wales is a georeferenced environmental-toponymy dataset comprising 3,757 settlement records across Wales. The resource links a reproducible settlement frame to deterministic lexical screening and settlement-level environmental attributes through stable identifiers. It contains 1,350 lexical detections across 1,294 settlements, generated from a frozen registry of 24 Welsh place-name elements, while retaining exact- and prefix-token matches and their provenance separately. Environmental attributes describe river and coastal proximity, elevation and local terrain context at multiple spatial scales, land cover and neighbourhood woody cover, with parallel terrain measurements derived from independent elevation products. The dataset is distributed as four relational tables accompanied by a field-level data dictionary, source-provenance register and licensing metadata. Technical validation confirms relational integrity, deterministic lexical reconstruction, documented environmental coverage, strong agreement between independent terrain sources and reproducible reconstruction of the frozen release. TEMPLAR Wales provides a reusable foundation for research in toponymy, linguistic geography, historical and environmental landscape studies, GIS and spatial data analysis without treating computational lexical detections as verified etymologies or contemporary environmental measurements as historical landscape reconstructions.
[NLP-41] Scaling Model-Generated Distillation Data Can Make Latent Teacher Traits More Recoverable
【速读】： 该论文旨在解决生成式模型在知识蒸馏过程中，大规模生成数据是否仅通过增加样本数量来提升学生模型性能的问题。传统观点认为，扩大生成数据规模可提高覆盖范围、降低噪声并增强学生模型的泛化能力。然而，本文揭示了一个被忽视的关键现象：即使训练数据与目标特征无关（off-task），更大的数据规模仍能促使学生模型更清晰地捕捉到教师模型中隐含的细微特定信号（trait-specific signals），即便这些数据从未显式提及该特征。其解决方案的关键在于，通过控制实验设计模拟亚阈值学习（subliminal learning），利用教师模型生成受限的非任务相关数据（如仅包含数字的补全内容），并考察学生模型在独立评估域中的行为表现。研究发现，随着生成数据规模的增加，教师所诱导的目标特征在学生模型中的表现显著增强，且这种效应在不同模型家族、特征类型及跨模型迁移场景中均具一致性。此外，对低秩适配器（LoRA）参数更新的分析也验证了这一趋势。这表明，在使用大规模生成数据进行蒸馏时，必须结合对目标特征的敏感性设计数据筛选与评估策略，即使数据看似无关或无害。
链接: https://arxiv.org/abs/2608.26958
作者: Zhichen Dong,Zhixuan Liu,Yuyu Fan,Xiangtian Li,Shuyang Zhang,Chao Yang
机构: Shanghai Jiao Tong University (上海交通大学); Shanghai Artificial Intelligence Laboratory (上海人工智能实验室); Fudan University (复旦大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Scaling model-generated data is usually viewed as improving distillation: more examples should increase coverage, reduce noise, and produce stronger students. We show a second effect: larger datasets can make subtle teacher-specific signals easier to detect in the trained student, even when examples are off-task and never mention the trait. In a controlled setup inspired by subliminal learning, a teacher induced to express a target trait generates restricted off-task data, such as number-only completions. Students trained on different amounts of independent off-task data are evaluated in a separate domain, with matched no-trait controls isolating target-specific transfer. Our main finding is that larger independent datasets make the teacher’s induced trait stand out more clearly in the student’s later behavior. Other plausible traits may also strengthen with scale, but the target usually grows more. When the small-scale student already favors the target, scaling mainly amplifies that behavior; when it favors a related or salient alternative, more data can shift behavior toward the intended trait. Analyses of learned LoRA updates show a parallel trend. These effects appear across model families, trait types, multi-trait settings, and cross-model transfer. Our results suggest that scaling generated distillation data should be paired with trait-aware curation and evaluation, even when the data appears off-task or benign.
[NLP-42] From Atomic to Agent ic: Towards Interpretable Evaluation of LLM s Agent ic Mathematical Capabilities EMNLP2026
【速读】： 该论文旨在解决当前大型语言模型（Large Language Models, LLMs）在数学推理任务中仅依赖最终答案进行评估所导致的诊断能力不足问题。现有数学基准测试多采用结果导向的评价方式，无法有效识别模型在推理过程中的逻辑缺陷或行为模式，因而难以指导模型向具备自主性（agentic）能力的鲁棒智能体演进。为此，论文提出一种面向过程层面的评估框架，其核心在于将数学求解中的代理行为与可复用的数学原子能力（mathematical atomic capabilities）进行结构化对齐。该框架通过设计涵盖规划、执行与反馈的文本及多模态任务，并依托自动化流水线实现高质量求解轨迹生成与细粒度标注，从而实现对模型过程性推理能力的精准刻画。实验表明，即使在端到端准确率相近的情况下，不同模型在代理能力谱系上仍表现出显著差异，验证了过程级评估对于揭示模型真实潜力、推动下一代数学智能体发展的关键作用。
链接: https://arxiv.org/abs/2608.26950
作者: Jiayi Kuang,Yinghui Li,Yunze Song,Keyu Chen,Zhifeng Shen,Yangning Li,Yidong Wang,Di Yin,Ruizhi Qiao,Xing Sun,Kai Jin,Ying Shen,Liang Lin,Philip S. Yu
机构: Tencent Youtu Lab; Sun Yat-sen University; University of Illinois Chicago; Pengcheng Laboratory
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: EMNLP 2026
Abstract:Large Language Models (LLMs) are evolving from performing end-to-end mathematical reasoning to integrating agentic intelligence. However, most existing math benchmarks evaluate only final answers. This outcome-oriented evaluation provides limited diagnostic value for identifying process-level failures or rigorous logic, failing to guide the transformation of LLMs into robust agents. To bridge this gap, we present a process-level benchmark designed to evaluate the inherent agentic mathematical reasoning abilities of LLMs. Our framework aligns problem-solving agentic behaviors with a structured taxonomy of reusable mathematical atomic capabilities. We design a comprehensive suite of planning, action, and feedback tasks across both textual and multimodal contexts, supported by an automated pipeline that synthesizes high-quality trajectories and produces fine-grained annotations via controlled LLM rewriting. Experiments reveal that models with similar end-to-end accuracy can exhibit markedly different agentic capability profiles. This demonstrates that process-level evaluation is crucial for interpreting the true potential of LLMs and guiding the development of next-generation mathematical agents.
[NLP-43] KinyaEmbed: Contrastive Sentence Embeddings for Kinyarwanda via Multi-Stage Curriculum Training
【速读】： 该论文旨在解决卢旺达语（Kinyarwanda）这一形态丰富的班图语系语言在现有通用句向量模型中表现不佳的问题，其根本原因是这些模型在预训练语料库中对卢旺达语严重欠表示。为应对这一挑战，作者提出KinyaEmbed——首个专为卢旺达语设计的句向量模型。其解决方案的关键在于采用四阶段课程学习（curriculum learning）策略，结合多种高质量标注数据与损失函数优化：第一阶段利用卢旺达官方公报中的约1.8万个同义句对及多温度尺度增强；第二阶段基于NLLB翻译的MNLI三元组进行蕴含结构微调；第三阶段通过英-卢旺达语OPUS-100翻译对实现跨语言表征对齐；第四阶段则使用从KinyaCOMET筛选出的2,936个高质量句对（置信度阈值0.8）进行精细化优化。最终通过七检查点集成模型，在SemRel2024-rw基准上达到Spearman ρ=0.7298，显著优于mE5-large和OpenAI text-embedding-3-large，并在新构建的无污染卢旺达语语义相似性任务基准Wiki-RW-STS上验证了其有效性，同时在文档聚类任务中取得最优轮廓系数（0.2146）。
链接: https://arxiv.org/abs/2608.26941
作者: Ireddi Rakshitha,Devavarapu Yashwanth,Ntakirutimana Pierre
机构: Barclays; Carnegie Mellon University (卡内基梅隆大学)
类目: Computation and Language (cs.CL)
备注: 14 pages, 3 figures, 5 tables
Abstract:We present KinyaEmbed, the first dedicated sentence embedding model for Kinyarwanda, a morphologically rich Bantu language spoken by over 12 million people in Rwanda. Existing multilingual embedding models such as LaBSE, mE5-large, and OpenAI text-embedding-3-large perform poorly on Kinyarwanda due to severe under-representation in their pre-training corpora. KinyaEmbed is built on KinyaBERT-large and trained via a four-stage curriculum using MultipleNegativesRankingLoss (MNRL): Stage 1 leverages ~18,000 paraphrase pairs from the Official Gazette of Rwanda with three temperature scales; Stage 2 fine-tunes on 715 NLLB-translated MNLI triplets for entailment structure; Stage 3 aligns representations using English-Kinyarwanda OPUS-100 translation pairs; Stage 4 refines with 2,936 high-quality pairs filtered from KinyaCOMET at quality threshold 0.8. We evaluate on SemRel2024-rw and introduce Wiki-RW-STS, a new contamination-free Kinyarwanda STS benchmark of 300 pairs derived from Kinyarwanda Wikipedia. A seven-checkpoint ensemble (all5+23A*2, with the final stage double-weighted) achieves Spearman \rho=0.7298 on SemRel2024-rw, surpassing mE5-large by 20.9% and OpenAI text-embedding-3-large by 41.0%. KinyaEmbed also achieves the best document clustering silhouette score (0.2146) across all evaluated models. All checkpoints, the KinyaCOMET filtered pairs, and the Wiki-RW-STS benchmark are publicly available.
[NLP-44] Mapping Written Words to Spoken Words in a Different Language Using Only Visual Grounding
【速读】： 该论文旨在解决在低资源语言环境下，如何利用视觉引导的口语数据构建跨语言词到语音映射的问题。具体而言，在缺乏标注转录文本的情况下，如何从图像-语音配对数据中自动识别特定英文关键词在印地语口语中的对应发音片段。其解决方案的关键在于提出一种基于自监督语音表征的对齐方法：首先通过现成的图像字幕系统自动获取图像的英文标签，随后利用自监督语音特征对齐同一关键词对应的印地语语音片段，并聚合对齐证据以识别重复出现的目标词语音段。实验表明，该方法在关键词检测与定位任务上优于先前基于注意力机制的端到端多模态神经模型，且引入负例样本可进一步提升性能。该研究证明了无需依赖转录文本或显式模型训练，即可直接从视觉语境中学习跨语言词-语音映射的可能性。
链接: https://arxiv.org/abs/2608.26925
作者: Gabriel Pirlogeanu,Dan Oneata,Horia Cucu,Herman Kamper
机构: Politehnica Bucharest (布加勒斯特理工大学); Stellenbosch University (斯泰伦博斯大学); Trinity College Dublin (都柏林三一学院)
类目: Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
备注: 9 pages, 5 figures, 5 tables, preprint, submitted to IEEE Transactions on Audio, Speech and Language Processing
Abstract:In many low-resource settings, even just eliciting speech for data collection is difficult. One promising approach has been to ask speakers to describe images. But how do we build models from such visually grounded speech data? Given a dataset of images with Hindi spoken captions, we consider how we can map a written English keyword to spoken realisations of that word in Hindi. Previous work trained end-to-end multimodal neural models. Instead, we explore a simpler alignment-based approach built on self-supervised speech representations. Written English tags are automatically obtained from images using off-the-shelf image captioning systems. Hindi utterances associated with the same keyword are then aligned (using self-supervised features), and alignment evidence is aggregated to identify recurring speech segments corresponding to the target word. Experiments evaluating keyword spotting and localization show that our alignment-based approach outperforms a previous attention-based neural model. We also show the benefit of incorporating negative examples during alignment. Our work demonstrates that cross-lingual word-to-speech mappings can be learned directly from visual grounding without transcriptions or explicit model training.
[NLP-45] abuLM: Morphology-Aware Tabular Pre-training for Low-Resource Languages
【速读】： 该论文旨在解决卢旺达语（Kinyarwanda）这一语法结构复杂的班图语系语言在表格数据表示学习方面缺乏专用资源的问题。由于卢旺达语拥有丰富的形态变化特征，且现有自然语言处理模型普遍缺乏对低资源、高形态语言的适配能力，导致其在表格问答（Table QA）等任务中表现受限。为此，论文提出TabuLM，首个基于卢旺达语表格数据预训练的语言模型，其关键创新在于：在已有的两层形态变换器架构KinyaBERT-large基础上，引入了行、列及单元格类型嵌入，并设计了表结构注意力偏置机制，以增强同一行与同一列间的注意力聚焦能力；同时提出了两种新型预训练目标——掩码单元恢复（Masked Cell Recovery, MCR）和列类型预测（Column Type Prediction, CTP），分别促进从行列上下文重建被掩码单元以及根据单元值预测列语义类型。模型在172个卢旺达政府公开数据门户中的表格（约3.5万个单元格）上进行预训练，并构建了首个原生卢旺达语表格问答基准集TabQA-kin（包含526个问答对，覆盖31张表格与四种问题类型）。实验表明，TabuLM在TabQA-kin上达到62.0%的精确匹配率，显著优于KinyaBERT-large（+5.7 EM点）及所有多语言基线模型（mBERT 49.3%，XLM-R 50.0%），提升达11.7–12.7个百分点。分析显示，表结构嵌入对比较与查找类问题贡献最大，而形态学感知能力则提供互补增益。
链接: https://arxiv.org/abs/2608.26923
作者: Ireddi Rakshitha,Devavarapu Yashwanth,Ntakirutimana Pierre
机构: Barclays; Carnegie Mellon University (卡内基梅隆大学)
类目: Computation and Language (cs.CL)
备注: 18 pages, 4 figures, 7 tables
Abstract:We present TabuLM, the first language model pre-trained on Kinyarwanda tabular data. Kinyarwanda is a morphologically rich Bantu language spoken by over 12 million people in Rwanda, yet lacks any dedicated tabular representation learning resource. TabuLM extends KinyaBERT-large, a two-tier morphological transformer, with additive row, column, and cell-type embeddings and a learned table-structure attention bias that sharpens same-row and same-column attention. Pre-training uses two new objectives: Masked Cell Recovery (MCR), which masks entire cells and forces reconstruction from row and column context, and Column Type Prediction (CTP), which predicts column semantic types from observed cell values. We pre-train on 172 Rwandan government tables (~35,000 cells) from NISR, RAB, REB, and MoH open-data portals, and introduce TabQA-kin, the first native Kinyarwanda table question-answering benchmark comprising 526 QA pairs across 31 tables and four question types. TabuLM achieves 62.0% exact match on TabQA-kin, outperforming KinyaBERT-large by 5.7 EM points and all multilingual baselines (mBERT 49.3%, XLM-R 50.0%) by 11.7-12.7 points. Analysis shows that structural table embeddings are most decisive for comparison and lookup questions, while morphological awareness provides complementary gains. Our code, data, and pre-trained checkpoint are publicly available.
[NLP-46] AraMS-28k: The Largest Publicly Released Line-Level Dataset of Historical Arabic Manuscripts with Margin and Insertion-Anchor Annotations
【速读】： 该论文旨在解决历史阿拉伯语手稿中缺乏大规模、高质量、细粒度标注的公开数据集这一关键问题，尤其针对手稿文本的非线性阅读顺序恢复、多书写体（Naskh、Ruq’ah、Maghrebi）兼容性以及音位标记缺失等挑战。其解决方案的关键在于构建AraMS-28k——首个公开发布的、基于真实历史阿拉伯手稿的线级（line-level）标注数据集，包含28,600条标注文本行（含主文与页边注），并首次引入插入锚点（insertion anchor）标注以恢复手稿真实的非线性阅读顺序。通过RefLAM（参考文本引导的标注流程），该数据集结合多模态大语言模型（multimodal-LLM）光学字符识别（OCR）与独立来源的纯净转录本进行对齐，并经由人工审核实现自动验证与专家监督的双重质量控制。此外，每条文本行同时提供带音位标记（diacritised）与音位归一化（diacritic-normalised）版本，以应对手稿书写中普遍缺乏音位符号的问题。数据集附带页面图像、细粒度标注及固定训练/验证/测试划分，采用CC BY-NC-SA 4.0许可，支持可复现的手稿识别、版面分析与阅读顺序重建研究。
链接: https://arxiv.org/abs/2608.26921
作者: Mohamed Guechaoui,Mohamed Diaa Zellagui,Souleyman Chaib,Sahraoui Dhelim
机构: Higher School of Computer Science (ESI-SBA), Sidi Bel Abbes, Algeria
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: Data and code available at this https URL and this https URL. Dataset: this https URL Code: this https URL
Abstract:We introduce AraMS-28k, the largest publicly released line-level dataset of genuine historical Arabic manuscripts, comprising 14 books, 3,043 pages, and 28,600 annotated text lines (27,971 main-text, 629 margin). Thirteen books are hand-copied manuscripts spanning three script traditions – Naskh, Ruq’ah, and Maghrebi – and one is a lithographed printed edition included to broaden format diversity. Each line is labelled as main-text or margin, and margin lines that have an unambiguous attachment point in the main text are further annotated with an insertion anchor, recovering the manuscript’s true non-linear reading order at line-level granularity – to our knowledge the first such annotation released for a historical Arabic manuscript corpus. Because reference transcriptions are fully vocalised while manuscript hands are typically undiacritised, we release both the raw diacritised transcription and a diacritic-normalised counterpart for every line. The dataset was constructed with RefLAM, a reference-grounded annotation pipeline that aligns multimodal-LLM OCR against independently sourced clean transcriptions and routes every line through human review, combining automatic verification with expert oversight. We describe the construction and quality-control process, present the annotation schema, report dataset statistics at both the corpus and per-book level, and provide baseline HTR results using Kraken and HATFormer, including a cross-script generalisation gradient from in-distribution pages to fully unseen books. AraMS-28k is released with page images, line-level annotations, and fixed train/val/test splits under CC BY-NC-SA 4.0 to support reproducible research on Arabic manuscript recognition, layout analysis, and reading-order recovery.
[NLP-47] Planting a Latent Variable in Natural-Looking Text: a More Realistic Test of Belief States in LLM s and Their Link to Concept Geometry
【速读】： 该论文旨在解决大语言模型（LLM）是否能够真实追踪潜在语义变量的“信念状态”（belief states），即在自然语言文本中对隐藏变量进行概率分布建模的问题。尽管已有研究在合成数据和个别案例中验证了这一假设，但缺乏在真实自然语言语料中的系统性证据，且未将信念状态的动态演化与模型特征空间的几何结构（如稀疏自编码器发现的概念解释性特征）建立实证关联。本文的关键解决方案是：在自然语言文本中植入一个可调控的潜在变量，并通过一个教师模型在生成过程中以隐式方式沿8个无关的稀疏自编码器方向施加控制，其激活遵循环形马尔可夫链。实验结果表明，一个小型变压器模型在训练后不仅能准确跟踪该潜变量的贝叶斯后验信念，还自发地将8个状态在特征空间中排列成与马尔可夫链顺序一致的环形结构，这为潜变量的统计动力学可决定概念几何形态提供了直接证据，揭示了模型内部表征的内在组织原则。
链接: https://arxiv.org/abs/2608.26887
作者: Alexandru-Iulius Jerpelea
机构: Columbia University(哥伦比亚大学)
类目: Computation and Language (cs.CL)
备注: 9 pages, 13 figures
Abstract:LLMs are thought to track “belief states,” i.e., running probability distributions over the latent variables that govern language (Shai et al., 2024; Sarfati et al., 2026), but so far this has only been comprehensively demonstrated on toy synthetic data and in a few isolated case studies. It has also never been empirically connected to the geometry of LLM features (the concepts interpretability finds in model activations). In this work, we plant a controllable latent variable inside natural-looking text. An LLM teacher writes ordinary text while we “subliminally” steer it along one of K = 8 unrelated sparse autoencoder directions at each token, with the active directions following a ring-shaped Markov chain. A small transformer model trained on this corpus does indeed track the Bayesian posterior belief about our planted latent variable. Moreover, it also arranges the 8 states themselves on a ring, in the exact order of the Markov chain, which is supporting evidence that a concept’s geometry can be formed by the statistical dynamics of the latent variable behind it.
[NLP-48] C-Unseen: Weak Signal Detection in Dynamic Temporal Knowledge Graphs via LLM Reasoning
【速读】： 该论文旨在解决现有弱信号检测方法在动态时间知识图谱（DTKGs）中难以捕捉语义与关系结构的问题。传统方法如基于关键词频率、主题建模或无类型图拓扑的方法，无法有效识别那些早期、低可见度但具有潜在重大影响的弱信号。其核心挑战在于如何从海量异构数据中提取出语义连贯且稀疏出现的子图模式，并判断其随时间演化的真实传播趋势。本研究提出C-Unseen框架，其解决方案的关键在于构建一个自解释的双模块系统：首先通过大语言模型（LLM）结合思维链（Chain-of-Thought）推理，识别与当前主流图谱叙事存在语义张力的稀疏子图；其次利用弱信号警报模块，追踪这些子图在连续时间快照中的持续性，从而甄别出真正具有预警价值的弱信号。该方法突破了传统技术对形式化特征的依赖，实现了对复杂语义关系和动态演化模式的精准建模，显著优于关键词、主题及图基基线方法。
链接: https://arxiv.org/abs/2608.26870
作者: Yassir Lairgi,Ludovic Moncla,Khalid Benabdeslem,Rémy Cazabet,Pierre Cléau
机构: INSA Lyon, CNRS, UCBL, LIRIS, UMR5205, 69621 Villeurbanne, France; GAUC, Lyon, France
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Social and Information Networks (cs.SI)
备注: Accepted at the AI4SE 2026 Special Track, held within the WISE 2026 Conference
Abstract:Weak signals are early, low-visibility indicators that precede significant changes before those changes become established. Existing detection methods, based on keyword frequency, topic modeling, or untyped graph topology, fail to capture the semantic and relational structure through which such signals manifest. In this paper, we propose C-Unseen, a self-interpretable framework for weak signal detection in Dynamic Temporal Knowledge Graphs (DTKGs). We define a weak signal as a rare, semantically coherent subgraph that proliferates across consecutive TKG snapshots. The framework operates through two modules: a Rare Subgraphs Extractor, in which an LLM identifies subgraphs whose content is in tension with the dominant snapshot narrative via chain-of-thought reasoning, and a Weak Signal Alerter, in which the persistence of these rare subgraphs is tracked across time steps to isolate true weak signals. Experimental results demonstrate that C-Unseen outperforms keyword-, topic-, and graph-based baselines.
[NLP-49] Evaluating Confidence-Gated Retrieval with Matched Trajectory Replay
【速读】： 该论文旨在解决生成式语言模型代理在决策过程中依赖置信度信号进行即时回答、检索额外证据或延迟响应时，缺乏对动作轨迹层面后果评估的问题。现有方法通常孤立地评估置信度，未能衡量其引发行为的长期影响。为此，作者提出“匹配轨迹重放”（matched trajectory replay）这一受控实验协议，通过固定候选答案状态、证据点、预算和行动成本，系统比较原始口语化置信度与事后等距校准（isotonic calibration）在多跳问答任务中的表现。关键发现在于：在相同数值承诺阈值下，校准改变了代理最终选择回答的问题，显著提升了已承诺答案的准确率（最高提升41个百分点），但同时降低了覆盖范围并增加了检索频率。整体准确率在HotpotQA上最高提升15个百分点，但在MuSiQue上反而下降17个百分点，表明校准使系统转向更保守、低风险的操作模式，而非真正提升答案质量或置信度排序能力。此外，预先在检索前拟合的校准映射虽能改善前两层检索深度的校准性能，但在第三层深度对所有三类模型均劣于原始置信度。尽管额外证据平均带来收益，但无法证明置信度能有效识别哪些具体实例将从进一步检索中获益。综上，校准可使承诺风险更具可解释性，但无法估计额外检索的预期收益，因此检索仍需独立的价值信息（value-of-information）或效用评估机制。研究强调，评估应报告未见样本上的校准性能、风险-覆盖率权衡及检索成本。
链接: https://arxiv.org/abs/2608.26846
作者: Prateek Chhikara
机构: University of Southern California(南加州大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 5 Tables, 5 Figures
Abstract:Interactive language-model agents use confidence signals to decide whether to answer immediately, retrieve additional evidence (from memory or external knowledge), or defer. Yet confidence is usually evaluated in isolation, without measuring the trajectory-level consequences of the actions it triggers. We propose matched trajectory replay, a controlled protocol for comparing confidence-to-action mappings. The protocol holds candidate answer states, evidence points, budgets, and action costs fixed. We use it to compare raw verbalized confidence with post-hoc isotonic calibration in a multi-hop question-answering system using Mistral, GPT, and Qwen models on HotpotQA and MuSiQue datasets. At the same numerical commitment threshold, calibration changes which questions agents ultimately commit to answering. Across all six model-dataset pairs, it increases accuracy among committed answers by up to 41 percentage points. However, it can reduce coverage and increase retrieval use. Overall accuracy improves by up to 15 percentage points on HotpotQA but falls by up to 17 percentage points on MuSiQue. These effects reflect a shift to a more selective, lower-risk operating point, not improved answers or confidence ranking. A calibration map fitted before retrieval improves held-out calibration through retrieval depths one and two, but is worse than raw confidence at depth three for all three models. Additional evidence helps on average, but this aggregate effect does not establish whether confidence identifies which individual episodes will benefit from another retrieval. Taken together, these results show that calibration can make commitment risk interpretable, but it does not estimate the expected benefit of another retrieval. Retrieval therefore requires a separate value-of-information or utility estimate. Evaluations should report held-out calibration, risk-coverage, and retrieval cost.
[NLP-50] SymbolLKG: Towards Verifiable Logical Reasoning via Logical Knowledge Graph and Symbolic Solvers
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在复杂多步逻辑推理任务中因缺乏严格验证机制而导致的幻觉（hallucination）与不一致性问题。现有方法如思维链（Chain-of-Thought, CoT）缺乏对推理过程的严谨验证，而传统的检索增强生成（Retrieval-Augmented Generation, RAG）又难以捕捉逻辑任务中隐含的结构化依赖关系。其解决方案的关键在于提出一种神经符号（Neuro-Symbolic）架构，通过构建基于本体的逻辑知识图谱（Logical Knowledge Graph, LKG），将逻辑规则与约束作为拓扑层面的一阶节点进行显式建模，从而精确刻画文本中提取出的依赖关系；同时设计了拓扑感知的混合检索机制与动态逻辑路由（Logic Router），能够根据任务特性智能调度至最优的符号求解引擎。该框架在多个逻辑推理基准测试中显著优于当前主流提示工程与RAG方法，实现了更高的推理准确率与可验证的推理路径。
链接: https://arxiv.org/abs/2608.26836
作者: Haizhao Fan,Yuchi Xiong,Jize Wang,Xinping Guan,Xinyi Le
机构: Shanghai JiaoTong University (上海交通大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 19 pages, 9 figures
Abstract:Large Language Models (LLMs) have demonstrated remarkable proficiency in natural language understanding, yet they struggle with strict multi-step reasoning, frequently suffering from hallucinations and inconsistency. Existing solutions like Chain-of-Thought (CoT) lack rigorous verification mechanisms, while standard Retrieval-Augmented Generation (RAG) often misses the complex, structural dependencies inherent in logical tasks. To bridge this gap, we propose a Neuro-Symbolic architecture that integrates a Logical Knowledge Graph (LKG) with dynamic solver routing. Specifically, we introduce an ontology-based LKG that treats logical rules and constraints as first-class topological nodes, enabling explicit modeling of dependencies extracted from text. We further design a Logic Router to dynamically dispatch tasks to the optimal symbolic engine, which is supported by a topology-aware hybrid retrieval mechanism. Experimental results on logical reasoning benchmarks demonstrate that our framework significantly outperforms state-of-the-art prompting and RAG baselines, delivering higher accuracy and verifiable reasoning paths.
[NLP-51] RuleWeaver: Benchmarking Rule-Centered Scenario Reasoning for Large Language Models EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在专业化领域应用中，对复杂规则驱动场景推理能力评估不足的问题。现有基准测试或仅关注输出层面的指令约束，或忽视规则在具体情境推理中的差异化作用，难以全面衡量模型对领域规则的理解与运用能力。为此，论文提出RuleWeaver——一个面向规则中心化场景推理的基准构建框架。其核心创新在于：从语料库中提取IF-THEN元规则（IF-THEN Meta Rules），通过逐步增强生成复杂规则，并将这些规则构造成具有真实情境背景的问答实例。该框架不仅评估最终答案的正确性，还引入基于评分量表的过程级评价指标，包括答案质量、规则召回率和规则精确率，从而实现对推理过程的多维度分析。实验结果表明，当前11个代表性大语言模型在复杂规则推理任务中仍表现有限，即使最优模型也仅达到理论最大评分的约50%，凸显了该任务的挑战性。
链接: https://arxiv.org/abs/2608.26832
作者: Bohan Yu,Shi-Yang Li,Pengfei Cao,Jun Zhao,Kang Liu
机构: University of Chinese Academy of Sciences (中国科学院大学); Institute of Automation, Chinese Academy of Sciences (自动化研究所); The Key Laboratory of Cognition and Decision Intelligence for Complex Systems (复杂系统认知与决策智能重点实验室)
类目: Computation and Language (cs.CL)
备注: Accepted by EMNLP 2026 Findings
Abstract:Large language models (LLMs) are increasingly applied to specialized domains, where effective use of domain expertise often requires reasoning over complex rules in concrete scenarios. However, existing benchmarks only partially evaluate this capability, as they either focus on output-level instruction constraints or overlook the distinct roles that rules play in scenario reasoning. To address these gaps, this paper introduces RuleWeaver, a benchmark construction framework for evaluating rule-centered scenario reasoning. RuleWeaver starts from corpus-derived IF-THEN Meta Rules, progressively augments them into complex rules, and composes these rules into rule-centered scenario QA instances. Beyond final-answer correctness, RuleWeaver further supports process-level evaluation through rubric-based answer quality, rule recall, and rule precision. Experiments on 11 representative LLMs show that current models still struggle with complex rule-centered scenario reasoning, with even the best-performing model achieving only around 50% of the maximum rubric score. We make our code and dataset available here: this https URL.
[NLP-52] Behavior2Trip: Towards Personalized Travel Planning via User Behavior Trajectory EMNLP2026
【速读】： 该论文旨在解决传统旅行规划代理（Travel Planning Agent）过度依赖用户显式指令或多轮澄清来获取偏好信息的问题，而忽视了用户历史行为中隐含的丰富偏好信号。这种对主动输入的依赖不仅增加了交互负担，也限制了旅行计划的个性化程度。为此，论文提出了一项新任务——行为感知旅行规划（Behavior-Aware Travel Planning），其核心在于通过分析用户的过往行为轨迹，直接推断其潜在偏好，并生成高度个性化的旅行计划。该研究的关键解决方案是构建了一个大规模中文在线旅游平台数据集——Behavior2Trip，包含11,400个实例，涵盖平均39.8条行为记录、涉及5个偏好维度的14个属性；并设计了基于强化学习的B2T-Agent框架，该框架能够利用用户行为序列，调用外部工具进行偏好对齐的信息检索，并结合内部记忆模块实现长期偏好建模。实验表明，即使在最困难的任务上，GPT-4.1的全约束通过率也仅为0.5%，而基于Qwen3-8B的B2T-Agent显著优于所有基线模型，且其训练出的模型在通用旅行规划基准（TravelPlanner）上表现超越GPT-4.1，验证了该方法在复杂任务中的有效性与良好泛化能力。
链接: https://arxiv.org/abs/2608.26807
作者: Zihao Cheng,Yingyu Shan,Hongru Wang,Zeming Liu,Xinyi Wang,Xiangrong Zhu,Yuhang Guo,Wei Lin,Yunhong Wang
机构: Meituan Inc.(美团); Beihang University (北京航空航天大学); Beijing Institute of Technology (北京理工大学); University of Edinburgh (爱丁堡大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted by EMNLP 2026 Findings
Abstract:Travel planning agents assist users in generating personalized travel plans by modeling their individual preferences. Existing agents either rely on explicit user instructions or engage in multi-turn clarification to elicit user preferences. However, both approaches overlook the rich behavioral signals latent in users’ past behaviors, which implicitly encode their preferences. This over-reliance on active user input increases interaction burden and limits plan personalization. To bridge this gap, we introduce a new task, Behavior-Aware Travel Planning, which infers user preferences directly from past behaviors and generates personalized travel plans. To facilitate research on this task, we introduce Behavior2Trip, a benchmark constructed from one of the largest Chinese online travel platforms, comprising 11,400 instances. Each instance represents an average of 39.8 past user behaviors spanning 14 attributes across 5 preference dimensions. We further propose B2T-Agent, a reinforcement learning-based agent that leverages user behavior trajectories, interacts with external tools for preference-aligned retrieval, and maintains an internal memory module. Experiments on Behavior2Trip show that GPT-4.1 achieves a full-constraint pass rate of only 0.5% on the hardest tasks, while B2T-Agent built upon Qwen3-8B outperforms all baselines, highlighting the substantial challenge of this task. Moreover, Qwen3-8B trained with B2T-Agent also outperforms GPT-4.1 on the TravelPlanner benchmark, demonstrating strong generalization. Code and data are available at this https URL
[NLP-53] Instruction Quality Matters: Refining Instructions for Effective Preference Learning
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在偏好学习（preference learning）中因指令质量低下而导致的偏好信号不充分问题。现有方法依赖成对响应进行模型优化，但其信息量本质上受生成这些响应的指令质量制约；低质量或模糊的指令会限制响应质量的分布范围，抑制优质“优选”响应的出现，从而削弱偏好信号的有效性。研究通过最佳与最差N个样本分析揭示，指令质量同时影响响应质量的上限与下限。为此，论文提出一种指令精炼（instruction-refinement）流程：利用奖励信号识别弱指令，并基于规则引导的大型语言模型（LLM）反馈进行修正，实现无需丢弃数据即可提升偏好数据质量。在多个模型与基准测试中，无论是离线还是在线偏好学习场景，该方法均显著优于原始数据及其它数据增强策略，且进一步分析表明，指令精炼可提升可实现的响应质量水平，并与以响应为中心的数据筛选策略形成互补。总体而言，指令质量被证实是决定大语言模型（LLM）对齐过程中偏好信号信息量的关键因素。
链接: https://arxiv.org/abs/2608.26779
作者: Seohyeong Lee,Hwaran Lee,Buru Chang
机构: Sogang University (西江大学); Korea University (高丽大学)
类目: Computation and Language (cs.CL)
备注: Preprint
Abstract:Preference learning optimizes models using response pairs, yet the informativeness of these pairs is fundamentally shaped by the instructions from which they are generated. We identify instruction quality as a hidden bottleneck in preference learning: low-quality or ambiguous instructions restrict the response-quality distribution, limiting strong chosen responses and weakening preference signals. Through Best- and Worst-of-N analyses, we show that instruction quality constrains both the ceiling and floor of sampled response quality. Motivated by this observation, we introduce an instruction-refinement pipeline that selects weak instructions using reward signals and revises them with rubric-guided LLM feedback, improving preference data without discarding examples. Across offline and online preference learning settings, experiments on multiple models and benchmarks show broad alignment improvements over original data and alternative data-improvement strategies. Further analyses indicate that instruction refinement raises achievable response quality and complements response-centric preference data curation. Overall, instruction quality emerges as a key factor governing how informative preference signals are formed for LLM alignment. Code is available at: this https URL
[NLP-54] Letters hide the truth from our eyes: English homophones have meaningfully different phonetic realizations
【速读】： 该论文旨在解决同音词（homophones）在实际语音实现中是否存在非时长相关的语音差异，以及这些差异是否可由语境中的语义信息预测的问题。以往研究表明，同音词的发音持续时间会随使用频率变化而调整，但本研究进一步探讨了除时长外的其他语音特征是否也存在系统性差异。研究通过对美国电视新闻广播中14,000个同音词实例进行定量分析，发现如“weight”与“wait”等同音词对在实际发音中存在可区分的语音实现差异，且这些差异能够根据语境中的语义信息进行预测。这一现象在控制时长因素后依然显著，表明语义驱动的语音变异具有独立作用。研究的关键在于采用时间归一化频谱图（time-normalized spectrograms）作为分析工具，能够精细揭示语音实现的细微差别，避免传统音标转写带来的信息遮蔽，从而更真实地呈现语音的物理属性。
链接: https://arxiv.org/abs/2608.26749
作者: Yu-Hsiang Tseng,Mirjam T. C. Ernestus,Louis F. M. ten Bosch,R. Harald Baayen
机构: University of Tübingen, Germany(图宾根大学, 德国); Radboud University, the Netherlands(奈梅亨大学, 荷兰)
类目: Computation and Language (cs.CL)
备注: 54 pages, 9 figures
Abstract:The distribution of spoken word duration of English homophones is known to co-vary with frequency of use. This study investigates whether other aspects of the phonetic realization of homophones also differ. A series of quantitative investigations of 14,000 homophone tokens in American television news broadcasts revealed that the tokens of homophone pairs such as \textitweight and \textitwait have different phonetic realizations, and that these can be predicted from their meanings in utterance context. These systematic differences remain even when taking duration-related variation into account. Time-normalized spectrograms emerged as an excellent tool for probing the fine details of phonetic realization, and obviate the need for phonetic transcriptions, which inevitably hide the phonetic truth from our eyes.
[NLP-55] Preserving General Capabilities during Domain Specialization with Uncertainty-Calibrated MOPD
【速读】： 该论文旨在解决大语言模型在垂直领域专业化过程中面临的领域专用性与通用能力之间的权衡问题，即在提升特定领域表现的同时往往导致推理、编程、指令遵循及创造性写作等通用能力的退化。其核心解决方案是提出一种校准不确定性的多教师在线蒸馏方法（Uncertainty-Calibrated MOPD），关键在于通过双重温度采样扩大候选轨迹池，并结合正向优势密度过滤筛选出具有更强正向学习信号的轨迹；进一步采用中心化对数似然（CLL）过滤机制，基于熵校准的教师支持度分数，按方向一致性概率保留更新令牌，从而有效识别并强化可靠的学习信号。实验表明，该方法在角色扮演和医疗领域专业化任务中分别将通用能力平均提升4.73%和10.84%，同时保持垂直领域的性能表现，消融与诊断分析也验证了所提轨迹级与令牌级机制针对原有缺陷的有效性。
链接: https://arxiv.org/abs/2608.26735
作者: Ziyuan Liu,Jiao Ou,Jian Liang,Ruiming Tang,Cheng Luo
机构: Kuaishou Technology(快手科技)
类目: Computation and Language (cs.CL)
备注:
Abstract:Specializing large language models to vertical domains improves domain-specific behavior but often degrades general capabilities such as reasoning, coding, instruction following, and creative writing. We study this domain–general trade-off in Multi-Teacher On-Policy Distillation (MOPD), where a specialized student is supervised on its own sampled trajectories by domain and general teachers. Standard MOPD faces two limitations: ordinary on-policy sampling rarely exposes tokens with large positive teacher–student advantages, while the advantage sign alone does not establish whether the resulting update direction is reliable. We propose uncertainty-calibrated MOPD to address these limitations. Dual-temperature sampling broadens the candidate trajectory pool, and positive-advantage-density filtering selects trajectories with stronger positive learning signals. Centered log-likelihood (CLL) filtering then computes an entropy-calibrated teacher-endorsement score and probabilistically retains token updates according to direction–endorsement consistency. Experiments on role-playing and medical-domain specialization show that our method improves the general-capability average over standard MOPD by 4.73% and 10.84% , respectively, while maintaining vertical-domain performance. Ablations and diagnostic analyses further confirm that the gains do not merely result from a larger rollout budget and that the proposed trajectory- and token-level mechanisms address their intended failure modes.
[NLP-56] owards Expert Financial QA via Self-Improving RAG ICLR2026
【速读】： 该论文旨在解决专家级金融问答系统中因缺乏事实验证与可审计性而产生的数值幻觉（numeric hallucinations）问题，以及传统单次遍历检索增强生成（RAG）系统在监管合规性方面存在的不足。其核心解决方案是提出自提升检索增强生成（Self-Improving RAG）框架，通过三个专业化代理——检索（Retrieval）、推理（Reasoning）和判断（Judge）——由一个协调器统一调度，并引入基于反馈的自我修正机制。当判断代理对答案评分低于动态阈值时，系统将触发分层重试策略：扩大检索范围、优化提示工程并放宽接受标准。在FinanceBench（SEC文件问答基准）上的实验表明，该方法在人工标注指导下达到86%的准确率，同时实现36.4%的Lazarus率，即近40%的初始错误答案可通过针对性重试修复。关键发现在于，即使采用固定的检索流程并依赖判断代理驱动的重试机制，也能取得优异性能，且无需动态路由，从而保证了全流程的可解释性；所有决策均记录置信度分数，为受监管的金融应用提供了必要的审计追踪能力。
链接: https://arxiv.org/abs/2608.26706
作者: Junjie Xiong,Shawheen Ghezavat,Aum Hirpara
机构: University of California, Berkeley(加州大学伯克利分校); California Polytechnic State University(加州理工州立大学); Hofstra University(霍夫斯特拉大学)
类目: Computation and Language (cs.CL)
备注: 17 pages, 2 figures. Accepted at the ICLR 2026 Workshop on Advances in Financial AI
Abstract:Expert-level financial question answering requires both grounded verification to catch numeric hallucinations and audit trails for regulatory compliance, attributes that standard single-pass RAG systems lack. We take a step toward this goal with Self-Improving RAG, a framework that decomposes document QA into three specialized agents (Retrieval, Reasoning, and Judge) coordinated by an orchestrator with feedback-driven self-correction. When the Judge Agent scores an answer below a dynamic threshold, the system triggers retry with escalated strategies: broader retrieval, more careful prompting, and relaxed acceptance criteria. We evaluate on FinanceBench (SEC filing QA), where Self-Improving RAG achieves 86% oracle-guided accuracy (measuring agreement with gold answers) with a 36.4% Lazarus Rate, recovering nearly 4 in 10 initially incorrect answers through targeted retry. A key finding is that a fixed retrieval pipeline with judge-driven retry achieves strong results without dynamic routing, providing full interpretability. Every decision is logged with confidence scores, enabling the audit trails required for regulated financial applications.
[NLP-57] Prag Align: Evidence-Sensitive Reply Assistance Across Chinese and Japanese Appropriateness Judgments
【速读】： 该论文旨在解决多语言环境下回复辅助（reply assistance）中面临的语言能力与文化情境适切性判断难题，尤其关注如何在跨语言交流中生成既符合语义又具备文化恰当性的回应。其解决方案的关键在于提出一种名为PragAlign的框架，该框架通过将上下文理解（context reading）与选择性澄清（selective clarification）相分离，实现对语言和文化双重因素的解耦处理。实验结果表明，在中文评估中，PragAlign显著优于直接生成（Direct）与规则驱动（Rule）两种基线方法；在日文评估中，尽管整体差异不显著，但PragAlign展现出最高的首选率（top-rank rate），且在5个场景中的4个与中文组达成一致选择，揭示了跨语言共享及语言特异性判断模式的存在。这一发现为设计兼具语言敏感性与文化适应性的回复辅助系统提供了实证依据。
链接: https://arxiv.org/abs/2608.26700
作者: Xin Zhong,Satori Hachisuka
机构: 未知
类目: Computation and Language (cs.CL)
备注: 5 pages, 5 figures, 2 tables
Abstract:Reply assistance in multilingual settings requires linguistic competence and culturally situated judgments of appropriateness. We present PragAlign, which separates context reading from selective clarification, and evaluate it alongside Direct and Rule. Nine native Chinese speakers judged Chinese materials, while three native Japanese speakers judged matched Japanese versions. In the Chinese evaluation, PragAlign received significantly better ranks than both baselines. In the Japanese evaluation, Direct had the lowest mean rank, PragAlign had the highest top-rank rate, and the omnibus difference was not significant. The groups selected the same top condition in 5 of 10 scenarios, including four shared PragAlign selections. The results identify shared and language-specific judgment patterns and inform reply assistance designed to support linguistic and cultural understanding.
[NLP-58] Scaling phoneme-based TTS augmentation for ASR: A unified pipeline and controlled study
【速读】： 该论文旨在解决基于文本到语音（TTS）生成的合成语音在自动语音识别（ASR）训练中如何有效提升性能的问题，尤其关注合成数据的质量与规模对最终识别效果的影响。其核心挑战在于：合成语音的监督信号效果受文本选择、参考语音质量及合成数据量三者共同影响，而现有方法缺乏系统性的优化策略。为此，论文提出一种统一的基于音素的TTS-to-ASR增强流水线，以多语言TTS模型为基础，采用从头训练的F5-TTS架构并引入语言标识（language-ID）条件控制，实现了跨语言的高效合成。该流水线的关键创新在于提出音素频率引导的选择方法（Phoneme-Frequency-Guided Selection, PFGS），通过利用真实ASR训练标签中估计出的音素频率对候选语句进行排序，从而优先选择更有利于模型学习的文本内容。实验结果表明，在固定合成比例下，PFGS相较于仅使用真实数据训练显著降低词错误率（WER），在12个测试集上优于真实数据基线，最大相对WER降幅达19.3%；同时，参考语音过滤进一步提升了意大利语和法语Common Voice数据集上的性能，分别降低0.29和0.59点绝对WER。这些结果揭示了合成规模、候选文本内容以及参考语音质量是影响TTS驱动的ASR增强效果的关键调控变量。
链接: https://arxiv.org/abs/2608.26697
作者: Zhen Wang,TianRui Wu,RongQi Han,Hao Wu,Wei Liang
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Synthetic speech provides scalable supervision for automatic speech recognition (ASR), but its benefit depends on the selected texts, reference speech, and amount of synthesized data. We present a unified phoneme-based TTS-to-ASR augmentation pipeline built around a multilingual TTS model trained from scratch using the F5-TTS architecture with language-ID conditioning. The pipeline combines language-specific grapheme-to-phoneme conversion, reference-speech filtering, candidate-text selection, synthesis, and matched ASR continuation. We further propose phoneme-frequency-guided selection (PFGS), which ranks candidate sentences using phoneme frequencies estimated from real ASR training labels. Experiments with separate monolingual ASR systems for Arabic, French, Italian, and Portuguese span 13 test sets. Across the synthesis-scale sweep, random augmentation improves over matched real-only continuation on 11 test sets. Under a nominal 60% synthesis budget, PFGS improves over real-only training on 12 test sets and over random selection on 9. Its largest relative word error rate (WER) reduction against random selection is 19.3%. With target texts and synthesis counts fixed, reference-speech filtering reduces absolute WER by 0.29 and 0.59 points on Italian and French Common Voice, respectively. These results identify synthesis scale, candidate-text content, and reference quality as important control variables in TTS-based ASR augmentation.
[NLP-59] Beyond Reflection: Affirmation as a Promising Behavioral Marker Associated with Quality in Text-Based Counseling EMNLP2026
【速读】： 该论文旨在解决生成式 AI 辅助的文本心理咨询中，哪些咨询师行为与更高对话质量相关这一实证问题。现有研究多聚焦于“反映（Reflection）”策略，借鉴动机访谈（Motivational Interviewing）框架，但缺乏对其他策略的系统性评估。本文通过多层次分析，基于由专业咨询师与实习生在真实场景下开展的大规模日文文本咨询数据集 KokoroChat，新增了咨询师策略标签和来访者困扰水平标注。研究发现，在所考察的多种策略中，“肯定（Affirmation）”相较于“反映（Reflection）”更稳定地与会话质量正相关。跨数据集迁移实验进一步表明，该质量信号在非专业支持者参与的英文语料库 ESConv 中亦部分存在。研究结果为咨询师培训及情感支持系统的设计提供了实证依据。
链接: https://arxiv.org/abs/2608.26689
作者: Michimasa Inaba
机构: The University of Electro-Communications (电气通信大学); 1-5-1, Chofugaoka, Chofu, Tokyo, Japan
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 Findings
Abstract:While AI-assisted text-based counseling is gaining attention, it remains empirically unclear which counselor behaviors are associated with higher dialogue quality. Existing research often focuses heavily on Reflection, borrowing frameworks from Motivational Interviewing. To address this gap, we conduct a multi-layered analysis using KokoroChat, a large-scale Japanese text counseling dataset conducted by professional counselors and trainees, newly annotated with counselor strategy tags and client distress levels. Our results show that, under the quality indicators used in this study, Affirmation is more consistently associated with session quality than Reflection among the analyzed strategies. Cross-dataset transfer experiments further suggest that this quality signal can be observed to some extent on ESConv, an English dataset with non-expert supporters. These findings provide empirical implications for counselor training and emotional support system design. We release the additional KokoroChat annotations and experimental source code at this https URL.
[NLP-60] FOCUS RePAIR: Mitigating Text Degeneration via Token-Level Guidance for Pruned Large Language Models ICML2026
【速读】： 该论文旨在解决大语言模型（LLM）剪枝后出现的文本退化问题，特别是重复循环现象，尽管此时困惑度（perplexity）和任务准确率等指标保持稳定。其核心问题是：剪枝虽有效压缩模型，但会加剧生成过程中的动态不稳定性，导致解码陷入局部重复状态。解决方案的关键在于从令牌级（token-level）视角将退化分解为“进入循环的风险”与“循环持续性”两个维度，并揭示循环持续性由采样集中合理替代项的“逃逸质量”（escape mass）决定。基于此，作者提出两种后剪枝微调的指导目标：FOCUS通过重加权教师模型高置信度区域的蒸馏以抑制信息泄漏；RePAIR则采用以起始点为中心的正负续写对及间隔损失，主动促进合理替代路径并避免过早锁定于重复循环。实验表明，二者均能显著减少重复并提升生成质量。
链接: https://arxiv.org/abs/2608.26676
作者: Junyoung Lee,Sehyeon Park,Shinhyoung Jang,Seonha Ryu,Hojeong Kim,Hyunsei Lee,Il Hong Suh,Yeseong Kim
机构: POSTECH(韩国浦项科技大学); Daegu Gyeongbuk Institute of Science and Technology (DGIST)(大邱庆北科学技术院); Hanyang University (汉阳大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to ICML 2026 as a Spotlight
Abstract:Pruning is a practical approach to compress large language models (LLMs), but it can amplify text degeneration, especially repetition loops, even when perplexity and task accuracy remain largely unchanged. In this work, we present a token-level analysis of this failure mode by viewing decoding as a dynamical process that enters and persists in a small set of recurrent contexts. Our analysis decomposes degeneration into loop entry risk and loop persistence, and shows that persistence is controlled by the escape mass assigned to plausible alternatives within the token sampling set. Motivated by these findings, we propose two token-level guidance objectives for post-pruning fine-tuning. FOCUS reweights distillation toward high-confidence teacher regions to suppress leakage, while RePAIR uses onset-centered positive/negative continuation pairs with a margin loss to promote plausible alternatives and prevent early commitment to repetition loops. Experiments on open-ended continuation and instruction-based generation show that both methods consistently reduce repetition and improve generation quality.
[NLP-61] Do LLM s Understand Personality? Rethinking Persona Fidelity Evaluation through Structured Behavioral Inference EMNLP2026
【速读】： 该论文旨在解决大语言模型在模拟多样化人类角色时，如何有效评估角色一致性（persona fidelity）的问题。现有评估范式存在两大局限：一是依赖整体式大语言模型评判者，易产生“整体判断幻觉”；二是采用静态心理测量量表，无法捕捉动态对话中情境依赖的角色一致性需求。为此，论文提出一种基于系统功能语言学（Systemic Functional Linguistics, SFL）的可解释评估框架PRISM（Persona Reasoning with Inverse SFL-based Modeling），其核心创新在于将角色一致性评估重构为结构化的逆向推理任务。PRISM从任务建构（Task Framing）、人际立场（Interpersonal Stance）和语言风格（Linguistic Style）三个功能维度出发，对目标角色条件下的证据进行维度特定估计，并融合多维信号生成可解释、可审计的评估结果。实验表明，相较于传统整体评判方法，PRISM在准确性与稳定性方面均显著提升，为角色一致性评估提供了更可靠的技术路径。
链接: https://arxiv.org/abs/2608.26674
作者: Mengfan Li,Zesheng Wei,Xuanhua Shi,Yang Deng
机构: Huazhong University of Science and Technology (华中科技大学); Singapore Management University (新加坡管理大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted by EMNLP 2026 main conference
Abstract:As large language models are increasingly deployed to simulate diverse human characters, ensuring persona fidelity, defined as the extent to which an agent’s behavior consistently reflects the psychological and stylistic characteristics of a target persona, has become a critical requirement. However, existing evaluation paradigms primarily rely on either holistic LLM-based judges, which are prone to "holistic appraisal hallucination’', or static psychometric inventories, which fail to capture the context-dependent fidelity required in dynamic dialogue. To address these limitations, we propose PRISM (Persona Reasoning with Inverse SFL-based Modeling), a psycholinguistically grounded framework that reformulates persona fidelity evaluation as a structured inverse inference task. Inspired by Systemic Functional Linguistics (SFL), PRISM decomposes persona fidelity into three functional dimensions: Task Framing, Interpersonal Stance, and Linguistic Style. It estimates dimension-specific evidence over a persona-conditioned label space and aggregates these signals into an interpretable and auditable evaluation process. Experiments show that PRISM yields more accurate and stable judgements than traditional holistic judging, providing a more reliable framework for persona fidelity evaluation.
[NLP-62] Meta-Learning Where to Allocate Experts: Task-Conditioned Layer-Wise Compression for MoEs
【速读】： 该论文旨在解决混合专家（Mixture-of-Experts, MoE）模型在实际部署中存在的一致性专家激活策略与动态任务需求不匹配的问题。具体而言，现有方法通常采用固定层数和任务的活跃专家数量（k），忽视了不同网络层的角色差异、专家冗余随深度变化以及任务难度对计算资源需求的影响。传统方法或基于离线静态分配，或依赖局部路由信号进行令牌级激活控制，缺乏对任务级上下文的全局感知。为此，本文提出MetaNet，一种支持集控制器（support-set controller），能够为每一层动态预测专家保留阈值（expert-retention threshold）与有界路由偏置（bounded routing bias），从而实现层内自适应的专家激活调控。其核心创新在于：在保持主干网络、专家模块及路由器参数冻结的前提下，通过可学习的元控制器实现灵活的精度-计算量权衡。实验表明，在DeepSeek-MoE-16B-Chat上，保守设置下平均仅激活3.61个专家（较固定k=6减少40%），而准确率仍接近原模型（MMLU: 0.489 vs. 0.474）；激进设置下激活数进一步降至2.28个（减少62%），准确率下降约3.7个百分点；且在未重新训练的情况下，基于MMLU训练的控制器可有效迁移至C-Eval任务，平均激活2.90个专家（减少52%），达到0.386的准确率。该方案的关键在于利用元学习机制实现任务与层间动态适配的专家激活控制，显著提升计算效率而不显著牺牲性能。
链接: https://arxiv.org/abs/2608.26650
作者: Rongfeng Wang,Shichao Weng,Zhiqiang Wang,Xinyu Liu,Yang Yi,Peilong Zhou,Hongwei Tang
机构: Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所); Nanjing College, University of Chinese Academy of Sciences(中国科学院大学南京学院); Nanjing Institute of Information Superbahn(南京信息超级枢纽研究院); Dobot Robotics(达博机器人); School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences(中国科学院大学交叉学科研究院); University of Chinese Academy of Sciences(中国科学院大学)
类目: Computation and Language (cs.CL)
备注: 18 pages, 3 figures, 9 tables
Abstract:Mixture-of-Experts (MoE) models route each token to a subset of expert networks, increasing capacity while keeping per-token computation sparse. In many deployed MoEs, the number of active experts is fixed across layers and tasks, although layer roles and expert redundancy vary with depth and demand varies with difficulty. Existing approaches address only part of this setting: layer-wise allocations are usually determined offline and reused for all tasks, while token-level methods vary expert activation using local routing signals without task-level context. We propose MetaNet, a support-set controller that predicts, for each layer, an expert-retention threshold and a bounded routing bias. The backbone, experts, and router remain frozen. On DeepSeek-MoE-16B-Chat, MetaNet provides a tunable accuracy-expert-activation trade-off. Relative to fixed k=6, a conservative setting activates 3.61 experts on average (40% fewer) and achieves comparable MMLU accuracy (0.489 vs. 0.474), whereas an aggressive setting activates 2.28 experts on average (62% fewer) with accuracy approximately 3.7 percentage points lower. The MMLU-trained controller also transfers to C-Eval without retraining, activating 2.90 experts on average (52% fewer than fixed k=6) at 0.386 accuracy.
[NLP-63] Information-Guided Frontier Decoding: Contextual Utility-Driven Commitment in dMLLM s EMNLP2026
【速读】： 该论文旨在解决扩散型多模态语言模型（dMLLMs）在解码过程中因掩码标记提交顺序不当而导致的生成质量下降问题。现有基于置信度的解码策略倾向于优先处理局部易预测的标记（如标点符号），但此类标记未必具有上下文语义价值，导致关键语义锚点被延迟或忽略，进而削弱上下文传播能力并加剧错误累积。其解决方案的核心是提出一种无需训练的信息引导前沿解码（Information-Guided Frontier Decoding, IGFD）策略，通过综合考虑标记置信度、邻域不确定性与结构化提交风险对候选词进行排序，从而优先确保高可靠性语义锚点的早期提交，同时推迟脆弱的结构化标记。此外，引入动态候选前沿机制，在相同解码预算下仅允许在局部可扩展区域中选择标记，进一步提升解码效率与上下文一致性。该方法无需额外训练、辅助模型或前向传播，实验证明其在多模态理解、推理、定位及幻觉抑制等基准测试中，相较于现有解码策略在多数场景和不同扩散型多模态大模型（dMLLM）架构上均表现出更优性能。
链接: https://arxiv.org/abs/2608.26641
作者: Xingyou Fang,Jingxing Zhong,Xiaosong Yuan,Xiaofeng Zhang
机构: Fuzhou University (福州大学); Jilin University (吉林大学); Shanghai Jiao Tong University (上海交通大学)
类目: Computation and Language (cs.CL)
备注: Accepted to Findings of EMNLP 2026
Abstract:Decoding quality in diffusion multimodal language models (dMLLMs) depends heavily on the order in which masked tokens are committed. Existing confidence-based strategies prioritize locally easy tokens, but confidence does not necessarily reflect contextual usefulness. As a result, structurally easy tokens such as punctuation may be committed before informative semantic anchors, weakening context propagation and increasing error accumulation. We propose Information-Guided Frontier Decoding (IGFD), a training-free decoding strategy that ranks candidates using token confidence, neighborhood uncertainty, and structural commitment risk. IGFD encourages early commitment of reliable semantic anchors while delaying fragile structural tokens, improving contextual support during decoding. A dynamic candidate frontier further constrains token selection to locally expandable regions under the same decoding budget. The method requires no additional training, auxiliary models, or extra forward passes. Experiments across multimodal understanding, reasoning, grounding, and hallucination benchmarks show that IGFD consistently outperforms existing decoding strategies across the majority of benchmarks and diffusion MLLM backbones under identical decoding budgets.
[NLP-64] Which Metrics Save the Most Human Annotation? Prediction-Powered Evaluation and Meta-Evaluation EMNLP2026
【速读】： 该论文旨在解决非可验证任务中人类评估成本高与自动指标存在偏差之间的矛盾问题。其核心挑战在于如何在保持评估可靠性的同时实现高效、低成本的系统比较。解决方案的关键是提出“预测驱动评估”（Prediction-Powered Evaluation, PPE）框架，该框架通过结合少量人工判断与大规模自动评分，实现数据高效的系统对比，并在理论上保证无偏性。该方法设计了参数化与非参数化两种推断程序，分析了配对与非配对实验设计在效率上的权衡，并在六组WMT数据集上进行了验证。此外，论文引入“预测驱动节省率”（Prediction-Powered Saving Ratio, PPSR）作为元指标，直接衡量自动指标在PPE框架下可节省的人工标注成本，相较于现有系统级元指标，PPSR能提供更优的判别力与稳定性。总体而言，该研究将自动指标重新定位为降低人工标注成本的工具，而非替代人类判断，具有广泛适用于各类非可验证任务的潜力。
链接: https://arxiv.org/abs/2608.26638
作者: Mingqi Gao,Anthony Sicilia,Weiyan Shi
机构: Northeastern University(东北大学); West Virginia University(西弗吉尼亚大学)
类目: Computation and Language (cs.CL); Machine Learning (stat.ML)
备注: Accepted at EMNLP 2026 (Main). Code available: this https URL
Abstract:Across various non-verifiable tasks, human evaluation is reliable but expensive, while automatic metrics are more scalable but often biased. Building on prediction-powered inference (PPI), we propose prediction-powered evaluation, a framework that combines limited human judgments with large-scale automatic scores to obtain data-efficient system comparisons that are provably unbiased. We develop parametric and non-parametric procedures, analyze the efficiency trade-off between paired and unpaired designs, and validate the framework on six WMT datasets. We further introduce the Prediction-Powered Saving Ratio (PPSR), a meta-metric that measures how much human annotation an automatic metric can save when used within prediction-powered evaluation. PPSR directly targets metric utility for prediction-powered evaluation and yields more discriminative and stable metric rankings than existing system-level meta-metrics. Overall, our new paradigm reframes automatic metrics as tools for reducing human annotation cost rather than replacing human judgment, and applies broadly to non-verifiable tasks.
[NLP-65] he Thousand-Graph Hypothesis: A Testable Hypothesis of Task-Conditioned Relation Materialization in Repository-Level Code Reasoning
【速读】： 该论文旨在解决大规模软件代码库因超出模型上下文长度限制而导致的知识表示难题，同时克服传统方法中将仓库知识显式训练进模型成本高且易过时、本地检索难以捕捉分散需求以及显式关系图谱需持续维护等瓶颈。其核心解决方案是提出一种仅依赖实体的外部接口，并在推理阶段根据任务条件动态生成关系（task-conditioned relation materialization），从而实现高效、灵活的知识调用。该方案采用两层索引结构，分离全局路由与局部实体聚焦，有效提升对分布式需求的感知能力；实验表明，在无需预先构建实体-关系边的情况下，基础模型、单层索引与双层索引分别达到92.1%、94.2%和95.6%的成功率，验证了该方法在保持低维护成本的同时显著提升了复杂代码任务的求解性能。
链接: https://arxiv.org/abs/2608.26602
作者: Fei Ding
机构: Alibaba Group(阿里巴巴集团)
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL)
备注: 9 pages, 2 figures,
Abstract:Large software repositories are often beyond model context limits. Training repository knowledge into models is costly and quickly stale, while local retrieval can miss scattered requirements, and explicit relation graphs add ongoing maintenance burden. We propose an entity-only external interface with task-conditioned relation materialization during inference. A two-layer index separates global routing from local entity focus and is evaluated on DeepSeek-V4-Flash and SWE-bench Verified. The base, one-layer, and two-layer conditions achieve 92.1%, 94.2%, and 95.6% success, respectively, under zero pre-built entity-relation edges.
[NLP-66] Not Just Reason Not Just Scan: Reinforcement Learning for Proactive Scientific Error Verification over Academic Paper EMNLP2026
【速读】： 该论文旨在解决多模态大语言模型（Multimodal Large Language Models, MLLMs）在科学自主研究中缺乏主动分析能力的问题，即如何在无预设问题或证据的情况下，自主检阅学术论文、构建全局证据视图并做出可追溯的判断。现有方法在无需指定问题与证据的验证任务上支持有限，难以实现真正意义上的自主科研辅助。为此，论文提出一种基于强化学习的科学错误检测框架VERA-RL，其核心在于通过“推理—验证—扫描”（Reason–Verify–Scan）的迭代流程，驱动模型进行证据驱动的可验证推理。关键创新包括构建包含4,300个匹配推理链的VERA-13K数据集，覆盖科研全流程中的六类科学错误，并引入细粒度奖励机制以优化推理完整性、证据对齐性与错误定位精度。实验表明，基于VERA-RL训练的Qwen3-VL-8B模型在“扫描”阶段的表现已接近旗舰级模型如Gemini 3 Pro和Qwen3-VL-235B-A22B，显著提升了模型在无监督情境下的科学推理能力。
链接: https://arxiv.org/abs/2608.26596
作者: Rongjin Li,Yuanxin Liu,Hao Zhou,Fandong Meng,Jie Zhou,Xu Sun
机构: Beijing University of Posts and Telecommunications (北京邮电大学); State Key Laboratory for Multimedia Information Processing, School of Computer Science, Peking University (北京大学计算机学院多媒体信息处理国家重点实验室); WeChat AI, Tencent Inc. (腾讯公司微信AI)
类目: Computation and Language (cs.CL)
备注: Accepted by EMNLP 2026 Findings
Abstract:Multimodal large language models (MLLMs) are increasingly capable scientific assistants, yet they remain far from fully autonomous research. This transition requires models to actively inspect academic papers, build global evidence views, and make traceable judgments without prespecified issues or evidence. However, existing work provides limited task paradigms or training studies for such issue- and evidence-absent verification. We study this challenge through scientific error detection, where models must determine whether errors exist and justify them with evidence-based reasoning. To fill this gap, we present VERA-RL, a reinforcement-learning formulation for scientific error detection over academic papers. Following a Reason–Verify–Scan progression, we construct VERA-13K, a 12,900-sample dataset organized into 4,300 matched chains, covering 6 scientific-error categories across the research workflow and broad natural-science domains. We further introduce fine-grained rewards for reasoning completeness, evidence alignment, and error precision. Training Qwen3-VL-8B with VERA-RL substantially improves verifiable reasoning, approaching flagship MLLMs such as Gemini 3 Pro and Qwen3-VL-235B-A22B on Scan.
[NLP-67] Benchmarking Clinical Decision Pathway Adherence in Large Language Models
【速读】： 该论文旨在解决当前医疗大语言模型（Medical Large Language Model, LLM）评估体系中对临床决策路径（Clinical Decision Pathway, CDP）遵循性评估不足的问题。现有基准主要关注最终答案的准确性，而忽视了模型生成决策过程是否符合临床实践指南（Clinical Practice Guidelines, CPGs）的要求，导致对模型真实临床适用性的评估存在偏差。为此，论文提出MEGA-CDP基准，其核心在于构建一个基于2,274份中英文临床实践指南、通过“指南到病例”（guideline-to-case）流程生成的42,353个临床病例数据集，每个病例均配有明确参考的指南遵循型决策路径。该基准支持单轮病例描述与多轮交互两种场景，并引入面向决策路径的一致性评估框架，以量化模型生成路径与指南要求之间的匹配程度。实验结果表明，当前主流16种医疗大模型在生成符合指南的决策路径方面仍表现不佳，凸显了发展以决策路径为导向的评估方法的必要性，也验证了MEGA-CDP在推动医疗大模型实现真正指南遵循性方面的关键价值。
链接: https://arxiv.org/abs/2608.26592
作者: Nuo Chen,Xinyang Jiang,Zilong Wang,Zhifei Zhang,Xiaoye Qu,Jiajun Deng,Yulan Guo,Cairong Zhao
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Following clinical decision pathways (CDPs) defined by clinical practice guidelines is essential for safe and reliable medical decision-making. However, existing medical large language model (LLM) benchmarks mainly evaluate final-answer accuracy, providing limited evaluation of models’ ability to adhere to guidelines. To address this gap, we introduce MEGA-CDP, a benchmark for evaluating whether medical LLMs can generate guideline-adherent CDPs using provided guidelines as references. MEGA-CDP is constructed from 2,274 English and Chinese clinical practice guidelines through a guideline-to-case pipeline, yielding 42,353 clinical cases with explicit reference CDPs. It supports both single-turn vignette and multi-turn interactive settings, and introduces a CDP-oriented evaluation framework for measuring pathway consistency. Experiments on 16 representative LLMs show that reliable clinical decision support remains challenging for current models, demonstrating the need for CDP-oriented evaluation and the value of MEGA-CDP for advancing guideline adherence in medical LLMs.
[NLP-68] Surgical Alignment in Knowledge Graph Training for Clinical Diagnosis with Large Language Models
【速读】： 该论文旨在解决如何将生物医学知识图谱（Biomedical Knowledge Graphs, KGs）的信号有效融入大语言模型（Large Language Models, LLMs）以支持临床诊断推理这一关键问题，尤其关注不同集成方法对模型优化过程与知识迁移行为的影响。其解决方案的关键在于提出一种基于优化几何的分析框架，引入梯度干预密度（Gradient Intervention Density, GID）和梯度畸变（Gradient Distortion, GD）两个指标，用以量化优化器对预训练模型参数更新的广度与分布特性。研究发现，在知识图谱判别任务中采用KL正则化的训练范式可产生稀疏、局部化的参数更新（称为“手术式对齐”，surgical alignment），而任务特定的监督微调（Supervised Fine-Tuning, SFT）则导致密集更新。控制消融实验表明，目标函数与KL正则化独立贡献于更新稀疏性，且稀疏更新范式虽在域内准确率上可能低于SFT，却显著提升模型推理质量。因此，该研究强调：评估知识图谱-大语言模型融合效果时，必须结合准确性与优化几何诊断，方能全面揭示不同训练范式下的知识迁移机制。
链接: https://arxiv.org/abs/2608.26587
作者: Saksham Khatwani,He Cheng,Majid Afshar,Dmitriy Dligach,Yanjun Gao
机构: University of Colorado Anschutz Medical Campus(科罗拉多大学安舒茨医学校区); University of Wisconsin Madison(威斯康星大学麦迪逊分校); Loyola University(洛约拉大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Biomedical knowledge graphs (KGs) offer structured medical knowledge that can ground large language model (LLM) reasoning in clinical diagnosis application, yet how KG signal should be integrated into LLMs remains an open question. We present a systematic study spanning five KG task formulations, three training paradigms, two KGs, and three base LLMs. At the task level, all paradigms improve over the non-finetuned baseline, but methods with comparable in-domain accuracy show substantially different knowledge transfer behavior. We introduce Gradient Intervention Density (GID) and Gradient Distortion (GD) to measure how broadly an optimizer modifies the pretrained model. GID and GD together reveal a clear divide: KG-judgment training under KL regularization produces sparse, localized updates (a regime we term as surgical alignment), while task-specific SFT produces dense ones. A controlled ablation shows that the objective and KL contribute to sparsity independently, and the paradigms that produce sparse updates also improve reasoning quality, even when their in-domain accuracy is lower than task-specific SFT. Assessing KG-LLM integration thus requires complementing accuracy with optimization-geometry diagnostics. Our implementation can be found at this https URL.
[NLP-69] J-Zero: Unified Challenger–Solver–Judge Co-Evolution from Zero Data
【速读】： 该论文旨在解决自演化语言模型在不可验证领域（unverifiable domains）中自我提升能力不足的问题，尽管在可验证领域已取得显著进展，但后者在缺乏明确评判标准的场景下仍面临挑战。其核心解决方案是提出一种统一的“挑战者-求解者-裁判”协同演化框架——J-Zero（Judge co-adaptation from Zero data），关键在于通过三者间的协同机制实现无监督的持续优化：挑战者与求解者通过对抗式交互共同进化，前者生成日益复杂的任务，后者则相应提升响应质量；而裁判（Judge）则通过预定义的偏好对（preference pairs）进行自适应更新，其判断依据并非主观评分，而是基于求解者回答相对于挑战者生成内容以及分解重构后答案相对于单次生成答案的相对表现，从而实现无需人工标注的数据驱动演化。该方法在可验证与不可验证领域分别平均超越基线4.2和8.0分，并可在至少十轮迭代中持续改进，而传统基线在两轮后即出现性能退化。
链接: https://arxiv.org/abs/2608.26582
作者: Gyouk Chu,Myeongho Jeon,Eunho Yang
机构: KAIST(韩国科学技术院)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Self-evolving language models have recently emerged as a promising path toward superintelligence, with the advantage of reducing the cost of human supervision. While considerable progress has been made in verifiable domains, self-evolution in unverifiable domains remains substantially less explored. We propose Judge co-adaptation from Zero data (J-Zero), a unified Challenger–Solver–Judge co-evolution framework that supports self-improvement across both domains. The Challenger and Solver co-evolve through an adversarial interaction: the Challenger generates increasingly difficult tasks, while the Solver learns to produce higher-quality responses to them. In parallel, the Judge co-adapts using preference pairs whose ordering is known in advance from how each response was produced, i.e., the Solver’s answer over the Challenger’s, and its decomposed-and-recombined answer over its one-shot answer, rather than from the Judge’s own scores. J-Zero outperforms the baselines by an average of 4.2 points on verifiable and 8.0 points on unverifiable domains, and continues to improve through at least ten iterations, whereas the baselines degrade after two.
[NLP-70] Visual Information-Guided Parallel Decoding for Diffusion Multimodal Large Language Models
【速读】： 该论文旨在解决扩散多模态大语言模型（dMLLMs）在生成过程中如何高效选择解码顺序以提升输出质量的核心问题。现有方法通常基于预测置信度或信息增益来决定解码顺序，但未能充分考虑输入图像的视觉信息对解码路径的影响。其解决方案的关键在于提出一种视觉信息引导采样器（Visual Information-Guided Sampler, VIG-Sampler），该方法通过量化候选词元对图像词元的注意力分布，优先选择那些与图像内容关联更强的词元进行解码；同时引入约束机制，惩罚与已选词元具有相似图像注意力分布的候选词元，从而确保每一步解码均带来更高的信息增益。实验表明，VIG-Sampler在7个图文生成与视觉问答基准上显著优于现有方法，尤其在图像描述任务中平均提升19.3 CIDEr分，且仅需一半的解码步数即可达到更优性能。
链接: https://arxiv.org/abs/2608.26580
作者: Insu Lee,Wooje Park,Wonseok Shin,Jinwoo Son,Byonghyo Shim
机构: 未知
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Diffusion multimodal large language models (dMLLMs) have recently emerged as a new decoding paradigm for multimodal generation. Starting from a fully masked sequence, dMLLMs progressively decode the sequence by unmasking a subset of the remaining masked positions at each step. Since the selected tokens serve as the prediction context for subsequent steps, deciding which tokens to decode is crucial to the quality of the final output. The most common strategy prioritizes tokens based on a certainty measure that tends to favor tokens frequently observed in the training data. Recent approaches instead order tokens according to their influence on subsequent predictions, but do not explicitly account for the input image. We propose the Visual Information-Guided Sampler (VIG-Sampler), which prioritizes tokens based on their attention to image tokens. We further impose a constraint that penalizes candidate tokens whose image-attention distributions are similar to those of previously selected tokens, thereby increasing the information gain of the decoded subset. Extensive experiments on 7 captioning and VQA benchmarks with 3 open-source dMLLMs demonstrate the effectiveness of VIG-Sampler, which outperforms the Info-Gain Sampler by an average of 19.3 CIDEr points across the captioning benchmarks and surpasses it on COCO Caption while using only half as many decoding steps.
[NLP-71] Double Trouble: Bilingual Pretraining Leaves Language-Conditioned Effects in Shared-Language Representations EMNLP2026
【速读】： 该论文旨在解决多语言模型在对比研究中普遍假设经对齐的嵌入空间具有可比性，而这一假设在仅解码器结构（decoder-only）模型中可能不成立的问题。其核心挑战在于：尽管经过共享英语词汇的嵌入对齐后，词元嵌入（token embeddings）表现出相似性，但模型深层隐藏状态（hidden states）在处理相同语言时仍存在显著差异，这种差异源于上下文建模过程中的内部表示分歧，而非输入表示本身。解决方案的关键在于揭示并验证嵌入对齐无法消除深层语义表示的异质性——即隐藏状态不匹配现象在所有八种语言中均持续存在，且随中间Transformer层加深而加剧，说明问题根植于上下文编码机制。因此，该研究强调，在进行探针分析、可解释性研究或跨语言迁移任务时，不能简单将对齐后的多语言模型视为可互换实体，必须考虑其内部表征的实际差异。
链接: https://arxiv.org/abs/2608.26576
作者: Anjishnu Mukherjee,Ziwei Zhu,Antonios Anastasopoulos
机构: George Mason University (乔治梅森大学)
类目: Computation and Language (cs.CL)
备注: Published as a conference paper at EMNLP 2026 (Main)
Abstract:When researchers compare multilingual models for probing, interpretability, or cross-lingual transfer, they often align embedding spaces and assume that shared-language representations are comparable. We show that this assumption can be premature for decoder-only models. We pretrain paired 310M-parameter models (one English-only, one bilingual) across eight typologically diverse languages, separately controlling for English exposure, total compute, and document overlap. After aligning on shared English vocabulary, we test held-out words and find that token embeddings look similar after alignment, but the deeper hidden states that the model uses for prediction do not. This gap holds for all eight languages and survives controls for document overlap and alternative alignment methods. This hidden-state mismatch grows through middle transformer layers, suggesting that it arises from contextual processing rather than the input representations where alignment is performed. Embedding alignment can mask real differences in how models internally represent a shared language, which matters for any downstream study that treats aligned models as interchangeable.
[NLP-72] Dependency-Aware Revocable Decoding for Efficient Diffusion Large Language Model Inference
【速读】： 该论文旨在解决扩散型大语言模型（diffusion large language models, dLLMs）在追求高并行解码速度时面临的生成质量下降问题。尽管通过迭代去噪实现多标记并行解码可显著提升效率，但早期生成错误易传播并污染后续上下文，导致整体输出质量受损。现有可撤销解码（revocable decoding）方法虽能重新评估并修正不可靠标记，却忽略了不可靠标记可能同时污染验证上下文这一关键缺陷。针对此问题，本文提出无训练依赖的依赖感知可撤销解码（Dependency-Aware Revocable Decoding, DARD），其核心创新在于将标记划分为掩码、候选和未掩码三种状态，并设计一种选择性上下文机制，在验证候选标记时主动排除可靠性较低的上下文成分，同时自适应调节其对后续解码的影响。实验结果表明，DARD在3个开源dLLMs上的12项文本与多模态基准任务中均显著优于当前主流方法，相较于Saber在Flickr30K数据集上实现了2.71倍的加速和4.35点的CIDEr分数提升，有效改善了速度-质量权衡（Pareto frontier）。
链接: https://arxiv.org/abs/2608.26574
作者: Wooje Park,Insu Lee,Minyoung Noh,Jaeyun Jang,Sungmin Lee,Kyuhong Shim,Byonghyo Shim
机构: Seoul National University(首尔国立大学); Sungkyunkwan University(成均馆大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Diffusion large language models (dLLMs) offer a promising alternative to autoregressive generation by decoding multiple tokens in parallel through iterative denoising. However, increasing decoding parallelism often degrades generation quality, as early errors can contaminate later contexts. Revocable decoding mitigates this issue by re-evaluating decoded tokens and remasking unreliable ones, but existing methods overlook that unreliable tokens may also corrupt the verification context itself. We identify this failure mode and propose Dependency-Aware Revocable Decoding (DARD), a training-free framework that separates tokens into masked, candidate, and unmasked states. DARD verifies candidate tokens using a selective context that excludes less reliable tokens and adaptively regulates their influence on subsequent decoding. Experiments across 12 textual and multimodal benchmarks on 3 open-source dLLMs show that DARD consistently improves the speed-quality Pareto frontier over recent revocable decoding methods, achieving a 2.71 \times speedup and a 4.35-point CIDEr score gain over Saber on Flickr30K.
[NLP-73] SPT: Skills as Pre-Training Data for Agent ic Language Models
【速读】： 该论文旨在解决生成式智能体（agentic）语言模型在后训练阶段依赖昂贵的工具调用轨迹和代理轨迹数据进行行为监督的问题，这类数据的获取需要复杂任务环境、执行与验证流程，导致覆盖广泛工具与任务的成本过高。为降低对高成本标注数据的依赖，论文提出一种名为“技能预训练”（Skill Pre-Training, SPT）的中段训练方法，其核心在于利用公开可获取的多文件技能包（skill packages）作为训练数据源。这些技能包蕴含可复用的工具语义与工作流信息，传统上仅用于推理时的上下文提示，而本文将其转化为预训练阶段的结构化语料。关键创新在于引入“引用插入”（Reference Insert）策略——通过将引用文件靠近其在主指令中的提及位置，有效保留技能包内文件间的关联性，从而提升模型对跨文件依赖关系的理解能力。实验结果表明，相较于在通用或轨迹数据上进行中段训练，SPT在多种模型规模和后训练方案下均显著提升智能体性能，同时基本维持原有通用语言能力；进一步的数据混合实验也证实，将技能数据与通用渐进语料结合可带来额外增益。研究证明，技能包是训练具备强工具使用能力的生成式语言模型的重要且高效的高质量数据来源。
链接: https://arxiv.org/abs/2608.26563
作者: Yufei Sun,Yudong Li,Yiming Cheng
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Agentic (tool-using) language models are mainly trained on tool-call traces and agent trajectories during post-training. These data provide direct behavioral supervision, but producing them requires task environments, execution, and verification, making broad tool and task coverage expensive. Publicly available skills offer another source of training data: they encode reusable tool semantics and workflows but are typically used only as inference-time context. We introduce Skill Pre-Training (SPT), a mid-training method that applies causal language modeling to SkillCorpus, a collection of public multi-file skill packages, optionally mixed with general data. To preserve relations among files within each package, we also introduce Reference Insert, a reference-aware assembly strategy that places supporting files near their mentions in the primary instruction. Experiments across multiple model scales and post-training recipes show that SPT consistently improves agentic performance over mid-training on general or trajectory data, while largely preserving general performance. Data mixture experiments show additional benefits from combining skill data with general annealing corpora. These results indicate that skill packages are a valuable data source for pre-training agentic language models.
[NLP-74] SPEAR: Distilling Domain-Adaptive Reasoning Skeletons via Sequential Symbolic Alignment in Reinforcement Learning
【速读】： 该论文旨在解决基于强化学习的知识蒸馏中一个关键难题：在使用稀疏的基于结果的奖励时，模型缺乏足够的逻辑引导；而采用昂贵的神经过程奖励模型（Neural Process Reward Model, PRM）虽可提供密集信号，但计算成本过高。其解决方案的关键在于提出SPEAR（Symbolic Process Evaluation and Alignment Reward），一种无需训练且可即插即用的序列级在线策略知识蒸馏过程奖励方法。SPEAR将自然语言推理轨迹映射为领域自适应的符号化里程碑，并利用最长公共子序列（Longest Common Subsequence, LCS）对齐学生模型的探索路径与教师模型的里程碑，从而生成密集、有序感知的奖励信号，实现无需外部神经验证器即可强制逻辑一致性。实验表明，SPEAR在数学、科学及常识推理任务中能通过高效的密集过程奖励有效弥合师生模型间的推理差距。
链接: https://arxiv.org/abs/2608.26550
作者: Zhuochun Li,Yuelyu Ji,Yiming Zeng,Daqing He
机构: University of Pittsburgh (匹兹堡大学); University of Connecticut (康涅狄格大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Reinforcement learning-based knowledge distillation has the potential to transfer complex reasoning from teacher to student models, yet it currently faces a critical dilemma: researchers must choose between sparse outcome-based rewards, which provide insufficient logical guidance, or expensive neural Process Reward Models (PRMs) for dense signals. We resolve this by introducing SPEAR (Symbolic Process Evaluation and Alignment Reward), a training-free and plug-and-play process reward method for sequence-level on-policy distillation. SPEAR projects natural-language reasoning traces into domain-adaptive symbolic milestones, providing an efficient proxy for process-level reasoning alignment. By utilizing the longest common subsequence (LCS) to align student explorations with teacher milestones, SPEAR provides a dense, order-aware reward signal that enforces logical consistency without the need for an external neural verifier. Our experiments across math, science, and commonsense reasoning tasks demonstrate that SPEAR effectively bridges the reasoning gap between student and teacher models via sequence-level distillation with efficient dense process rewards. Our code and data are available at: this https URL.
[NLP-75] DuMateBench: Evaluating Autonomous Agents in Complex Real-World Workflows
【速读】： 该论文旨在解决现有自主代理（autonomous agent）评估基准在真实场景适应性方面的不足问题，即当前多数基准将任务按应用或能力维度割裂评估，且测试环境过于理想化，无法反映实际部署中复杂的动态环境。为此，论文提出DuMateBench——一个基于大规模生产级代理平台匿名化用户会话重建的真实会话基准，完整保留任务前的交互历史、持久化配置与工作区状态，并经人工验证确保真实性。该基准包含200个任务，覆盖8类典型场景和17个细粒度能力类别，多数任务需多能力协同完成。研究通过在隔离Docker容器中引入三类真实环境复杂性（资源不足、不稳定、噪声干扰），并采用混合确定性与大语言模型作为裁判（LLM-as-Judge）的评估协议，对五种代表性代理框架与四种先进大语言模型进行实验。结果揭示了严格任务完成率存在显著差距，且鲁棒性、效率与诊断分析表明，代理在环境扰动下的表现由大语言模型能力与代理框架特性共同决定。其核心解决方案在于构建兼具真实性和复杂性的评估体系，以更准确地衡量代理在现实世界中的综合性能。
链接: https://arxiv.org/abs/2608.26546
作者: Zechun Niu,Yukun Zhao,Jiaxin Zhang,Xu Shen,Jinhua Si,Han Tian,Can Xu,Yunfan Song,Jiaxin Mao,Yansong Gao,Yuchen Li,Jianmin Wu,Lingyong Yan,Shuaiqiang Wang,Dawei Yin
机构: Renmin University of China (中国人民大学); Shandong University (山东大学); Independent Researcher (独立研究员); Michigan State University (密歇根州立大学); Nankai University (南开大学); East China Normal University (华东师范大学); Imperial College London (帝国理工学院); Baidu, Inc. (百度公司)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Autonomous agents are increasingly adopted to complete complex, multi-tool workflows in real-world settings. However, existing benchmarks typically separate tasks by application or capability and evaluate agents in environments that are cleaner and more stable than those encountered in practice. We introduce DuMateBench, a real-session benchmark reconstructed from anonymized and privacy-screened user sessions collected from a large-scale production agent platform. Each task preserves the relevant pre-solution interaction history, persistent configurations, and workspace state, and is then validated through human verification. The resulting benchmark comprises 200 tasks spanning 8 broad scenarios and 17 fine-grained capability categories, with most tasks requiring multiple capability coordination. We execute these tasks in isolated Docker containers injected with three forms of real-world environmental complexity: Insufficient, Unstable, and Noisy, and assess performance using a hybrid deterministic and LLM-as-Judge evaluation protocol. Experiments across five representative autonomous-agent frameworks paired with four state-of-the-art LLMs reveal substantial gaps in strict task completion. Complementary robustness, efficiency, and diagnostic analyses further show that performance under environmental perturbations is jointly shaped by the capabilities of the LLM and the surrounding agent framework. The code and data are publicly available at this https URL.
[NLP-76] Multi-Expert Conformal Risk Control for Pairwise LLM Judging in Open-Ended Dialogue EMNLP2026
【速读】： 该论文旨在解决在开放式对话场景中，基于大语言模型（LLM）作为裁判（LLM-as-a-Judge）进行成对评估时的置信度控制与评分一致性问题。现有方法如共形风险控制（Conformal Risk Control, CRC）虽能通过拒答机制在决策阈值上控制风险，但其性能受限于单一专家评分尺度的局限性，尤其在异质化专家群体中难以实现高覆盖率。本文提出的关键解决方案是引入多专家协同机制，核心在于通过多专家聚合从源头净化评分函数，以增强鲁棒性与适应性。具体而言，作者设计了两种多专家CRC方法：基于得分平均的Score Averaging和基于决策投票的Decision Voting，分别在评分层与决策层实现聚合。然而，二者在异质化专家面板下仍受限于统一阈值无法匹配各专家不同评分尺度的问题。为此，论文进一步提出边际校准共现共识（Marginal-Calibrated Conformal Consensus, MC3），其关键创新在于利用初始阈值比率捕捉各专家特有的评分尺度，并联合优化一个在校准与测试阶段均一致应用的统一决策函数 $ C_t(x) $，从而保持交换性的同时实现跨专家尺度的自适应校准。实验结果表明，MC3在三个数据集上均显著提升了异质化专家面板下的覆盖率与准确性，验证了其在复杂评估场景中的有效性。
链接: https://arxiv.org/abs/2608.26529
作者: Ming Cheng,Yusheng Dai,Qiuhong Ke,Zhaolin Chen,Lizhen Qu
机构: Monash University(蒙纳士大学)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 Main Conference
Abstract:In this paper, we explore multi-expert Conformal Risk Control (CRC) algorithms for pairwise LLM-as-a-Judge evaluation in open-ended dialogue. Our core insight is that multi-expert aggregation offers a complementary remedy to CRC: whereas CRC controls risk at the decision threshold through abstention, aggregation sanitizes the scoring function at its source. Guided by this, we first design two multi-expert CRC methods: Score Averaging and Decision Voting, which aggregate at the score and decision levels, respectively. While both strategies outperform single-expert methods on homogeneous expert panels, on heterogeneous LLM judges they remain risk-valid but recover only limited coverage, because a uniform threshold cannot match the experts’ distinct scoring scales. To resolve this issue, we further propose Marginal-Calibrated Conformal Consensus (MC3): it captures distinct per-expert scales via initial threshold ratios, while jointly tuning a unified decision function C_t(x) applied identically in both calibration and test, thereby preserving exchangeability. To evaluate our framework, we construct Panel, a 1,800-pair human pairwise-preference benchmark for open-ended dialogue. It is built on responses generated by four open-weight LLMs over dialogue contexts from three domains (ESConv, MSC, DREAM), with full logit access. In experiments, we find that both Score Averaging and Decision Voting substantially improve accuracy and acceptance rate on homogeneous panels. Notably, MC3 extends these gains to heterogeneous panels by accommodating distinct per-expert scoring scales across all three datasets.
[NLP-77] Sycophancy Suppression Can Impair Rational Updating: Anti-Sycophancy Should Preserve the Ability to Update EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）中存在的“奉承行为”（sycophancy）问题，即模型在用户质疑时倾向于修改自身回答以迎合用户意见。这种现象可能源于两种不同机制：一种是模型无依据地迎合用户（Unsupported-Yielding），另一种则是基于用户反馈中蕴含的有效信息进行合理修正（Rational-Updating）。现有研究主要聚焦于抑制前一种非理性迎合行为，却忽视了其对后一种理性更新能力的潜在损害。为此，论文提出一种两轮评估框架，可独立测量这两种行为。实验结果表明，反奉承干预措施在降低Unsupported-Yielding的同时往往牺牲Rational-Updating，反之亦然，即使两者联合优化也难以避免此权衡。机制分析揭示，二者共享相似的内部表征基础：驱动两类行为的MLP神经元与注意力头存在显著重叠，且其对应的控制方向正向对齐。进一步开展的初步正交化控制探索显示，可在一定程度上实现选择性提升，但效果依赖于模型架构。综上，论文认为反奉承问题不应被简单视为抑制性目标，而应作为选择性优化问题来处理，有效干预策略需在保留Rational-Updating能力的前提下减少Unsupported-Yielding。
链接: https://arxiv.org/abs/2608.26511
作者: Huanhuan Ma,Henry Peng Zou,Chengze Li,Enze Ma,Yunyue Su,Philip S. Yu
机构: University of Illinois Chicago(伊利诺伊大学芝加哥分校); National University of Singapore(新加坡国立大学)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 Findings. Code and data: this https URL
Abstract:Large language models often exhibit sycophancy, revising their answers to align with users when users push back. Such answer flips, however, can arise from different causes. One possibility is that the model simply aligns with the user’s feedback in order to satisfy them. Another is that the feedback genuinely contains useful evidence, prompting the model to update its answer in a rational way. We distinguish them as Unsupported-Yielding and Rational-Updating. Prior work focuses primarily on suppressing Unsupported-Yielding, while overlooking its effect on Rational-Updating. We address this gap with a two-turn evaluation framework that measures the two behaviors separately. Across representative training-time and inference-time interventions, we find that anti-sycophancy methods often encounter a trade-off in which reducing Unsupported-Yielding can sacrifice Rational-Updating, and vice versa, even when the two objectives are optimized jointly. Mechanistic analysis suggests that the two behaviors share an internal substrate: the MLP neurons and attention heads driving them overlap substantially, and their associated steering directions are positively aligned. We further conduct a preliminary orthogonalized steering exploration, which yields modest, backbone-dependent selectivity gains. Overall, our results suggest that anti-sycophancy should be treated not as a simple suppression problem, but as a selectivity problem, where effective interventions should preserve Rational-Updating while reducing Unsupported-Yielding.
[NLP-78] A Single Suffix to Break Them All: Basin-Aware Jailbreaks for Merged Model Families EMNLP
【速读】： 该论文旨在解决模型合并（model merging）过程中潜在的安全风险问题，尤其关注在所有参与合并的微调模型均经过独立安全对齐的情况下，仍可能暴露于新型越狱攻击（jailbreak attack）的风险。传统观点认为，只要各子模型本身是安全的，其合并结果也应保持安全，但本文揭示了这一假设的局限性：即使所有子模型均经过对齐，合并过程仍可能暴露预训练基础模型（pretrained foundation model）中固有的、未被察觉的越狱漏洞。其解决方案的关键在于提出一种名为基底感知越狱（Basin-Aware Jailbreak, BAJ）的新方法，该方法将越狱提示生成建模为在合并空间中的极小-极大优化问题，旨在生成可跨共享同一预训练主干的多个合并模型迁移的对抗性后缀（adversarial suffixes）。实验表明，BAJ在多种预训练主干和合并设置下均表现出高且稳定的迁移成功率，并能有效绕过现有防御机制，凸显了模型合并场景下对齐安全性的脆弱性。
链接: https://arxiv.org/abs/2608.26506
作者: Yu Zhe,Yixin Tan,Junhao Wei,Wang Chen
机构: RIKEN AIP(理化学研究所人工智能研究中心); Institute of Science Tokyo(东京科学研究所); Zhejiang University(浙江大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Accepted by EMNLP findings 2026
Abstract:Model merging enables combining multiple fine-tuned models without additional training, but its safety implications remain poorly understood. Prior work primarily attributes merging risks to unsafe constituent models, implicitly assuming that merging individually aligned models preserves safety. In contrast, we show that model merging reveals a previously overlooked jailbreak risk rooted in the pretrained foundation model, even when all constituent models are individually safety-aligned. Motivated by this observation, we study a new threat setting where an attacker constructs jailbreak prompts that generalize across merged models sharing the same pretrained backbone, without access to the exact merging coefficients or constituent checkpoints. To exploit this phenomenon, we propose \textbfBasin-Aware Jailbreak (BAJ), which formulates jailbreak generation as a min–max optimization over the merging space to produce transferable adversarial suffixes across merged model families. Experiments across diverse backbones and merging settings show that BAJ achieves consistently high transfer success rates and remains effective under existing defenses.
[NLP-79] Compositional Generalization via Structural Identification in a Category-Theoretic Framework
【速读】： 该论文旨在解决传统评估中对组合泛化（compositional generalization）仅依赖模型准确率所带来的局限性问题，转而关注在未见的COGS（Compositionality in Generalization and Structure）示例中，哪些结构或词汇识别机制能够使其在训练所见结构的基础上具有可接受性。其解决方案的关键在于将句子建模为从句法地址到词汇项的函子（functor），并通过选择性坍缩（selective collapses）诱导出柯伊兰扩展（Kan extensions），从而实现观测到的语义关联的传播。该方法揭示了21种组合泛化类型中，不同识别模式决定了示例的可接受性，而残余失败则对应于未被支持的结构模板。该基于数据的诊断框架无需训练预测模型，即可刻画训练语料库在特定识别假设下所允许的结构范围。
链接: https://arxiv.org/abs/2608.26465
作者: Akihiro Maeda,Thomas Seiller,Yohei Oseki
机构: The University of Tokyo(东京大学); CNRS(法国国家科学研究中心)
类目: Computation and Language (cs.CL); Machine Learning (stat.ML)
备注: 11 pages
Abstract:Compositional generalization is usually evaluated through model accuracy. We instead ask which structural or lexical identifications make held-out COGS examples admissible from the structures observed in training. Sentences are represented as functors from syntactic addresses to lexical tokens, and selective collapses induce Kan extensions that propagate observed associations. Across 21 COGS generalization types, admissibility follows distinct identification profiles, while residual failures separate unsupported structural templates. These data-side diagnoses characterize what the training corpus licenses under specified identifications, without training a predictive model.
[NLP-80] Diff Mining: Logit Differences Reveal Finetuning Objectives ICLR2026
【速读】： 该论文旨在解决生成式语言模型在微调（Fine-tuning）过程中行为演化不透明的问题，即难以准确识别微调后模型究竟学到了哪些新行为或特征。随着模型能力的提升，这种不可解释性可能导致意外或有害行为的出现，因此亟需一种有效的机制来揭示微调所引入的核心特征。其解决方案的关键在于提出一种名为Diff Mining的框架，通过对比微调模型与基础模型在参考语料上的输出logits差异，自动识别出在微调过程中被显著增强的显著词元（salient tokens），从而形成反映微调目标的“指纹”。该方法仅依赖于模型输出的logits，无需访问模型内部结构，具备良好的可扩展性，适用于大规模模型。框架包含两个模块化阶段：首先计算上下文相关的logits差异，其次通过Top-K频率法或非负矩阵分解（Non-negative Matrix Factorization, NMF）对差异信号进行聚合，以分离出多个潜在的微调目标并生成可解释的词元集合。实验表明，Diff Mining在微调领域检测、偏见识别等任务中均显著优于现有模型差异分析方法，展现出在模型审计与可解释性分析方面的巨大潜力。
链接: https://arxiv.org/abs/2608.26462
作者: Greg Kocher,Robert West,Clément Dumas,Julian Minder
机构: EPFL; MATS
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 37 pages, 7 figures. ICLR 2026 Workshop: Principled Design for Trustworthy AI. Code available at this https URL
Abstract:Finetuning has become the gold standard for refining existing behaviors and inducing new ones in language models, yet it often remains unclear exactly which behaviors emerge during this process. As models grow ever more capable, understanding finetuning better becomes increasingly important, particularly since unwanted behaviors may arise during finetuning. In this paper, we introduce Diff Mining, a simple yet effective framework for identifying what a finetuned model has learned by comparing its logits to those of its base model. Diff Mining effectively surfaces salient tokens that are amplified in the finetuned model, serving as a fingerprint of its training – even on text unrelated to the finetuning domain. Unlike many existing model diffing methods which require model internals, Diff Mining only needs access to output logits and scales to large models. The framework consists of two modular stages: (i) extracting per-context logit differences between the finetuned and base models on a reference corpus, and (ii) aggregating the resulting signals to construct an interpretable token set representing the finetune. For aggregation, we explore both a simple Top-K frequency method and a Non-negative Matrix Factorization (NMF)-based approach for disentangling multiple finetuning objectives into distinct token clusters. Empirically, Diff Mining succeeds across diverse settings: on finetune domain detection, it significantly outperforms state-of-the-art model diffing methods both in identifying relevant tokens and in downstream performance when an interpretability agent is given access to the extracted token set; on models with injected biases, it identifies more than one third of the biases without targeted probing. Overall, our framework shows promise in developing auditing tools to detect finetuning objectives.
[NLP-81] Vowel Signs Are Not Letters: A Pre-tokenization Ceiling on Multilingual Tokenizer Fertility
【速读】： 该论文旨在解决基于字节级BPE（Byte-level BPE）的分词器在处理音节文字（abugida）语言时因继承GPT-2的仅包含字母的词界定正则表达式（\pL+）而导致的严重分词缺陷问题。其核心问题是：在音节文字中，元音以变音符号（combining marks）形式书写，而现有分词器将每个元音符号视为独立词汇单元，导致每个词被过度拆分，从而显著增加词元数量（即“肥力”fertility），且这种缺陷无法通过增大词汇量或调整语料组成来缓解。解决方案的关键在于引入一种对变音符号敏感的词边界识别机制（mark-aware word class），使分词器能够正确识别音节结构，避免在元音符号处错误切分。作者通过理论分析证明了这一缺陷导致的训练无关下界（training-free lower bound on fertility），并在26种语言的平行语料库上验证了17个音节文字均受此影响，肥力倍数从1.47×（藏语）至9.02×（泰语）不等，而拉丁、西里尔、韩文和汉字系统不受影响（肥力为1.00）。实验表明，仅通过修复分词器中的字符类定义，即可在尼泊尔语上实现从4.78降至1.58个词元/词的显著改善；当训练语料中尼泊尔语占比从5%增至95%时，修复后的分词器表现出稳定的性能变化，而未修复版本几乎不变，清晰区分出结构性缺陷与数据不足。三组相同规模（268M）模型对比显示，修复后的分词器在同等算力下实现4.43%更低的保留集比特/字节，并在更高算力下仍保持优势。此外，对HuggingFace 3,479个热门文本生成模型的普查发现，63.3%使用了仅含字母的词类定义，占总下载量的72.5%，揭示了该问题的广泛性。尽管GPT-4o已采用标记感知词类，但本文首次系统量化其价值、提供诊断方法、明确适用脚本范围、评估部署现状，并公开发布了一个包含65,536个词条的尼泊尔语-英语分词器及其可复现的基准测试框架，支持用户在普通笔记本电脑上从公开数据重新生成所有实验结果。
链接: https://arxiv.org/abs/2608.26449
作者: Sajal Regmi,Siddhartha Pudasaini,Chetan Phakami Pun
机构: Karela Technologies Inc.
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 14 pages, 2 figures, 12 tables. Code, tokenizer and reproduction harness: this https URL and this https URL
Abstract:Byte-level BPE tokenizers that use the HuggingFace ByteLevel pre-tokenizer inherit GPT-2’s word regex, where a word is defined as \pL+, one or more Unicode letters. In abugida scripts, vowels are written as combining marks; this pattern therefore splits each word at every vowel sign. Since BPE merges only within a pre-token, those splits persist through training regardless of vocabulary size or corpus composition. We formalise this effect as a training-free lower bound on fertility. Across 26 languages from a parallel corpus, every one of the 17 abugidas is affected, ranging from 1.47x (Tibetan) to 9.02x (Thai), whereas Latin, Cyrillic, Hangul, and Han show exactly 1.00x. For 5 languages, matched tokenizer pairs that differ only in this character class fall within 2.2% of the predicted floor, scoring 4.78 versus 1.58 tokens per word on Nepali. When the Nepali share of the training corpus is swept from 5% to 95%, the broken tokenizer barely shifts at all (1.7%) while the fixed one shifts 33.9%, which separates a structural ceiling from a data shortage without needing to inspect any code. We train three 268M models that differ only in their tokenizer; the fixed variant achieves 4.43% lower held-out Nepali bits per byte at equal compute, and it still leads when given the same bytes with 1.59x the compute. A census of 3,479 HuggingFace repositories finds the letters-only word class present in 63.3% of the most-downloaded text-generation models, accounting for 72.5% of their downloads. GPT-4o’s o200k pattern already uses a mark-aware word class, making the repair itself prior art. We quantify its value, show how to recognise its absence from symptoms alone, map which scripts it reaches, measure how widely it is deployed, and release a 65,536-entry Nepali-English tokenizer with a harness that regenerates every number here from public data on a laptop.
[NLP-82] Dont Overthink Dont Underthink: Toward Adaptive Reasoning in Agent ic AI
【速读】： 该论文旨在解决当前生成式AI（Generative AI）在复杂任务中因推理资源分配不当而导致的性能瓶颈问题，尤其聚焦于智能体系统（agentic AI systems）中动态演化任务需求下推理行为的失调现象。现有方法多依赖静态或预设的推理控制机制，如固定的令牌预算、事前难度估计或激活空间干预，且常在孤立的推理基准上评估，难以适应智能体在规划、工具调用、记忆检索及多智能体交互等过程中持续变化的推理需求。这种不匹配会导致推理过度或不足，进而引发不必要的计算开销、延迟增加、计划漂移、工具滥用或求解不完整等问题。论文的核心观点是：下一代智能体系统的根本挑战并非“进行多少推理”，而是如何根据任务演进动态、自适应地分配推理资源。研究通过分析过推理（over-reasoning）与欠推理（under-reasoning）这两种典型误配模式，在MATH-500和GAIA公开验证基准上的实证表明，过推理虽带来更高的计算成本却未显著提升准确率，而欠推理则始终伴随错误或不完整的解决方案。这一发现揭示了开发可自适应调节推理强度与时机的机制对于提升智能体系统效率与可靠性的重要性，为未来研究指明了方向。
链接: https://arxiv.org/abs/2608.26442
作者: Md Jueal Mia,M. Hadi Amini
机构: Florida International University (佛罗里达国际大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Recent advances in Large Language Models (LLMs) have shown that increased inference-time reasoning can improve performance on complex tasks. However, many existing approaches rely on fixed or preallocated reasoning controls, such as fixed token budgets, pre-execution difficulty estimates, or activation-space interventions, and are often evaluated on standalone reasoning benchmarks rather than full agentic workflows. These assumptions may not hold in agentic AI systems, where reasoning requirements evolve dynamically through planning, tool use, memory retrieval, and agent-to-agent interactions. Consequently, reasoning can become either excessive or insufficient, resulting in unnecessary computation, increased latency, planning drift, excessive tool use, or incomplete solutions. We argue that a major challenge for next-generation agentic AI is not merely how much reasoning a language model should perform, but how it should allocate reasoning according to evolving task demands. We characterize over-reasoning and under-reasoning as recurring failure modes of misallocated reasoning and evaluate them on MATH-500 and the GAIA public validation benchmark. Using tool-decision latency, token consumption, token-limit exhaustion, and answer correctness, our results suggest that cases classified as over-reasoning are associated with higher computational cost without proportional accuracy gains, whereas cases classified as under-reasoning are consistently associated with incorrect or incomplete solutions. These findings motivate future research on adaptive reasoning mechanisms for agentic AI.
[NLP-83] AfriSwitch: A Benchmark for In-the-Wild African Code-Switched Speech Recognition
【速读】： 该论文旨在解决非洲多语种对话中常见的代码切换（Code-Switching）现象在自动语音识别（ASR）系统中的建模难题。现有ASR系统普遍假设输入为单语种，且评估基准多为人工整理的单语数据，无法反映真实场景下多语言混合语音的复杂性。为此，研究提出了AfriSwitch——一个包含61.36小时人工转录的野外真实代码切换语音数据集，涵盖16种非洲语言及语言变体，附带逐句级英文片段标签、每句话的代码混合指数（Code-Mixing Index, CMI）及切换点数量。研究表明，非洲语言间的代码切换行为在“切换频率”与“混合平衡度”两个独立维度上存在显著差异，单一数值无法准确刻画语言的代码切换程度。在对五种开源与商用多语种ASR系统进行零样本测试时，其词错误率（WER）远高于同语言在单语基准上的报告结果，最佳系统平均WER达35.93%，无一系统在任一语言上低于24%。关键发现是：针对非洲语境的专项训练数据比模型规模或名义语言覆盖范围更能有效预测系统性能，表明本地化数据驱动的适配是提升跨语言语音识别能力的核心。
链接: https://arxiv.org/abs/2608.26434
作者: Gabrial Zencha Ashungafac,Busayo Awobade,Tobi Olatunji
机构: Intron Health
类目: Computation and Language (cs.CL)
备注:
Abstract:Code-switching is pervasive in bilingual African conversation, yet most ASR systems assume monolingual input and are evaluated on curated monolingual benchmarks. We present AfriSwitch, a 61.36-hour human-transcribed benchmark of in-the-wild code-switched speech spanning 16 African languages and language varieties, released with switch-level English span tags, perutterance Code-Mixing Index (CMI), and switch-point counts. Corpus statistics show that mixing behaviour varies widely across African languages along two largely independent axes: how often speakers alternate, and how balanced the mixture is. No single scalar captures how code-switched a language is. Benchmarking five open and commercial multilingual ASR systems zero-shot yields word error rates far above published monolingual figures for the same languages, with the best system averaging 35.93% WER and no system falling below 24% on any language. Africa-targeted training, not model scale or nominal language coverage, best predicts performance.
[NLP-84] SpeechGym: An Audio-Native Gym for Training Voice Agents via Reinforcement Learning
【速读】： 该论文旨在解决语音代理（voice agent）在真实语音交互场景中难以有效训练的问题，核心挑战在于现有方法依赖文本模态进行训练，而实际应用中语音代理需通过纯语音完成多轮对话与工具调用，导致训练与部署环境不一致。其解决方案的关键在于提出SpeechGym——一个音频原生的智能体环境，支持两个全模态模型在原始音频上直接对话，完全摒弃外部自动语音识别（ASR）和文本转语音（TTS）模块及专有语音API，使整个交互过程保持端到端可微分。该设计使得语音交互模态成为唯一变量，同时保留了原有任务、工具与成功判定标准，从而实现真正的语音原生训练。研究发现，语音代理的失败主要源于感知误差（如听错参数值）而非推理能力不足，且行为偏差（如未经授权的操作）亦可被环境自动标注并用于监督学习。针对训练信号稀疏性问题，采用基于每轮操作的成功奖励机制（per-turn process reward），显著提升梯度更新的方差，克服了传统仅基于结果的强化学习（GRPO）因多数轨迹均失败而导致的梯度匮乏问题。实验表明，经此方法训练后的模型无需额外调优即可在独立构建的语音基准上实现任务成功率翻倍，并将一个开源模型从排行榜末位提升至第二名，同时减少对话轮次与令牌消耗。
链接: https://arxiv.org/abs/2608.26432
作者: Jiajun Fan,Jingyuan Li,Prashanth Gurunath Shivakumar,Jia-Hong Huang,Qi Luo,M. Maruf,Ivan Bulyko,Ge Liu,Roger Ren
机构: 未知
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Voice agents must call tools and hold multi-turn dialogue entirely through speech, yet the dominant paradigm trains them in text. Existing frameworks either cascade TTS and ASR around a proprietary voice API, where gradients cannot flow and per-call cost makes on-policy reinforcement learning prohibitive, or stay in text: they measure voice agents but cannot improve them. We present SpeechGym, an audio-native agentic environment in which two omni-modal models converse in native audio, with no external ASR or TTS and no API boundary, over the unmodified tasks, tools and success check of an established text agentic benchmark, so that the interaction modality is the only variable and the loop stays local and trainable end to end. Audio agentic capability does not follow from audio understanding. The failures speech introduces are perceptual rather than reasoning deficits: the agent picks the right tool and the right argument slot but fills it with a value misheard from the waveform, and that single error cascades into a failed call, a retry of the same call, and a wasted step budget. A second failure is behavioural: under an insistent caller the agent performs an unauthorised write and ends the episode believing it helped. Both are trainable, because the environment labels them for free: a call with a misheard argument fails against the database while a correct one succeeds. The obstacle is sparsity, not signal. Outcome-only GRPO is gradient-starved here, since almost every rollout group fails identically, while a per-turn process reward crediting each successful tool call restores variance to nearly every group. Trained this way, the agent transfers with no further tuning to an independently implemented voice benchmark, more than doubling task success and carrying an open-weights model from last place to second on that leaderboard, while using fewer turns and tokens than before training.
[NLP-85] he Latent Diagnostic Taxonomy: A Framework for Constructing Classifiers and Diagnosing Their Decisions Applied to Prompt Injection Detection
【速读】： 该论文旨在解决生成式 AI 模型在面对提示注入攻击（prompt injection attack）时，其分类器所作出的高置信度决策缺乏鲁棒性的问题。具体而言，现有方法往往无法区分哪些高置信度预测是真正可靠的，哪些是因模型对特定输入标记（token）的敏感性而产生的脆弱性响应。为此，论文提出了一种名为“潜在诊断分类法”（Latent Diagnostic Taxonomy）的框架作为解决方案，其关键在于：首先通过交叉验证优化嵌入维度，构建一个非固定先验的降维分类器；其次，识别出仅占训练样本约29%的潜在支持向量（latent support vectors），这些向量对应于能显著改变分类器输出的关键提示片段；最后，基于这些敏感标记及其对应的攻击幅度，构建诊断分类体系，从而实现对提示的细粒度分类与处理建议。该分类法可将提示划分为三类：可安全依赖分类器判断的提示、需标记为启发式偏差或启发式覆盖的提示，以及需转交人工或安全审查的上下文不足提示。实验表明，高达77%的高置信度决策在移除单个标记后即失效，且这种脆弱性可归因于两类不同故障模式——置信度校准失败与可被利用的捷径学习。针对每一类诊断结果，论文进一步提出了相应的修复策略，实现了从检测到治理的端到端闭环。
链接: https://arxiv.org/abs/2608.26423
作者: Jaturong Kongmanee,Smile Thanapattheerakul
机构: Trend Micro(趋势科技)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR)
备注: 10 pages, 5 figures
Abstract:This paper proposes a framework for constructing a classifier as a safeguard layer, and for developing a complementary diagnostic that identifies which of the classifier’s confident decisions can be trusted. This framework, the Latent Diagnostic Taxonomy, consists of (i) constructing a dimensionality-optimized classifier, in which the embedding dimensionality is empirically selected via cross-validated performance rather than fixed a priori, (ii) locating a relatively small set of latent support vectors (~ 29% of total training examples) representing influential prompts for identifying tokens that alter the classifier’s predicted labels, and (iii) utilizing such tokens and their associated attack magnitudes for constructing a diagnostic taxonomy. This diagnostic taxonomy provides an end-to-end guideline for flagging prompts that require different treatments: rely Safely on the classifier’s decision; flag Heuristic Bias and Heuristic Override cases; route Insufficient Context cases for further human/safety review. Applying the framework to a classifier trained on a public prompt injection dataset, we find that a substantial fraction of its confident decisions (~ 77%) are not robust to removing a single token, and that this brittleness separates into two distinct failure patterns: a confidence calibration failure and a genuinely exploitable shortcut. For each zone of the taxonomy, we also recommend strategies for remediating diagnosed prompts. We illustrate the framework as a series of steps, demonstrating how each step operates.
[NLP-86] LowRankArena: A Standardized Evaluation Platform for SVD-Based LLM Compression
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）中基于奇异值分解（Singular Value Decomposition, SVD）的低秩压缩方法在现有研究中缺乏可比性的问题。由于以往评估采用不同的基准测试、不一致的压缩比率及多样的实验设置，导致难以将低秩压缩本身的效果与辅助技术的影响分离，进而无法准确判断性能提升是源于方法本身的改进还是评价协议差异所致。为应对这一挑战，论文提出LowRankArena——一个标准化的SVD-based LLM压缩评估平台，其关键在于统一任务版本、统一精度下的压缩预算、一致的对比范式以及可复现的推理测量流程，并公开超过3 TiB的压缩检查点以支持可复现研究。通过该平台对五种代表性SVD方法的系统审计发现，在标准化协议下，先前报告的性能优势具有高度条件依赖性：最优方法和性能层级随模型架构与保持率变化而转移；多项选择准确率可能掩盖困惑度（perplexity）的显著退化；名义上的低秩节省并未带来稳定且显著的端到端加速效果，其实际收益具有工作负载依赖性且通常有限。
链接: https://arxiv.org/abs/2608.26389
作者: Zishan Shao,Lixun Zhang,Kangning Cui,Wenhao Wu,Jinhee Kim,Yixiao Wang,Ting Jiang,Hancheng Ye,Qinsi Wang,Fan Yang,Danyang Zhuo,Yiran Chen,Hai Li
机构: Duke University (杜克大学); Wake Forest University (维克森林大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:SVD-based low-rank compression has become a fast-growing direction for reducing the memory and computational cost of large language models (LLMs). However, meaningful comparison across existing studies remains difficult as prior evaluations use varied benchmarks, inconsistent ratios, and diverse setups, often failing to isolate low-rank effects from auxiliary techniques. As a result, it remains unclear whether reported gains reflect method-level improvements or differences in evaluation protocol. This lack of comparability highlights the need for a unified, reproducible evaluation platform. To address this problem, we present LowRankArena, a standardized evaluation platform for SVD-based LLM compression. LowRankArena unifies task versions, uniform-precision compression budgets, comparison regimes, and inference measurements, and provides a reproducible pipeline with over 3 TiB released compressed checkpoints. Using LowRankArena, our aligned audit of five representative SVD methods reveals that prior findings are highly conditional under standardized protocols: clear leaders and performance tiers shift across backbones and keep ratios, multiple-choice accuracy can hide large perplexity degradation, and nominal low-rank savings yield workload-dependent and often limited end-to-end speedups. Our code is available at: this https URL.
[NLP-87] Co-Evolving Structured Knowledge and Reasoning in Language Models
【速读】： 该论文旨在解决现有检索增强型方法在利用非结构化文本进行知识补充时引入无关上下文、难以控制检索信息，以及结构化知识库构建成本高且推理能力脆弱的问题。其核心解决方案是提出KBevo——一种协同演化的框架，通过端到端联合优化知识库构建与问答推理过程，以问答任务的准确率作为奖励信号驱动两者共同进化。该方法的关键在于将知识库的质量提升与推理成功率直接耦合，使得推理成功能够反向优化知识库的构建质量，从而生成更庞大、连接更紧密、答案可达性更高的知识结构，显著提升了组合式事实推理能力与可控性，优于传统的检索基线方法。
链接: https://arxiv.org/abs/2608.26386
作者: Ryan Thomas Noonan,Linxi Zhao,Menghan Xu,Akanksha Sarkar,Mihir Mishra,Dongyoung Go,Kilian Q. Weinberger,Yoav Artzi,Jennifer J. Sun
机构: Cornell University (康奈尔大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: COLM2026. Code available at this https URL
Abstract:Retrieval-augmented methods improve factual accuracy by grounding language models in external knowledge, but retrieving over unstructured text often introduces irrelevant context and offers limited control over the retrieved information. Structured knowledge bases offer a more controllable alternative, yet they are expensive to construct and often brittle to reason over. To address these limitations, we propose KBevo: a co-evolving framework that jointly learns to construct a structured knowledge base and reason over it for knowledge-intensive question answering. By optimizing both components end-to-end with QA outcome rewards, our method enables reasoning success to directly improve the quality of the constructed knowledge base. This leads to larger, better-connected knowledge structures with higher answer reachability, while also improving compositional factual reasoning and controllability compared to standard retrieval baselines.
[NLP-88] Why RAG s Hallucinate: Penalty-Aware Evaluation of Retrieval-Augmented Generation Systems with Knowledge-Gap Canaries
【速读】： 该论文旨在解决当前检索增强生成（RAG）系统在评估中因“基于数量的准确率”（volume-based accuracy）而产生误导性结果的问题：即系统倾向于通过猜测来提高回答数量，从而获得更高的准确率评分，即使其答案缺乏事实依据。这一现象导致系统在知识库无法支持回答时仍强行生成内容，造成幻觉（hallucination）风险。其解决方案的关键在于提出一种惩罚感知评估框架（penalty-aware evaluation framework），包含三个核心机制：（1）采用非对称评分机制（正确+1、错误-4、拒绝回答0），以显著惩罚错误生成；（2）引入知识缺口探测器（knowledge-gap canaries），即答案明确不在知识库中的问题，用于识别由参数化记忆产生的无依据生成；（3）构建故障归因流水线（failure-attribution pipeline），可区分检索失败、生成失败与拒答策略失败。实验结果显示，尽管各系统在正确回答率上高度趋同（97.0%-98.0%），但在知识缺口探测器上的违规率差异高达六倍（16.7% vs. 98.1%），表明系统间真正关键差异在于“是否在不应回答时仍强行回答”。该框架通过惩罚感知评分重新排序系统性能，且结果在不同惩罚强度（k=1至k=9）下保持稳定，有效揭示了系统真实可靠性。所有代码、配置、对话记录与评审投票均公开，支持独立审计。
链接: https://arxiv.org/abs/2608.26385
作者: Alden Do Rosario,Hussein Younes,Felipe Pires
机构: CustomGPT.ai
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 13 pages, 1 figure, 5 tables. Code, full per-request logs, and all judge votes: this https URL
Abstract:Volume-based accuracy rewards retrieval-augmented generation (RAG) systems for guessing: a system that answers everything outscores one that declines when its knowledge base cannot support an answer. Building on the confidence-target analysis of Kalai et al. (2025), we present a penalty-aware evaluation framework for deployed RAG products, combining (i) asymmetric scoring (correct +1, wrong -4, abstain 0), (ii) knowledge-gap canaries, questions whose answers are verifiably absent from the knowledge base, so that any answer constitutes ungrounded generation from parametric memory, and (iii) a failure-attribution pipeline that separates retrieval, generation, and abstention-policy failures. Applying the framework to three commercial RAG systems and a no-retrieval baseline on SimpleQA-Verified (1,000 questions x 3 repeats, graded blind by a cross-family three-judge panel with 98.9% unanimity), we find that accuracy when answering is closely clustered across systems (97.0-98.0%), while canary violation rates differ roughly sixfold (16.7% vs. 98.1%). The systems are separated less by what they answer correctly than by whether they answer at all when they should not, and penalty-aware scoring reorders the volume-based ranking accordingly; the reordering is stable across penalty settings from k=1 to k=9. All code, configurations, transcripts, and judge votes are released for independent audit.
[NLP-89] Survival-Guided Length Control for Efficient Diffusion Language Models EMNLP2026
【速读】： 该论文旨在解决扩散语言模型（Diffusion Language Models, DLMs）在文本生成过程中因序列长度固定或依赖启发式停止规则而导致的冗余去噪步骤问题，进而影响推理效率。其核心解决方案是将序列结束长度的选择建模为对结束标记（end-of-sequence token）的离散时间生存分析（survival analysis）问题，并提出一种即插即用、无需额外训练的长度预测器，可无缝集成至任意现有DLM中。实验表明，基于生存分析引导的长度解码策略在推理速度上最高提升7倍，同时保持任务准确率不变；此外，研究发现同一数据集内预测长度存在显著差异，表明模型性能对所选长度高度敏感。
链接: https://arxiv.org/abs/2608.26374
作者: Ivan Kobyzev,Abbas Ghaddar,Yufei Cui
机构: Huawei Noah’s Ark Lab, Montreal Research Center (华为诺亚方舟实验室，蒙特利尔研究中心)
类目: Computation and Language (cs.CL)
备注: EMNLP 2026 (Main Conference)
Abstract:Diffusion language models (DLMs) generate text by iteratively denoising masked sequences, but standard decoding either fixes the sequence length or relies on ad hoc stopping rules, often leading to unnecessary denoising steps. We recast length selection as a discrete-time survival problem over the end-of-sequence token and propose a plug-in, training-free length predictor that can be added to any existing DLM. Across reasoning and code-generation benchmarks, survival-guided length decoding speeds up inference by up to 7 times while preserving task accuracy. We further find that predicted lengths vary widely even within the same dataset, making model performance sensitive to the chosen length.
[NLP-90] Knowledge-Verified Emergent Deception in LLM Agents Under Conflicting Incentives
【速读】： 该论文旨在解决在用户与部署方利益冲突的场景下，大型语言模型作为自主代理是否仍能保持诚实的问题，尤其关注当代理明知用户应得权益却因部署方利益驱动而选择否认时的欺骗行为。其核心挑战在于区分虚假陈述是源于知识缺失（无知）还是故意欺骗（幻觉或蓄意说谎）。为此，论文提出KnownLieBench——一个基于知识验证的基准测试框架，通过中立探针首先确认代理是否知晓用户的合法权利，再引入激励机制以评估其在压力下是否做出虚假声明。该基准覆盖8个客户服务领域、112个真实情境，采用多轮对话形式并集成信任追踪客户代理，能够分离仅由激励驱动的欺骗行为与受明确指令诱导的欺骗。实验表明，不同模型家族和领域间表现出显著差异的“涌现欺骗”现象；进一步的后训练研究表明，以诚实性为导向的微调可有效降低激励下的欺骗行为，而基于欺骗程度的微调虽提升谎言在无激励控制对话中的成功率，却未增加实际激励下的说谎频率。通过在评分前严格验证代理对权利的认知状态，KnownLieBench有效缓解了“说谎”与“不知情”之间的混淆，为代理诚实性的严谨审计与定向引导提供了可靠工具。
链接: https://arxiv.org/abs/2608.26372
作者: Zheyuan Liu,Weiliang Zhao,Xiangchi Yuan,Ningshan Ma,Yue Huang,Meng Jiang
机构: University of Notre Dame(圣母大学); Columbia University(哥伦比亚大学); Georgia Institute of Technology(佐治亚理工学院); Massachusetts Institute of Technology(麻省理工学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: A benchmark for knowledge-verified emergent deception in LLM agents under conflicting incentives
Abstract:Large language models are increasingly deployed as autonomous agents serving users on behalf of companies, placing them in settings where user and deployer interests can conflict. When an agent knows that a user is owed something its deployer would prefer to deny, does it remain honest? Answering this is difficult because false statements can reflect either ignorance or hallucination rather than deception. To address this challenge, we introduce KnownLieBench , a knowledge-verified benchmark that first confirms through a neutral probe that an agent knows a user’s entitlement, and then evaluates whether it makes false claims once an incentive to deny that entitlement is introduced. Specifically, KnownLieBench covers eight customer-service domains and 112 grounded cases, conducts multi-round dialogues with a trust-tracking customer agent, and separates deception emerging from incentive alone from deception produced under explicit instruction. Across eighteen proprietary and open-weight models, emergent deception varies substantially across model families and domains. We further use the benchmark for post-training, finding that honesty-directed fine-tuning reduces deception under incentive, while deception-graded fine-tuning increases lie success on honest-control dialogues without increasing lie frequency under incentive. By verifying entitlement knowledge before scoring deceptive behavior, KnownLieBench reduces the confound between lying and not knowing and enables more rigorous auditing and steering of agent honesty.
[NLP-91] Cross-lingual Representation Learning via Centroid Intervention Fusion EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在多语言任务中表现不均衡的问题，尤其针对低资源语言的跨语言迁移能力不足。现有推理时干预（inference-time intervention）方法通常为每对源-目标语言学习独立的映射投影，导致可扩展性差且难以实现跨语言知识共享。本文提出中心点干预融合（Centroid Intervention Fusion, CIF）框架，其核心创新在于将多个多语言干预投影整合为一个共享的语言通用算子，实现了跨语言知识的统一与复用。该方法在多语言常识推理、自然语言推理、事实编辑及机器翻译等多个基准上均显著优于现有最优的成对干预基线，平均提升达+3.378个百分点，并有效提升了低资源语言的表现。
链接: https://arxiv.org/abs/2608.26357
作者: Wei Sun,Marie-Francine Moens
机构: KU Leuven (鲁汶大学); Celestijnenlaan 200A 3001 Heverlee, Belgium
类目: Computation and Language (cs.CL)
备注: EMNLP 2026 (Main)
Abstract:Large language models (LLMs) exhibit uneven multilingual performance, especially when dealing with low-resource languages. Inference-time intervention offers a lightweight way to improve cross-lingual transfer by modifying the hidden states produced by the LLMs during the forward pass, without updating model parameters. However, existing cross-lingual intervention methods typically learn separate projections from source to target languages, which limits scalability and prevents knowledge sharing across languages. We propose Centroid Intervention Fusion (CIF), a projection fusion framework that consolidates multiple multilingual intervention projections into a single language-shared operator. Across multilingual commonsense reasoning, natural language inference, factual editing, and machine translation benchmarks, CIF outperforms the strongest prior pairwise intervention baseline by up to +3.378 pp on average across four model backbones, while supporting performance gains for low resource languages. The code is available at this https URL.
[NLP-92] MoganColBERT-TR: A Late-Interaction Multi-Vector Retrieval Model for Turkish
【速读】： 该论文旨在解决土耳其语（Turkish）场景下晚交互式检索（late-interaction retrieval）中模型表示能力与效率之间的平衡问题，特别是在缺乏高质量标注数据的情况下如何有效构建高性能的多向量检索模型。其核心挑战在于：传统单向量嵌入模型（single-vector embedding）在处理复杂查询时存在信息压缩损失，而直接训练大规模多向量模型又面临数据与计算成本高昂的问题。解决方案的关键是提出MoganColBERT-TR，一个基于已预训练的MoganBERT-embed编码器进行轻量化微调的多向量检索模型，通过仅需一阶段知识蒸馏（distillation）的方式，将交叉编码器（cross-encoder）教师模型的软标签（soft scores）迁移至目标模型。该方法利用字符级和句级划分的标题-段落对以及两个土耳其语问答检索数据集构建训练数据，并引入基于排名跳过（rank-based skipping）、组掩码（group mask）和余弦天花板（cosine ceiling）相结合的硬负样本挖掘策略以提升检索精度。实验在五个未参与训练的土耳其语BEIR基准数据集上进行，均采用零样本评估（zero-shot），结果显示该模型以1.49亿参数规模达到37.36的平均得分（nDCG@100为35.53，nDCG@10为31.81），在性能上优于两倍规模的ColmmBERT-base-TR及基准最大模型，仅次于mLateOn，尤其在长查询任务（ArguAna-TR）上仍存在可改进空间，表明其在保持高效性的同时显著提升了土耳其语多向量检索的准确率。
链接: https://arxiv.org/abs/2608.26344
作者: Furkan Yilmaz,Habibe Aleyna Tasdemir,Muhammed Faruk Gozay
机构: 未知
类目: Computation and Language (cs.CL)
备注: 12 pages, 6 tables. Model weights: this https URL
Abstract:We previously reported a ModernBERT encoder trained from scratch for Turkish (MoganBERT-TR) and a single-vector embedding model built on top of it (MoganBERT-embed). This work introduces the third model in that lineage: MoganColBERT-TR, a multi-vector retrieval model that, instead of compressing a query or a document into a single vector, represents it at the token level through a 768-128 projection and scores it with MaxSim late interaction. The model is not trained from scratch: the embedding model’s encoder is taken as the starting point and adapted to the ColBERT objective with a single-epoch distillation phase. Training data is produced from two sources - title-to-passage pairs carved out of our own pretraining corpus in the character domain and at sentence boundaries, and two Turkish question-based retrieval sets - and is distilled from the soft scores of a cross-encoder teacher (bge-reranker-v2-m3) over one positive and seven mined negatives. We show that in hard negative mining, rank-based skipping alone is insufficient and must be combined with a group mask and a cosine ceiling. Evaluation is carried out with the official pipeline of TurkColBERT, a benchmark built for Turkish late-interaction retrieval (PLAID index, exact MaxSim), on five Turkish BEIR datasets; none of them appears in our training pool, so all five results are clean zero-shot. With 148.9M parameters, MoganColBERT-TR reaches an overall score of 37.36 (35.53 nDCG@100, 31.81 nDCG@10) averaged over the five datasets and finishes second among the five models compared: it outperforms the twice-as-large ColmmBERT-base-TR on four of five datasets and by +3.05 overall, and the benchmark’s largest model by +12.30. The gap to the leading model (mLateOn) is concentrated on ArguAna-TR, the dataset with by far the longest queries.
[NLP-93] Neuro-symbolic PRM: Enhancing Scientific Reasoning via Structured Traces and Symbolic Verification
【速读】： 该论文旨在解决工具增强型大语言模型（Tool-augmented Large Language Models）在定量科学、技术、工程与数学（STEM）任务中多步推理时存在的关键残余缺陷：即中间推理步骤虽在语法上正确、数学上可执行且单位一致，但缺乏语义上下文的合理支撑（contextually ungrounded）。现有方法要么依赖无法评估语义意图的形式化验证器，要么使过程奖励模型（Process Reward Model, PRM）同时承担算术与逻辑校验的双重负担，导致性能受限。为此，本文提出一种神经符号框架，将推理过程形式化分解为两个独立维度——符号有效性（Symbolic Validity, V）与语义接地性（Semantic Groundedness, G）。其中，符号有效性通过确定性符号验证器作为硬过滤器严格保证；而语义接地性则通过在验证器接受的样本流形上条件训练PRM来评估。为高效训练该PRM，作者引入了一种新颖的数据合成策略——反事实符号扰动（Counterfactual Symbolic Perturbation, CSP），能够算法生成在形式上通过验证器但逻辑上错误的“硬负例”（hard negatives）。在推理阶段，采用验证器优先的约束搜索机制，在确保验证器覆盖操作的执行一致性的同时，仅依赖PRM对语义接地性进行排序。该方法精准针对强工具使用型大模型的核心残余错误类型，显著提升了推理可靠性，且避免了以往框架中冗长的启发式设计。
链接: https://arxiv.org/abs/2608.26329
作者: Yuxin Zi,Cong Xu,Suparna Bhattacharya,Martin Foltin,Amit Sheth
机构: AI Institute of South Carolina(南卡罗来纳州人工智能研究所); HPE Labs(惠普实验室); Indian AI Research Organisation(印度人工智能研究组织)
类目: Computation and Language (cs.CL)
备注:
Abstract:While tool-augmented Large Language Models have significantly improved multi-step reasoning in quantitative STEM tasks, a critical residual failure mode remains: intermediate reasoning steps that are syntactically well-formed, mathematically executable, and unit-consistent, yet contextually ungrounded. Current approaches either rely on formal verifiers that cannot assess semantic intent, or burden Process Reward Models (PRMs) with the dual task of checking both arithmetic and logic. In this paper, we propose a neuro-symbolic framework that cleanly decouples reasoning into two formal dimensions: Symbolic Validity ( V ) and Semantic Groundedness ( G ). We guarantee V by construction using a deterministic symbolic verifier acting as a hard filter. To assess G , we train a PRM conditionally on the verifier-accepted manifold. To train this PRM efficiently, we introduce Counterfactual Symbolic Perturbation (CSP), a novel data synthesis strategy that algorithmically generates constraint-preserving hard negatives (steps that perfectly pass the verifier but are logically flawed). At inference, we deploy a verifier-first constrained search that guarantees execution consistency for verifier-covered operations while relying on the PRM solely to rank semantic grounding. By targeting the exact residual error class of strong tool-using LLMs, our method significantly improves reasoning reliability without the sprawling heuristics of prior frameworks.
[NLP-94] How Unlikely Is “Unlikely”? Assessing Verbal Probability Perception Across Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在生成和理解口语化概率表达时是否存在语义一致性的问题，特别是这些表达是否与人类对不确定性的认知相匹配。其核心解决方案是通过基于人类基准的词到数值映射任务，对19个模型在两种条件下（强制单数值响应与要求解释）进行系统性跨模型评估，并引入一种新颖的双向往返测试以检验模型内部表征的一致性。关键发现在于：尽管模型在整体上高度复现了人类口语概率认知的结构特征——如词语顺序保持、三个锚点被准确恢复，且“可能”（possible）表现出最高方差和跨模型分歧，这与人类中该词的双峰解释一致——但模型在负向概率表达（如“unlikely”、“improbable”）上存在系统性偏高现象。此外，要求解释虽降低了模型内部的方差，却加剧了模型间的差异，而双向往返实验揭示出模型间存在明显分层，前沿模型能维持更连贯的双向表征。因此，该研究表明，尽管大语言模型能够模拟人类概率语言的认知结构及其固有偏差，但在负向概率表达上存在系统性偏离，这对人机之间概率语言交互的应用场景具有重要启示。
链接: https://arxiv.org/abs/2608.26327
作者: Christos Petridis,Konstantinos Pelechrinis,Zoran Obradovic
机构: Temple University (坦普尔大学); University of Pittsburgh (匹兹堡大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 10 pages
Abstract:Large language models increasingly produce and interpret verbal probability expressions, yet whether these expressions carry consistent meaning across models (or match human perceptions of uncertainty) remains unknown. We present a systematic cross-model evaluation using a word-to-number mapping task grounded in established human benchmarks. Eleven uncertainty expressions were presented to 19 models under two conditions, forced single-number response and explanation elicitation, alongside a novel bidirectional roundtrip test of internal consistency. LLMs track the human benchmark with surprising fidelity: word ordering is preserved, three anchor points are recovered, and possible'' shows the highest variance and cross-model disagreement of any expression tested, consistent with its documented bimodal interpretation in humans. However, models show a systematic upward bias for negative expressions such as unlikely’’ and ``improbable.‘’ Explanation elicitation reduces within-model variance while increasing between-model divergence, stabilizing individual models at the cost of inter-model consensus, and the roundtrip experiment reveals clear stratification, with frontier models maintaining coherent bidirectional representations. LLMs thus reproduce the structure of human verbal probability cognition, including its biases, while diverging systematically at the negative end—with implications for any setting where humans and models exchange probabilistic language.
[NLP-95] When Is Noise Response Universal? Tokenization as the Hidden Variable in Language Models
【速读】： 该论文旨在解决文本神经模型在输入受到噪声（如拼写错误、OCR错误或词语缺失）干扰时性能下降的问题。其核心发现是：模型对噪声的鲁棒性差异主要由训练目标决定，而非模型架构；具体而言，在词级噪声下，不同架构的模型表现出相似的性能退化曲线，而在字符级噪声下则出现显著分化。关键机制在于分词方式——单个字符的修改会触发分词器对周围词汇的重新切分，从而引发更剧烈的词元序列扰动，远超整词丢失的影响。基于此，研究提出可通过引入特定尺度的噪声进行增强训练，以在不依赖噪声评估的前提下有效提升模型在目标噪声水平下的鲁棒性。
链接: https://arxiv.org/abs/2608.26319
作者: Yefan Tao,Gerald Friedland,Luyang Kong
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 14 pages, 8 figures
Abstract:The performance of textual neural models often degrades when their inputs are corrupted by noise such as typos, OCR errors, or dropped words. We study the degradation rate across neural models, both sentence embeddings and decoder-only LLMs, and find that how consistent it is depends on the scale of the noise: under word-level noise, models with very different architectures decline along nearly the same curve, while under character-level noise they separate. We further identify the determining factor to be the training objective, not the architecture: eight encoders spanning six pretraining paradigms are scattered initially, and collapse onto a common curve after a short contrastive training recipe. We trace the word/character split to tokenization: a single character edit forces the tokenizer to re-segment the surrounding word, disturbing the token sequence far more than dropping a whole word does. This finding and its underlying mechanism provide a practical means to predict a model’s robustness to noise without any noisy evaluation, and to install robustness at a chosen noise scale through noise-augmented training.
[NLP-96] On Scope Classification and Current Knowledge-Editing Benchmarks: A Negative Result with INLAY as a Gradient-Free Case Study
【速读】： 该论文旨在解决现有知识编辑评估基准在衡量记忆编辑作用范围决策（scope decision）方面的根本缺陷。具体而言，当前主流的基于记忆的知识编辑评估方法无法有效评估在给定查询时，哪些已存储的编辑应被应用这一关键问题。其核心问题是：现有基准设计存在结构性缺陷——它们属于反事实型评测（counterfactual benchmarks），其评估目标要求模型输出“编辑后”的答案，而在此设定下，直接使用参数化知识进行回答在逻辑上即为错误，导致任何依赖于“是否应用编辑”的路由决策（routing decision）均无法获得正向反馈。作者通过构建无需梯度的编辑器INLAY，实现了对每条查询的精确真值标注（ground truth），并在三个数据集、三种输入条件下对1,689个查询执行所有候选路由动作分析，发现最优的路由策略（由理想路由器选择）与静态单行策略在所有九个实验单元中表现一致（差异小于0.0001），表明任意动态路由机制所能带来的最大收益接近于零。此外，拒绝响应（abstention）从未成为胜出策略，进一步验证了基准设计的不可靠性。研究通过人为构造缺失条件（即对一半样本隐藏查询自身的编辑记录），使整体可提升空间从+0.0000增至+0.0420，并首次赋予拒绝响应以获胜机会，从而直接证实了问题源于基准设计本身。该结论不仅适用于作者提出的INLAY系统，更普遍适用于整个“作用范围分类器”（scope-classifier）家族所依赖的现有评估体系。同时，研究也揭示了自身系统中的两个潜在漏洞，但经自审计确认，这些缺陷并未影响已发表的核心结果。
链接: https://arxiv.org/abs/2608.26292
作者: Aditya Pratap Singh
机构: 独立研究者(Independent Researcher)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 12 pages, 5 figures, 6 tables. Code and data: this https URL
Abstract:Every memory-based knowledge editor in the SERAC lineage depends on a scope decision: given a query, does a stored edit apply? We report that current knowledge-editing benchmarks cannot measure this decision at all. Using INLAY, a gradient-free editor we built to obtain exact per-query ground truth (the model is frozen, edits live in an external addressable memory, and applying an edit is a bias added along one token’s unembedding direction at decode time), we execute every candidate router action on 1,689 queries spanning three datasets and three input conditions. An oracle router choosing the best action every time ties a one-line static policy to four decimal places in all nine dataset-by-condition cells: the maximum attainable gain of any per-query routing method is 0.00 points. Abstention is the sole winning action zero times out of 1,689. The cause is structural: these are counterfactual benchmarks whose evaluation question asks for the post-edit answer, so answering from parametric knowledge is wrong by construction, and a benchmark without negatives cannot reward a classifier’s ability to reject. This generalizes beyond our system to the whole scope-classifier family the benchmarks are used to evaluate. We confirm the mechanism directly: constructing the missing condition ourselves, by withholding a query’s own edit from the index for half the sample, moves pooled headroom from exactly +0.0000 to +0.0420 and gives abstention its first wins. We also report where INLAY itself does not win (WISE beats it on Qwen2.5-7B CounterFact, and retrieval-augmented generation beats every method we tested, INLAY included, on rigorously matched RippleEdits), and disclose two bugs found during a self-audit of our own routing machinery, neither of which changed a published headline number outside noise.
[NLP-97] Comparing Chunking and Embedding Strategies for Turkish RAG Systems CEC2026
【速读】： 该论文旨在解决形态丰富的语言（如土耳其语）在检索增强生成（Retrieval-Augmented Generation, RAG）系统中，文档分块（chunking）策略与嵌入（embedding）模型选择对问答性能影响缺乏系统研究的问题。其核心解决方案在于通过一个完全交叉的实验设计，系统评估三种分块策略（固定长度、语义分块、布局感知的Docling）、五种嵌入模型及两种生成器大语言模型（LLM）在三类不同版式文档上的表现。关键发现表明：分块策略决定了嵌入模型选择的重要性——布局感知分块显著缩小了先进嵌入模型间的性能差距，使前三名嵌入模型在统计上无显著差异，表明语言专精并未带来可测量的检索优势；生成速度较快的模型并非更准确；且最优配置依赖于内容类型，布局感知分块对含表格的文档提升显著优于纯文本。因此，尽管个别组件表现优异，但其组合并不构成最佳整体配置，最终系统达到87.0%的最高性能。
链接: https://arxiv.org/abs/2608.26192
作者: Mustafa Sertaç Türkel,Fatma Nur Korkmaz,Ahmet Tuğrul Bayrak
机构: Ata Technology Platforms(Ata技术平台)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted to INTCEC 2026. This is the author’s pre-print version. The final authenticated version will be available through the conference proceedings
Abstract:How documents are segmented into retrievable chunks and how those chunks are embedded strongly affect Retrieval-Augmented Generation (RAG) quality, yet neither has been systematically studied for morphologically rich languages such as Turkish. We compare Turkish document question answering across three chunking strategies (fixed-length, semantic, and layout-aware Docling), five embedding models, and two generator LLMs, over three documents with contrasting layouts. The fully crossed design yields 9,000 graded question-answer evaluations, each scored by an independent judge model, and component comparisons are tested by paired McNemar tests under Holm correction. Four findings follow. The chunking strategy determines how much the embedding choice matters: layout-aware chunking compresses the spread between the modern embedding models to about a point. The three leading embedding models are statistically indistinguishable, so language specialization yields no measurable retrieval advantage. The faster generator is not the more accurate one. And the preferred configuration depends on content type, since layout-aware chunking helps documents containing tables far more than prose. The best individual components therefore do not compose into the best complete configuration, which reaches 87.0%.
[NLP-98] Invocation-Level Reliability of Tool-Using Agents
【速读】： 该论文旨在解决工具使用型智能体在多步推理任务中因早期错误（如选择错误工具或形成错误推理链）而导致下游结果被隐蔽性污染的问题。其核心挑战在于现有评估方法在面对模型自主运行（free-running）上下文时，无法有效区分工具调用正确性与推理路径的可靠性，尤其在任务深度增加后，模型自身错误会显著侵蚀其真实能力。解决方案的关键在于揭示了传统“精确匹配评分”机制的根本缺陷：一旦模型偏离预设的黄金轨迹（gold trajectory），后续所有判断均基于模型从未观测到的工具常量生成的固定答案，导致模型无法从中学习恢复，从而使得严重性（severity）和恢复能力（recovery）参数在统计上被评分规则强制锁定——严重性被固定为0，恢复能力则因无观测信号而不可见。作者提出一种名为“条件状态评分”（conditional-on-state scoring）的修正方法，该方法可对缓存的生成结果进行事后重评，无需额外计算成本，能够将严重性估计从边界值解放至内部区间（+0.149, +0.316），从而实现对模型真实鲁棒性和纠错能力的准确测量。
链接: https://arxiv.org/abs/2608.26189
作者: Afiya Noorain,Subhranshu Mohanty,Amritesh Banerjee,Abhijit Dasgupta
机构: SAI International School, Bhubaneswar, Odisha, India; University of Massachusetts Amherst, Amherst, Massachusetts, USA; SP Jain School of Global Management, Mumbai, India
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Tool-using agents fail two ways: choosing the wrong tool, or forming wrong arguments, and an early failure of either kind can silently corrupt everything downstream. We measure a correct-invocation rate that separates the two, under both a clean teacher-forced context and the model’s own free-running context, on five open-weight models over contamination-free multi-step tasks (depths 1-8). By depth 6, roughly 70% of a model’s own clean-context capability is lost to its own earlier mistakes (L6 = 0.686, 0.684). Our central finding concerns the measurement itself. Under exact-match scoring against a fixed gold trajectory, a propagation model’s severity and recovery parameters are not merely hard to estimate - they are fixed by the scoring rule. Severity is forced to its boundary (0 of 869 poisoned steps correct); recovery is structurally unobservable (0 of 580 poisoned steps returned on-track, against an expected 0.0058 by chance). Both follow from one mechanism: post-divergence, the gold value is generated by tool constants the model never sees, so it is information the model cannot derive. A fit run anyway returns 0.92 and 0.73 for a quantity that is exactly 1.000 - confident numbers for a parameter the scoring rule already determined. We give the mechanism and a remedy, conditional-on-state scoring, applied retrospectively to cached completions at zero additional cost, which un-pins severity to interior estimates excluding zero (+0.149, +0.316).
[NLP-99] When the Canonical Completion Is Wrong: Formalizing and Measuring the Jump in Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）是否能够完成从证据到新公理体系的归纳性跳跃（abductive jump）这一核心争议问题。当前学界对此存在分歧：一方认为LLMs在结构上无法实现此类跳跃，而另一方则质疑该论断的机制基础与实证支持。然而，由于缺乏对“跳跃”行为的严格形式化定义及可量化的评估方法，该争论长期难以定论。为此，本文提出一个四步形式化框架来系统定义和测量跳跃行为，重点在于第二步——即在何种条件下放弃默认补全（canonical completion）并正确转向非默认解。研究将跳跃实例定义为具有机器可验证证书的有限扩展问题，要求存在唯一且不同于默认补全（由左/右凯恩扩展，Kan extensions，给出）的正确解。通过证明跳跃实例的适定性，并建立一族可无枚举地生成任意难度实例的通用定理，研究实现了对跳跃能力的量化评估。实验在9个经认证的跳跃实例上对4个前沿模型进行测试，结果显示，在全部248次约束实验中，模型的凯恩默认率（Kan-default rate）均为零，表明模型始终能主动放弃默认补全，完成跳跃。高难度失败主要源于推理预算耗尽或约束错误，而非回归默认解。因此，关键结论是：第二步（放弃默认）并非瓶颈；若模型确有跳跃能力缺失，其根源可能在于生成约束条件或构建新框架的能力不足。
链接: https://arxiv.org/abs/2608.26187
作者: Dai Shi,Xiaoyu Li,José Miguel Hernández-Lobato
机构: University of Cambridge (剑桥大学); University of New South Wales (新南威尔士大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Logic in Computer Science (cs.LO)
备注:
Abstract:Whether large language models (LLMs) can perform the abductive leap from evidence to a new system of axioms, commonly referred to as a jump, has recently attracted considerable debate. A prominent position holds that LLMs are structurally incapable of such jumps, while recent studies challenge both its mechanism and its evidence. However, the debate remains difficult to settle, since the field still lacks a formal definition of the jump and a measure to test either side. In this paper, we develop a formal account of the jump in four steps and measure the second. The steps ask what the default completion of partial data is, when abandoning it is forced, when the abandonment is correct, and how successive jumps compound. Specifically, we define a jump instance as a finite extension problem with a machine-checked certificate that a correct completion exists, is unique up to renaming, and differs from the canonical completion of the data. The canonical completion is given by the left and right Kan extensions and is also what models produce without constraints, so it serves as the default. We prove that jump instances are well-posed and establish a family theorem that certifies instances of unbounded difficulty without enumeration. We further formalize when a jump is correct and how successive jumps compound. Finally, we run the measurement on nine certified instances and four frontier models. The Kan-default rate is zero in all 248 constrained trials, so the models do jump at this step and abandon the excluded default every time. Failures at higher difficulty stem from exhausted reasoning budgets or constraint errors, never from reverting to the default. These results indicate that the second step is not the bottleneck. If the disputed incapacity is real, it lies in generating the constraints or inventing the framework. Code can be found at: this https URL.
[NLP-100] Investigating the Influence of Prompt and Response Languages on LLM Content Generation
【速读】： 该论文旨在解决多语言环境下提示（prompt）与响应（response）语言对大型语言模型行为影响的问题，特别是探讨不同语言组合的提示如何系统性地影响模型输出的长度、语义保真度及跨语言关键词重叠。其解决方案的关键在于通过多维度量化分析揭示提示语言的非中立性：研究发现，使用挪威语作为提示语时，英文输出的响应长度平均缩短约37%；而使用英语作为提示语时，挪威语输出的响应长度则平均缩短约41%。尽管存在显著的词数差异，但基于LabSE的语义相似度仍保持较高水平，且软Jaccard（soft Jaccard）指标揭示了原始Jaccard无法捕捉的深层概念重叠，表明语义内容未因长度压缩而严重损失。此外，不同模型间效应量存在异质性，进一步说明提示语言对模型行为具有系统性塑造作用，这对多语言提示工程设计具有重要启示。
链接: https://arxiv.org/abs/2608.26186
作者: Thi Thanh Nhan Nguyen,Mai Khoi Tieu,Michael A. Riegler,Pål Halvorsen,Thu Nguyen
机构: Université de Technologie de Compiègne (贡比涅技术大学); Norwegian University of Science and Technology (挪威科技大学); SimulaMet (西穆拉科技); HUTECH University (胡志明市技术大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:This study examines how prompt and response language influence the behavior of large language models. Using five models, we evaluated answers to 68 non translation questions across four language conditions: English to English, English to Norwegian, Norwegian to Norwegian, and Norwegian to English. After removing refused items, the dataset contains 1348 responses. We measure length differences with Cohen d, semantic fidelity with LabSE cosine similarity, and cross lingual keyword overlap with both raw and soft Jaccard. Prompt language has a strong effect on response length. With English output, Norwegian prompts shorten responses by about thirty seven percent. With Norwegian output, English prompts shorten responses by about forty one percent. The largest cross lingual contrast shows a reduction in word count but a smaller reduction in tokens, reflecting tokenizer differences. Despite variation in length, semantic similarity remains high, and soft Jaccard reveals substantial conceptual overlap that raw Jaccard does not capture. Effect sizes vary across models, indicating heterogeneity. Prompt language is not neutral and systematically shapes output length and lexical realization, with implications for multilingual prompt design.
[NLP-101] PACEShop: Evaluating Personalized Actionable Compositional and Evidence-grounded Shopping Assistants
【速读】： 该论文旨在解决当前购物助手从传统排序式商品列表向结构化决策支持系统演进过程中，缺乏统一、可衡量的评估标准问题。随着系统需整合用户情境（shopping context）、产品证据（product evidence）与下一步行动建议（next-step guidance），推荐体验的质量不仅取决于语言流畅性，更依赖于个性化、可操作性、组件间一致性及证据可追溯性等多维度指标。现有评测体系虽覆盖了个性化、事实对齐（grounding）或大模型作为评判者（LLM-as-a-judge）中的部分维度，但未能形成对结构化购物助手响应的联合评估目标。为此，作者提出PACE（Personalized, Actionable, Compositional, and Evidence-grounded）评估框架，其核心在于定义一个任务匹配的、多维度协同的评估标准。解决方案的关键在于构建两个关键工具：PACEShop——一个包含22,625条结构化记录的基准数据集，涵盖可控人物画像、可审计证据池、良/劣标注及黄金缺陷类别与位置标注；以及PACEJudge——一种无需训练的评判协议，通过结构化输出契约（structured output contract）实现对诊断性字段的精准捕捉。实验表明，通用评判模型虽能识别整体质量，却无法还原PACE所需的诊断细节；而PACEShop使这些失败可验证，PACEJudge则在不重新训练的前提下显著提升了对用户源一致性、跨组件一致性、证据对齐性以及缺陷类别与位置闭合性的判断能力，揭示出真实购物助手评估的关键在于设计适配任务需求的输出契约，而非单纯依赖更强的模型架构或标量提示。
链接: https://arxiv.org/abs/2608.26180
作者: Weimin Lyu,Chen Luo,Guangrui Li,Yaochen Xie,Dhineshkumar Ramasubbu,Arief Koesdwiady,Wanqiu Long,Hansu Gu,Yutong Chen,Zheshen Wang,Dakuo Wang,Yi Liu
机构: Amazon(亚马逊); Northeastern University (东北大学); Expedia(易培途)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Shopping assistants are shifting from ranked product lists toward structured decision support, where systems must synthesize shopper context, product evidence, and next-step guidance into a coherent recommendation experience. This changes the unit of evaluation: a fluent response can still fail by ignoring shopper context, contradicting itself across components, or leaving defects too vague to localize. Existing personalization, grounding, and LLM-as-a-judge benchmarks cover pieces of this problem, but they do not define a joint evaluation target for structured shopping-assistant responses. We formulate this missing evaluation target as PACE: Personalized, Actionable, Compositional, and Evidence-grounded evaluation. We instantiate PACE with two artifacts: PACEShop, a benchmark dataset that makes the target measurable through 22,625 controlled records with structured personas, auditable evidence pools, GOOD/BAD labels, and gold defect family and location annotations; and PACEJudge, a training-free judging protocol that makes the target reportable through a structured output contract. Our experiments show that generic judges can recognize broad quality but fail to recover the diagnostic fields required for PACE; PACEShop makes these failures verifiable, and PACEJudge improves persona-source, cross-component, grounding, and family/location closure without retraining, showing that realistic shopping-assistant evaluation requires a task-matched output contract rather than only a stronger backbone or scalar prompt.
[NLP-102] A Multi-Framework Comparison of Outline Stages in Long-Form Generation with LLM s
【速读】： 该论文旨在解决大语言模型在长文本生成中面临的根本性局限问题，特别是700亿参数模型在输出超过16,000个词元时出现的“长度坍缩”现象，以及多章节故事生成中常见的“中间迷失”（lost-in-the-middle）导致的属性漂移问题。现有“先提纲后写作”的范式虽被广泛采用，但以往研究多聚焦于最终生成文本的评估，混淆了提纲与成文两个应解耦的评价对象。为此，作者构建了一个统一的头对头基准测试，覆盖7种代表性长文本生成框架，并在单章、多章和全书三个生成粒度层面进行评估；提出一种基于锚点的大型语言模型作为裁判（LLM-as-a-judge）协议，直接以源文本为参照，在五级锚定量表上对提纲质量进行量化评估。实验结果表明，无任一框架在所有场景下均占优，性能表现高度依赖于框架内在输出形式与目标粒度的匹配程度：SuperWriter在受限长度的单章模式下表现最优，但在全书生成模式下优势显著下降。更重要的是，提纲评分与成文评分之间的相关性仅为中等水平，有力支持了提纲与写作过程应解耦评估的理论原则。由于计算资源限制，当前写作侧评估仅覆盖部分样本，后续工作将扩大样本规模并引入跨模型评估者，以增强统计推断能力。
链接: https://arxiv.org/abs/2608.26177
作者: Yifan Song
机构: Taiyuan Institute of Technology (太原学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 20 pages, 13 tables
Abstract:Long-form generation exposes fundamental limitations of large language models. Even 70B-parameter models exhibit length collapse at 16k-token outputs, and multi-chapter stories frequently trigger the attribute drift characteristic of the lost-in-the-middle'' effect. The outline-first, write-later’’ paradigm has gained wide adoption, yet existing research evaluates the final writing rather than the outline itself, conflating two evaluation objects that should be decoupled. We construct a unified head-to-head benchmark covering 7 representative long-form generation frameworks across 3 generation granularities – single-chapter, multi-chapter, and whole-book – and propose an anchor-based LLM-as-a-judge protocol that directly assesses outlines against the source text on a 5-point anchored scale. Across 21 framework-granularity cells, no single framework dominates; performance depends on the match between a framework’s intrinsic output form and the target granularity. SuperWriter ranks first in the length-constrained single-chapter mode, but this advantage degrades in whole-book mode. The outline-side ranking correlates only moderately with the writing-side ranking, supporting the outline–writing decoupling principle. Compute constraints limit the writing-side evaluation to a subset of cases; follow-up experiments will expand the sample size and add cross-model evaluators to enable stronger statistical inference.
[NLP-103] Knowledge Cards: Structured Knowledge for AI Systems
【速读】： 该论文旨在解决当前人工智能（AI）系统在决策支持中缺乏可审查、结构化的知识表示问题，尤其针对代理型AI（agentic AI）系统在从输入到输出之间关键推理层的透明性缺失。现有文档形式如模型卡（model cards）、数据卡（data cards）和系统卡（system cards）虽能描述模型行为、训练数据及部署风险，但均未涵盖系统所依赖的核心概念、概念间关系及其推理逻辑。对于模式识别任务而言，这一知识空白尚可容忍；但对于需基于结论自主行动的代理型AI系统，该缺失正是将原型验证与可信赖的生产级应用之间分隔开的关键环节。为此，论文提出“知识卡”（Knowledge Card）这一新型结构化文档，用于以形式化领域本体（formal domain ontology）为根基，对单一限定概念（如特定故障模式、合规义务或流程决策）进行建模，完整记录涉及实体与关系、推理链条、推理失效条件以及每项主张的溯源信息，并由领域专家审核签发。该方案的核心在于通过可审计、可验证的知识表示，实现人机协同的可信推理，从而提升复杂决策场景下AI系统的可靠性与可解释性。初步原型已在能源与制药领域构建，其数据模式已作为公开草案向社区开放协作。
链接: https://arxiv.org/abs/2608.26176
作者: Liliana Ferreira
机构: Mondegreen.ai(莫德格林人工智能公司)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 11 pages
Abstract:AI systems whose outputs inform real decisions, and increasingly consequential ones, require something that current documentation practice does not provide: a structured, inspectable representation of the knowledge they need to ground, contextualize, and reason about those decisions, ideally reviewed and signed off by a domain expert. Established documentation artefacts already capture important aspects of an AI system. Model cards describe how a system behaves, data cards describe what it was trained on, and system cards describe the risks of a deployed system. None of them addresses the layer between inputs and outputs, more precisely, the concepts a system holds, the relationships it models, and the patterns of reasoning it applies. For pattern-recognition tasks this gap is tolerable. For agentic AI, where systems act on their conclusions, it is the step that most often separates a promising proof of concept from an operational solution an organisation can rely on. This paper introduces the Knowledge Card, a structured artefact that captures validated knowledge about a single bounded concept in a form that experts can review, organisations can audit, and AI systems can reason over. For one concept, such as a specific failure mode, a compliance obligation, or a process decision, a Knowledge Card records the entities and relationships involved, the reasoning that connects them, the conditions under which that reasoning no longer holds, and the provenance of every claim, all grounded in a formal domain ontology and signed off by a domain expert. Initial prototype cards have been built in the energy and pharmaceutical domains. The schema is released as a public draft for community engagement.
[NLP-104] Lost in Compression: A Controlled Cross-Lingual Audit of Extractive Prompt Compressors
【速读】： 该论文旨在解决多语言场景下生成式 AI (Generative AI) 中提示压缩（prompt compression）技术在非英语语言中效率显著下降的问题，即压缩是否能够缓解甚至扩大不同语言间固有的令牌开销溢价（token premium）。研究发现，尽管非英语语言本身已存在1.3–1.8倍的令牌成本溢价，但当前主流的提示压缩方法——尤其是基于英语监督训练的压缩器（如LLMLingua-2 XLM-R/mBERT、Kompress-v2）——在非英语语种上表现严重退化，导致压缩后的上下文利用率远低于英语，尤其在低保留率（keep-rate）下，立陶宛语仅保留10–24%、中文几乎无法保留有效信息。其关键发现在于：压缩性能的差异并非由模型架构决定，而是强烈依赖于压缩器的训练数据语言分布；仅当压缩器在多语言数据上进行训练（如XProvence v1）时，才能避免该语言转移差距。此外，研究揭示在长上下文任务中，激进的有监督压缩策略反而使部分非英语语言的压缩后内容效用低于无上下文输入，而“先翻译再压缩”（translate-then-compress）的流水线在三种语言中以约一半的令牌成本达到或超越原生压缩效果。因此，解决方案的关键在于采用多语言训练的压缩器，并结合翻译预处理策略以实现跨语言一致且高效的提示压缩，同时强调非英语环境下的安全压缩预算应显著低于英语。
链接: https://arxiv.org/abs/2608.26175
作者: Mantas Lukauskas
机构: Hostinger; Kaunas University of Technology (Kaunas大学科技)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 17 pages, 6 figures. Code and artifacts: this https URL
Abstract:Extractive prompt compression promises to cut LLM inference costs by removing low-information tokens, and learned compressors such as LLMLingua-2 report strong results on English benchmarks. Most other languages already pay a token premium: the same content costs 1.3-1.8x more tokens than in English. We ask whether compression closes or widens this gap. Using fully parallel data in ten languages spanning five scripts, with controls budget-matched in the target model’s tokenizer, we audit four learned compressors against four deterministic baselines, on eleven target models from ten vendors (over 250,000 evaluation calls). Three of the compressors are trained with English supervision (LLMLingua-2 XLM-R/mBERT; Kompress-v2 from the production Headroom stack); the fourth, XProvence, is trained multilingually. First, the transfer gap is real, replicates across target models and compressor backbones, and is strongly rate-dependent: at a 0.33 keep-rate English retains 57-62% of normalized context utilization while Lithuanian retains 10-24% and Chinese essentially none, despite Chinese having the smallest token premium. Second, the gap tracks compression supervision data, not architecture. All three English-trained compressors show it, deterministic methods show no comparable gap, and the multilingually trained XProvence v1 shows none. Its v2 release, retrained on translated data, empties 92% of Chinese contexts at its aggressive threshold without any warning. Third, in a harder long-context setting, aggressive learned compression drives compressed contexts to or below no-context utility in three of five non-English languages. A translate-then-compress pipeline matches or beats native compression at roughly half the token cost in three of five tested languages. We release all code, compressions, and model outputs. Safe compression budgets are much smaller outside English.
[NLP-105] Hallucinations in LLM s: A Lifecycle-Based Survey of Causes Detection Mitigation and Prevention
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）中幻觉（hallucination）问题在高风险场景（如医疗、法律和科研）下的可控性与可靠性挑战。现有研究多聚焦于幻觉的检测或缓解，而本文提出一种基于生命周期的系统性框架，将幻觉划分为数据相关、训练相关和推理相关的三类，与LLM开发流程相匹配。其解决方案的关键在于构建一个贯穿模型全生命周期的结构化分析体系，涵盖幻觉成因、检测方法及针对性缓解策略，并通过多维度评估基准数据集的适用性，为研究人员和实践者提供标准化、可操作的诊断与治理路径，从而实现对幻觉的系统性识别、控制与预防，推动更安全、可靠的生成式人工智能（Generative AI）系统发展。
链接: https://arxiv.org/abs/2608.26168
作者: Naveen Lamba,Sanju Tiwari,Manas Gaur
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:The lifecycle of hallucination in LLMs is a concept that enables building solid frameworks on the control and reliability of LLMs in high-stakes environments, including health, legal, and scientific research. Although previous surveys have primarily focused on detection or mitigation, this survey provides a lifecycle-based overview of the hallucinations in the LLMs, their cause, detection, mitigation, and this http URL propose a three-fold categorization of hallucinations across the LLM lifecycle: data-related, training-related, and inference-related, which is consistent with the lifecycle of the development of the LLM. Each of these stages is discussed regarding the cause of hallucinations, their detection, and the ways they can be addressed under specific mitigation or prevention interventions. In addition, we discuss the available benchmark data using a number of parameters so as to establish their suitability in identifying, restricting and managing hallucinations. The survey provides researchers and practitioners with a standardized framework to understand, diagnose, and cure hallucinations in a systematic system to present actionable data to build safer and more reliable LLMs.
[NLP-106] Using Poly-Encoders for Computationally Efficient Automated Creativity Assessment
【速读】： 该论文旨在解决自动化创造力评估中存在的计算资源消耗大且实际准确性不足的问题。传统方法往往依赖于高成本的模型或人工标注，难以在实际场景中推广。其解决方案的关键在于采用Poly-Encoder架构，通过微调小型预训练BERT编码器实现高效且精准的创造力评估。该方法利用轻量级模型结构，在保持与微调大型语言模型（Large Language Models, LLMs）相当性能的同时，显著降低计算需求。实验结果表明，基于BERT族模型与Poly-Encoder的组合在科学创造性思维测试数据集上达到了与人类评分者高达r = 0.74（95%置信区间[0.73, 0.75]）的皮尔逊相关系数，验证了其有效性。该研究实现了高性能与计算效率之间的平衡，为在教育等场景中实现可扩展、低成本的自动化创造力评估提供了可行路径。
链接: https://arxiv.org/abs/2608.26165
作者: Sam Grouchnikov,Phillip Gregory,Jiho Noh
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted at AIED 2026. The final authenticated version is available online at this https URL
Abstract:Automated creativity assessment has been a long standing challenge, with traditional methods often being resource intensive or lacking practical accuracy. We introduce a novel approach by using Poly-Encoder for computationally efficient and accurate automated creativity assessment. We fine-tuned a Poly-Encoder on a public dataset from the Scientific Creative Thinking Test, comprised of approximately 18,000 human-rated question responses. Our method leverages small pre-trained BERT encoders, achieving performance comparable to fine-tuned Large Language Models while significantly reducing computational demands. Experiments with the BERT-family models and poly-code counts achieved Pearson correlations of up to r = 0.74, 95% CI [0.73, 0.75] with human raters, matching the performance of resource intensive LLMs. This study bridges the gap between high performance and computational efficiency, potentially enabling widespread implementation of automated creativity assessment on accessible consumer-grade hardware. With some limitations, our findings suggest that Poly-Encoders are a promising alternative to LLMs for practical, scalable creativity assessment in various contexts, especially educational.
[NLP-107] Mutual Debiasing via Dual-Seed Comparison for Probabilistic Sampling in Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在生成随机变量时存在的高保真概率采样难题，其核心问题是LLMs在生成过程中普遍存在系统性偏差，导致目标概率分布失真。现有方法通常依赖单一自生成的种子，易继承模型固有的偏差。本文提出双种子对比（Dual-Seed Comparison, DSC）机制作为解决方案，其关键在于利用两个独立生成的LLM种子，通过比较其字符级序值构建比特序列，并将该序列转换并归一化为伪均匀随机变量，最终通过逆累积分布函数（inverse CDF）映射至目标分布。该方法无需外部工具，具有透明性与可解释性，实证表明在96%的测试场景中显著优于现有方法；此外，基于DSC对比算子的任务适配变体在多选题生成和属性约束型文本到图像提示中进一步提升了分布控制能力。
链接: https://arxiv.org/abs/2608.26161
作者: Zihao Guo,Hongtao Lv,Chaoli Zhang,Laiguo Yin,Lei Liu,Yonghui Xu,Lizhen Cui
机构: Shandong University (山东大学); Zhejiang Normal University (浙江师范大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 28 pages, 4 figures, 13 tables
Abstract:Although Large Language Models (LLMs) demonstrate remarkable capabilities in reasoning and decision-making, high-fidelity probabilistic sampling remains a persistent challenge. When generating random variables, LLMs consistently exhibit systematic biases that warp the target probability distributions. Current approaches often rely on a single, self-generated seed, which inherits model-specific biases. To overcome this vulnerability, we introduce Dual-Seed Comparison (DSC), a transparent, tool-free protocol that utilizes two independent LLM-generated seeds to neutralize bias. DSC compares the character-level ordinal values of the two seeds to construct a bit sequence, converts and normalizes this sequence into a pseudo-uniform variate, and then maps the variate to the target distribution through the inverse cumulative distribution function (CDF). Empirical results show that DSC substantially outperforms existing methods across 96% of evaluated settings. Beyond direct sampling, task-adapted variants based on the DSC comparison operator improve distributional control in MCQ generation and attribute-constrained text-to-image prompting.
[NLP-108] Self-Generated Text Recognition: Quality Heuristics Cross-Task Transfer and Downstream Bias in LLM Evaluation
【速读】： 该论文旨在解决生成式 AI (Generative AI) 模型自我生成文本识别（Self-Generated Text Recognition, SGTR）能力对人工智能安全机制构成的潜在威胁问题。具体而言，当大语言模型（LLM）被用作评估者或监控者时，若其具备识别自身输出的能力，可能导致偏见判断或合谋行为，从而破坏系统的公正性与安全性。现有研究对当前模型是否具备显著的SGTR能力存在矛盾结论，本文通过系统分析实验设计中的关键变量（即“操作化”策略，operationalizations），揭示了导致结果分歧的核心原因。研究发现，模型在不同评估格式（成对对比与单独评估）、对话结构（候选文本置于用户标签或助手标签中）以及生成任务领域（如编程与摘要）下的识别准确率存在显著差异。同时，研究验证了“质量启发式”——即模型倾向于将高质量文本归因于自身——是主要混淆因素。此外，研究发现通过监督微调（SFT）提升某一种评估配置下的SGTR性能可泛化至其他配置，并且训练后的模型在AlpacaEval框架中作为评判者时更偏好自身输出。这些结果表明，尽管存在干扰因素，部分模型已具备实际的SGTR能力，且此类训练会引发更广泛的自我识别与自我偏好倾向。因此，论文强调应将SGTR纳入安全关键型AI系统的设计考量之中，并建立相应的监测机制。
链接: https://arxiv.org/abs/2608.26159
作者: Jesse St. Amand,Callum Canavan,Sohaib Imran,Joseph Hewson,Aaron Lutz,Shi Feng,Puria Radmard,Lennie Wells
机构: Anthropic (Anthropic); Google (谷歌); Meta (Meta); OpenAI (OpenAI); Alibaba (阿里巴巴); DeepSeek (DeepSeek); Moonshot (Moonshot); University of Cambridge (剑桥大学); George Washington University (乔治华盛顿大学); MARS (MARS); Geodesic Research (Geodesic Research)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 22 pages, 9 figures (4 main body, 5 appendix), 11 tables (1 main body, 10 appendix)
Abstract:Self-Generated Text Recognition (SGTR)–the ability of an LLM to identify its own outputs–poses risks to AI safeguards that rely on LLMs as evaluators or monitors. Specifically, an LLM may recognize outputs from other copies of the same model and make biased judgments or collude outright. Prior work has drawn conflicting conclusions about whether current models possess significant SGTR capabilities. We reconcile these findings by identifying key experimental design choices–which we term operationalizations–that drive divergent results. Evaluating 13-21 models across six operationalizations, we find that accuracy varies substantially with evaluation format (pairwise vs. individual assessments of text), conversation structure (presenting candidate text in user tags vs. assistant tags), and the domain of the task used to generate candidate text (e.g., coding vs. summarization). We corroborate previous observations that a quality heuristic–models attributing authorship to text they perceive as higher quality–is a dominant confound. We also find that improving a model’s SGTR performance via SFT in one evaluation configuration can generalize to others. Training for SGTR additionally causes models to prefer their own outputs when acting as a judge in the AlpacaEval framework. Finally, we discuss the implications of our evaluations for the safety of future AI systems: our work suggests that, despite confounds, some models possess practical SGTR capabilities, and that training a model for SGTR in one setting can affect its self-recognition and self-preference more generally. We conclude that SGTR should be monitored and considered in the design of safety-critical AI applications.
[NLP-109] VFA: Empowering Multilingual MLLM s via Vision-Free Adaptation
【速读】： 该论文旨在解决多模态大语言模型（Multimodal Large Language Models, MLLMs）在非英语语种上的性能瓶颈问题，其核心挑战在于高质量非英语图像-文本对数据稀缺且标注成本高昂，导致现有模型仍以英语为主导。传统方法通过直接在多语言文本上进行微调虽可提升语言能力，但易破坏视觉-语言对齐关系并引发灾难性遗忘。为此，论文提出无需依赖视觉信息的适配框架——视觉无关适配（Vision-Free Adaptation, VFA），其关键创新在于将多语言语言能力增强与视觉对齐过程解耦：首先在多语言文本数据上微调基础大语言模型（LLM），生成多语言任务向量（multilingual task vector）；随后将其与已具备视觉-语言对齐能力的多模态模型的任务向量进行组合融合，共享同一骨干网络。实验结果表明，VFA在五个多模态大模型、六个多语言多模态基准上均实现一致性能提升，同时保持原始模型的通用多模态与纯文本处理能力，并仅需不到2%的文本数据即可显著缩小与全量多模态训练模型的差距，展现出卓越的数据效率。
链接: https://arxiv.org/abs/2608.26155
作者: Yixia Li,Yaqing Shi,Zhiwen Ruan,Dongdong Zhang,Lingjie Jiang,Shaohan Huang,Yun Chen,Guanhua Chen,Furu Wei
机构: Southern University of Science and Technology; Microsoft Research Asia; Shanghai University of Finance and Economics; Peking University; MoE Key Laboratory of Interdisciplinary Research of Computation and Economics
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Multimodal large language models have advanced rapidly, yet most remain English-centric, as scaling multilingual multimodal instruction tuning is limited by the scarcity and high cost of high-quality non-English image-text supervision. Although multilingual text data is abundant, naive textual fine-tuning can disrupt vision-language alignment and induce catastrophic forgetting. We propose Vision-Free Adaptation (VFA), a framework that decouples multilingual language enhancement from visual alignment by composing complementary task vectors over a shared LLM backbone. Specifically, we fine-tune a base LLM on multilingual text data to derive a multilingual task vector, which is then merged with the vision-aligned task vector of an MLLM. Experiments on five MLLMs across six multilingual multimodal benchmarks show consistent improvements while preserving both general multimodal and text-only capabilities. Moreover, using less than 2% of the text data, VFA narrows the gap to the fully multimodal-trained model, demonstrating its data efficiency.
[NLP-110] Evaluating AI Generated Summaries for Cancer Patients
【速读】： 该论文旨在解决生成式AI（Generative AI）在癌症患者照护应用中生成医疗摘要时存在的准确性、临床相关性及安全性问题。其解决方案的关键在于构建一个双评估框架，结合领域专家（包括肿瘤科医生和面向患者的护理人员）的高质量人工评估与大语言模型作为评价者（LLM-as-a-judge）的自动化评估，从而系统识别摘要中的遗漏与微小错误，并基于反馈迭代优化提示工程（prompt design）、内容可追溯性（grounding）以及安全防护机制（safety guardrails），以提升生成摘要的临床可信度与实用性。
链接: https://arxiv.org/abs/2608.26154
作者: Muhammad Aurangzeb Ahmad,Kim Shyu,Leon Oliver,Fergus Sleight,Paul Landau
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) are increasingly being integrated into digital health platforms to generate summaries of complex medical data. Although these models can improve patient engagement and communication, these systems also raise concerns about accuracy, faithfulness, and safety in clinical contexts. In this study, we evaluate AI-generated summaries within a cancer patient care application using a dual assessment framework. Human domain experts, including oncology clinicians and patient-facing care staff, provided ground-truth evaluations of summary quality along dimensions of accuracy, clinical relevance, and readability. In parallel, we employed LLMs serving as evaluators (LLM-as-a-judge). Some limitations were identified in the generated summaries e.g., occasional omissions and minor inaccuracies. These were systematically analyzed and used to iteratively improve prompt design, grounding, and safety guardrails.
[NLP-111] owards Interpretable Depression Detection: Linking Acoustic Features to DSM-5 Indicators
【速读】： 该论文旨在解决抑郁症诊断过度依赖主观自评报告而可能遗漏真实行为特征的问题，提出了一种基于语音声学特征与DSM-5抑郁行为指标之间可解释关联的透明化链接框架（Linkage Framework）。其解决方案的关键在于通过显式映射语音声学特征（如音调变异性、停顿、语速）至临床行为指标，实现对心理运动改变和注意力困难等核心症状的可解释性量化分析，同时在通用硬件（commodity hardware）上实现本地化运行，保障用户隐私。初步实验在DAIC-WOZ数据集上的结果表明，语音特征与DSM-5指标间存在方向一致的关联，验证了该方法的设计合理性，为后续在纵向数据集上的验证及多模态融合扩展奠定了基础。
链接: https://arxiv.org/abs/2608.26148
作者: Jonas Länzlinger,Katharina O.E. Müller,Burkhard Stiller,Bruno Rodrigues
机构: 未知
类目: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注:
Abstract:Depression affects millions worldwide, yet diagnosis relies on subjective self-reports that may miss authentic behavior. This paper presents an approach linking speech acoustics to DSM-5 depressive-behavior indicators through a transparent Linkage Framework. Unlike black-box models, the framework explicitly maps acoustic features (pitch variability, pauses, speech tempo) to clinical indicators, enabling interpretable, indicator-level outputs. The system runs locally on commodity hardware (HW) to preserve privacy. Preliminary evaluation on DAIC-WOZ shows directionally consistent associations between acoustic features and DSM-5 indicators for psychomotor change and concentration difficulty, supporting the design rationale. Future work will validate on longitudinal datasets and extend multimodal integration while maintaining edge constraints.
[NLP-112] CARE: Causally-Aligned Reasoning Exploration for Medical Large Language Models ECCV2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在医学推理任务中因专家标注数据稀缺与成本高昂而导致的训练瓶颈问题，同时克服传统基于结果的强化学习方法在医学场景下存在的自回归信用分配失败与梯度方差爆炸等关键挑战。这些问题常导致模型陷入“正确答案、错误推理”（Right Answer, Wrong Reason）的陷阱，即模型通过捕捉数据集中的虚假相关性或捷径而非基于临床逻辑进行推理。为应对这一问题，本文提出了一种理论严谨的内在经验筛选框架——因果对齐推理探索（Causally-Aligned Reasoning Exploration, CARE），其核心在于构建高质量训练轨迹的双重严格条件：一是因果充分性（Causal Sufficiency），通过基于一致性验证的自校验机制模拟do-演算干预，有效消除梯度偏差；二是邻近可学性（Proximal Learnability），利用动态熵边界筛选处于模型“最近发展区”内的经验样本，实现方差受限的优化。在此基础上，CARE采用双流目标函数，结合策略组相对探索与难度加权的经验回放机制，对筛选后的高质量经验进行优化。大量在多模态及纯文本医学基准上的实验表明，CARE显著优于现有先进方法，在大幅减少“正确但不一致”的推理现象的同时，提升了训练稳定性与模型可解释性。
链接: https://arxiv.org/abs/2608.26147
作者: Yucheng Zhou,Peng Luo,Qianning Wang,Chengzhong Xu,Jianbing Shen
机构: University of Macau (澳门大学); Auckland University of Technology (奥克兰理工大学)
类目: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: ECCV 2026
Abstract:Large Language Models (LLMs) have shown strong potential for medical reasoning, yet the scarcity and cost of expert-annotated data constrain their progress. While reinforcement learning offers a scalable alternative, standard outcome-based methods in medicine often suffer from autoregressive credit assignment failure and gradient variance explosion. This leads to the “Right Answer, Wrong Reason” trap, where models inadvertently reinforce spurious correlations and dataset shortcuts rather than valid clinical deduction. In this work, we propose Causally-Aligned Reasoning Exploration (CARE), a theoretically grounded framework for intrinsic experience curation. CARE is built upon two rigorous conditions for high-quality training trajectories: Causal Sufficiency, which utilizes an agreement-based self-verification mechanism to mimic do -calculus interventions and effectively debias gradients; and Proximal Learnability, which employs dynamic entropy bounds to select experiences within the model’s zone of proximal development for variance-bounded optimization. These rigorously filtered experiences are optimized via a dual-stream objective that combines on-policy group-relative exploration with difficulty-weighted experience replay. Extensive experiments on diverse medical multimodal and text-only benchmarks demonstrate that CARE consistently outperforms other strong competitors, substantially reducing correct-but-inconsistent reasoning and improving training stability.
[NLP-113] Vagdhenu: A Vrutta (Meter) Aware Shloka-to-Chant (TTS) System for Sanskrit
【速读】： 该论文旨在解决梵语偈颂（shloka）在文本转语音（TTS）系统中高保真吟诵（parayana recitation）生成的问题，核心挑战在于如何精确建模梵语独特的韵律（vrutta）、音系特征及声调规律。传统印度语言语音合成模型因采用天城文（Devanagari）书写系统而引发元音“schwa”删除等偏差，且难以区分梵语中如visarga sandhi的发音变体、alpaprana与mahaprana的送气对立以及齿音、卷舌音和腭音擦音的细微区别。为此，其解决方案的关键在于构建一个基于韵律感知的前端处理管道：首先通过卡纳达语正字法映射规避天城文导致的音素丢失；其次引入符合梵语音系学规则的精细化音素处理模块；最后设计一种基于半参考规则（half-reference rule）的韵律感知机制，能够自动识别诗句的格律（vrutta）并匹配最贴合的参考吟诵片段。研究发现，在自填充流匹配（flow-matching）架构中，文本侧的韵律调节器因梯度消失而失效，仅参考音频片段与语音引导重训练是有效的韵律控制手段。实验表明，该系统在五小时克隆数据集上达到接近4.6的专家平均意见得分（MOS），显著超越前代模型（StyleTTS2、VITS2、Matcha-TTS），并成功部署于包含5183句、约17.5小时视频内容的32章语料库及覆盖12部经典共约18000句的音频应用，成果已开源。
链接: https://arxiv.org/abs/2608.26146
作者: Prathosh A P
机构: Indian Institute of Science, Bengaluru (印度科学研究所，班加罗尔)
类目: Computation and Language (cs.CL)
备注:
Abstract:We present Vagdhenu, a vrutta (meter) aware shloka-to-chant system for Sanskrit: a text-to-speech system that maps a metrical verse to its chanted parayana recitation at high fidelity. This is an experience report, not a new architecture. We take an off-the-shelf flow-matching TTS backbone and a large-scale neural vocoder, and add the components a faithful Sanskrit chant pipeline needs: a frontend that routes Sanskrit through Kannada orthography to avoid the Hindi-style schwa deletion that Devanagari triggers in Indic models; a frontend that obeys subtle Sanskrit phonology (visarga sandhi with its jihvamuliya and upadhmaniya allophones, the aspiration contrast of alpaprana and mahaprana, and the dental, retroflex, and palatal sibilants kept distinct); and a vrutta-aware mechanism that detects the meter and picks an exactly matched reference under a half-reference rule. We report a negative result that shaped the system: in a self-infilling flow-matching backbone, a text-side prosody conditioner is architecturally inert, because the model recovers pitch from the context mel and the embedding gets no gradient; the reference clip and a voice-steering retrain are the only working prosody levers. We also report a comparative lineage across four families (StyleTTS2, VITS2, Matcha-TTS, and the flow-matching backbone), where each earlier family hit a ceiling on conjuncts or prosody that a five-hour clone cleared at an expert MOS near 4.6. The system shipped two deployments: a 32-chapter, 5183-verse video corpus (about 17.5 hours) and an audio app covering about 18000 verses across 12 books. We release the frontend, inference and training code, weights, a single-speaker chant dataset, and an interactive demo. Subjects: Computation and Language (cs.CL) Cite as: arXiv:2608.26146 [cs.CL] (or arXiv:2608.26146v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2608.26146 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Prathosh A. P. [view email] [v1] Sun, 28 Jun 2026 20:29:16 UTC (20 KB) Full-text links: Access Paper: View a PDF of the paper titled Vagdhenu: A Vrutta (Meter) Aware Shloka-to-Chant (TTS) System for Sanskrit, by Prathosh A PView PDFHTML (experimental)TeX Source view license Current browse context: cs.CL prev | next new | recent | 2026-08 Change to browse by: cs References Citations NASA ADSGoogle Scholar Semantic Scholar export BibTeX citation Loading… BibTeX formatted citation loading… Data provided by: Bookmark checked="checked"class=“labs-tab-input”> Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv’s community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) mathjaxToggle(); We gratefully acknowledge support from our major funders, member institutions, , and all contributors. About Help Contact Subscribe Copyright Privacy Accessibility Operational Status (opens in new tab) Major funding support from
[NLP-114] Why Current XAI Is Not Enough for Arabic NLP: A Critical Survey of the Explainability Gap
【速读】： 该论文旨在解决阿拉伯语自然语言处理（Arabic NLP）领域在可解释人工智能（XAI）方面的三大核心问题：方法论鸿沟、任务覆盖不均以及语言特性忽视。具体而言，现有阿拉伯语XAI研究高度依赖少数后验解释技术（如LIME、SHAP、注意力可视化和显著性分析），缺乏诊断性、反事实推理、探针测试、基于理由的解释及以人为中心等多样化方法；任务范围集中于分类任务（如情感分析、仇恨/攻击性语言检测、虚假新闻识别与垃圾信息过滤），而对生成、检索、翻译、摘要、结构化预测及对话等任务的解释支持不足；更重要的是，多数解释仅关注影响决策的词元（token），却未能充分揭示阿拉伯语特有的语言现象，如形态学复杂性、附着词（clitics）、方言变异、标准语与口语并存（diglossia）、拼写歧义、符号标记（diacritics）、代码切换、命名实体、文化指涉以及古典与宗教语体等。论文提出一个涵盖任务、方法、语言单位、语言变体、解释目标与评估实践的系统性分类框架，并主张构建以语言学为基础的阿拉伯语可解释AI研究范式，强调解释必须忠实于阿拉伯语作为语言、文化与社会技术复合体的本质特征。其解决方案的关键在于建立一种融合语言学深度与可解释性需求的、面向多模态与跨任务的系统性研究框架，推动阿拉伯语XAI从“通用解释”迈向“语言学可信解释”。
链接: https://arxiv.org/abs/2608.26144
作者: Salima Lamsiyah,Ruslan Mitkov
机构: University of Luxembourg, Luxembourg; University of Alicante, Spain
类目: Computation and Language (cs.CL)
备注: 11 pages
Abstract:Explainable AI (XAI) is now a major theme in NLP; however, Arabic NLP remains under-explained in three connected senses. First, there is a method gap: Arabic XAI relies heavily on a small set of post-hoc techniques such as LIME, SHAP, attention visualization, and saliency, while broader NLP XAI offers richer diagnostic, counterfactual, probing, rationale-based, and human-centered methods. Second, there is a task gap: existing Arabic XAI work is concentrated in classification tasks, especially sentiment analysis, hate/offensive language detection, fake news, and spam, with weaker coverage of generation, retrieval, translation, summarization, structured prediction, and dialogue. Third, there is a linguistic gap: many explanations identify influential tokens, but rarely explain Arabic-specific phenomena such as morphology, clitics, dialectal variation, diglossia, orthographic ambiguity, diacritics, code-switching, named entities, cultural references, or Classical and religious registers. This critical structured survey synthesizes the reviewed literature on Arabic XAI across text, speech, and multimodal settings. We argue that Arabic NLP does not only need explanations of model decisions; it needs explanations that are faithful to Arabic as a linguistic, cultural, and sociotechnical object. We introduce a taxonomy of tasks, methods, linguistic units, varieties, goals, and evaluation practices, and propose a research agenda for linguistically grounded Arabic XAI.
[NLP-115] Beyond Accuracy: A Qualitative Analysis of Vision-Language Models for Hate Speech Detection in Memes
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在识别具有复杂语境依赖性的仇恨表情包（hateful memes）时面临的挑战。尽管当前先进的视觉-语言模型（Vision-Language Models, VLMs）在多模态任务中表现优异，但其普遍忽视语境、反讽等细微线索，导致对仇恨内容的误判或漏判。本研究的关键解决方案在于通过定性分析四种前沿VLM模型（LLaVA-7B、Qwen-VL、GPT-4o mini与Claude 3 Haiku），在零样本（zero-shot）和少样本（few-shot）提示设置下评估其输出的合理性与推理过程，重点考察上下文框架如何影响模型判断，并深入剖析模型生成的解释性理由，从而揭示其在处理仇恨表情包时的思维逻辑与内在局限，为提升模型对隐含恶意内容的理解能力提供关键洞见。
链接: https://arxiv.org/abs/2608.26143
作者: Muhammad Jawad Chowdhury,Adiba Hasan,Ishrak Hossain,Shahriar Ivan,Sabbir Ahmed
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Memes have turned out to be a powerful tool through which individuals share their ideas concerning contemporary social and political problems. Their anonymity, as well as their ability to go viral, make them a powerful medium for spreading hate. It remains very difficult to identify such complex and context-dependent hate speech. Although they display excellent performance on multimodal tasks, vision-language models (VLMs) tend to ignore context, irony, and other subtle cues that play a key role in identifying hateful memes. In this work, we present a qualitative analysis of four state-of-the-art VLMs: LLaVA-7B, Qwen-VL, GPT-4o mini, and Claude 3 Haiku. We evaluate these models under zero-shot and few-shot prompting to examine how contextual framing influences their outputs. Our analysis goes beyond simple classification accuracy and focuses on a qualitative evaluation of the models’ generated justifications, providing a more in-depth understanding of their thought processes and constraints when dealing with hateful memes.
[NLP-116] Position Is All You Need: A Free Lunch Token Compression Strategy for MLLM -based Referring Expression Segmentation ICML2026
【速读】： 该论文旨在解决基于多模态大语言模型（Multimodal Large Language Models, MLLMs）的指代表达分割（Referring Expression Segmentation, RES）任务中，由于令牌压缩（token compression）方法导致性能显著下降的问题。尽管现有研究在提升RES性能方面取得了进展，但其高昂的计算开销仍是主要瓶颈，而这一问题尚未得到充分探索。本文通过系统评估典型令牌压缩方法发现，传统压缩策略在RES任务中引发显著性能退化，根源在于其忽略了视觉令牌的位置嵌入与局部空间结构信息的重要性。研究进一步揭示，在RES任务中，保持原始位置嵌入和局部邻域空间结构是令牌压缩的关键需求，远超其他任务对位置信息的依赖程度。基于此洞察，作者提出PAYN——一种即插即用、无需训练的令牌压缩方法，仅依赖位置信息进行压缩：通过保留每个局部邻近区域中分布均匀的令牌，并严格维持原始位置索引，从而确保空间关系的一致性。在多个主流RES基准上的实验结果表明，PAYN在保持低计算成本的同时，显著优于现有压缩方法，验证了“位置即一切”（position is all you need）在基于MLLM的RES任务中的有效性。
链接: https://arxiv.org/abs/2608.26142
作者: Yuhan Liu,Yixiong Zou,Yuhua Li,Ruixuan Li
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted by ICML 2026
Abstract:Referring Expression Segmentation (RES) aims to generate pixel-wise segmentation masks from complex and implicit textual queries. While recent advances in Multimodal Large Language Models (MLLMs) have substantially boosted RES performance, their prohibitive computational overhead remains a critical bottleneck, which, however, is rarely explored. To fill this gap, we first evaluate typical token compression methods on this task and observe a surprising performance degradation. In this paper, we aim to understand this phenomenon for a solution. By extensive experiments, we find that token compression for RES requires preserving the original position embeddings and local neighboring spatial structures, indicating that visual token position information is far more critical than in other tasks. Building on this insight, we ask: Can we design the token compression method purely based on the position information? Therefore, we propose PAYN, a plug-and-play, training-free token compression method that relies solely on position information. PAYN retains tokens that are adequately distributed in every local neighboring region while strictly preserving original positional indices, thereby maintaining spatial relational consistency. Experiments on multiple RES benchmarks demonstrate that our method outperforms existing token compression methods, verifying that position is indeed all you need for token compression in the MLLM-based RES task. Codes are avaliable at this https URL.
[NLP-117] AdaThinking-E: One-Token Entropy Regulation for Adaptive Thinking
【速读】： 该论文旨在解决多模态大语言模型在文档推理任务中因对所有问题统一应用深度思考过程而导致的计算开销过大问题。当前方法缺乏对问题复杂度的动态感知能力，使得简单任务也经历冗余推理，不仅降低响应效率，还可能损害基准测试中的准确性。其解决方案的关键在于提出AdaThinking-E——一种基于强化学习的自适应思考框架，通过单个标记（token）的熵值调控实现思维决策的智能化。核心思想是利用关键决策标记处预测概率分布的熵来量化模型对是否启动思考的置信度，进而构建熵驱动的奖励机制：训练过程中模型从高熵探索阶段逐步收敛至低熵、具有确定性和泛化能力的决策策略。该机制使模型能够内生地识别何时需要进行深度推理，无需人工标注难度或外部干预。大量实验表明，该方法在保持复杂任务高精度的同时，显著提升了简单任务的处理效率，实现了性能与效率的双重优化。
链接: https://arxiv.org/abs/2608.26141
作者: Zining Wang,Tongkun Guan,Boming Chen,Zhentao Guo,Jianqiang Liu,Chao Jin,Chen Duan,Kai Zhou,Pengfei Yan,Wei Shen,Xiaokang Yang
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Multimodal large language models have demonstrated strong document reasoning capabilities by incorporating explicit thinking processes. While this capability significantly improves performance on challenging tasks, current models apply such deep reasoning uniformly to all questions, resulting in unnecessary computational overhead for simple task. This not only degrades user experience but also negatively impact accuracy on benchmark datasets. We identify the critical need for adaptive thinking mechanisms that can intelligently determine when to engage reasoning based on question complexity. To address this, we propose AdaThinking-E, a novel reinforcement learning framework that learns adaptive thinking through one-token entropy regulation. Our key insight is that model confidence in the decision to engage thinking (or not) can be quantified through entropy analysis of the predicted probability distribution at critical decision tokens. This observation motivates our entropy-governed reward mechanism: the training process naturally transitions from high-entropy exploration, where the model experiments with different thinking strategies, to low-entropy convergence with confident, generalizable decision-making policies. Crucially, this approach enables models to intrinsically discover when to think without requiring manual intervention or external difficulty labels. Extensive experiments demonstrate that our approach enables models to be both accurate on complex problems and efficient on simple ones across diverse document tasks.
[NLP-118] Affix Cache for Diffusion Large Language Models
【速读】： 该论文旨在解决扩散型大语言模型（Diffusion Large Language Models, DLLMs）在高效推理过程中面临的缓存复用难题。与自回归模型可重用共享前缀的键值（Key-Value, KV）缓存不同，DLLMs由于采用双向注意力机制，导致共享上下文中的标记（token）与生成过程中的动态标记之间存在耦合，使得直接复用整个上下文的KV缓存会引入错误，而完全重新计算又代价高昂。为此，论文提出ACache——一种面向前缀和后缀（affix）的缓存复用机制，其核心在于通过量化生成掩码标记对上下文的影响程度，识别出少量关键的“锚点标记”（Anchor Tokens），仅对这些关键标记进行选择性重计算，其余非关键部分则保留原有缓存。该方法基于Fast-dLLM实现，在仅重计算约20%的前缀/后缀标记情况下，有效恢复了因直接缓存复用造成的精度损失。此外，基于Nano-vLLM构建的共享前缀原型系统表明，ACache可将重计算延迟降低高达55.7%，并使端到端吞吐量提升至1.68倍。因此，解决方案的关键在于：通过精准识别影响生成结果的少数关键标记，实现高精度与高效率之间的平衡。
链接: https://arxiv.org/abs/2608.26140
作者: Kaihua Liang,An Zhong,Xin Tan,Zafar Ayyub Qazi,Hong Xu,Jian Weng,Marco Canini
机构: KAUST(沙特阿美大学); CUHK(香港中文大学); LUMS(巴基斯坦拉合尔管理科学大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 15 pages, 7 figures
Abstract:Diffusion Large Language Models (DLLMs) enable non-autoregressive decoding and bidirectional context modeling, but efficient inference remains challenging. Unlike autoregressive systems, whose key-value (KV) cache can be reused for shared prefixes, DLLMs couple the KV states of shared context tokens with evolving generated tokens through bidirectional attention, making naive cache reuse stale while full recomputation is expensive. We present ACache, an affix-oriented cache reuse mechanism for shared text spans in DLLMs beyond prefixes. ACache identifies a small request-specific subset of critical affix tokens, called Anchor Tokens, by measuring their influence on masked generation tokens, and selectively recomputes the KV states of only these tokens while reusing the remaining affix cache. Built on Fast-dLLM, ACache recovers the accuracy loss caused by direct affix-cache reuse across different settings when recomputing around 20% of affix tokens. We also build a shared-prefix prototype on top of the Nano-vLLM engine, showing that ACache reduces recompute latency by up to 55.7% and improves end-to-end throughput by up to 1.68 \times .
[NLP-119] Syntax vs. Semantics: How Transformers Learn Deep Dependencies
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在训练过程中对深层语义依赖关系（deep semantic dependencies）学习机制不明确的问题，尤其关注其优化动态中为何结构化推理能力呈现突发式涌现而非渐进式发展。其核心解决方案在于提出一个机理框架，将模型学习过程建模为“表层统计”（Surface Statistics）与“深层语义”（Deep Semantics）之间的竞争机制。关键发现是“梯度饥饿”（Gradient Starvation）现象：在早期优化阶段，稀疏的深层语义依赖所对应的误差信号被主动抑制，导致结构化推理难以有效学习，并最终表现为一种突变式的相变现象。该框架进一步揭示了思维链（Chain-of-Thought, CoT）策略的有效性根源——通过将中间推理步骤显式外化为具体标记（tokens），CoT可绕过隐式推理中固有的梯度抑制机制。研究在从简化Transformer到实际部署模型（如Llama-3.1-8B、Qwen2.5-Coder-7B）的多尺度实验中验证了该理论。基于此理论，作者设计了一种拓扑对齐的对比目标函数，显式修正梯度几何结构，在变量绑定任务上相较标准交叉熵微调实现了超过2倍的性能提升。
链接: https://arxiv.org/abs/2608.26139
作者: Jiangrui Zhao,Xiaoting Du
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models demonstrate remarkable syntactic fluency, yet the optimization dynamics governing their acquisition of deep semantic dependencies remain poorly understood. We propose a mechanistic framework that models this learning process as a competition between Surface Statistics and Deep Semantics. Our theoretical analysis identifies a ``Gradient Starvation" phenomenon where the error signals for sparse semantic dependencies are actively suppressed during early optimization. This suppression impedes the learning of structural reasoning and causes its emergence to manifest as a sudden phase transition. Furthermore, this framework offers a mechanistic basis for the effectiveness of Chain-of-Thought (CoT) strategies. By externalizing intermediate reasoning steps into concrete tokens, CoT effectively bypasses the suppression regime inherent to implicit reasoning. We validate these findings across scales ranging from toy transformers to production models (Llama-3.1-8B, Qwen2.5-Coder-7B). Finally, guided by this theory, we propose a topology-aligned contrastive objective that explicitly rectifies the gradient geometry. Experiments on variable binding tasks demonstrate that our method achieves an improvement that is over 2x larger than that obtained via standard cross-entropy fine-tuning.
[NLP-120] Cross-Platform Generalisation Failure in Mental Health Natural Language Processing: A Five-Axis Fairness Audit of Transformer Models on Social Media
【速读】： 该论文旨在解决生成式AI在跨平台心理状态识别任务中存在显著性能退化与公平性失效的问题，尤其关注模型在不同社交媒体平台（如Reddit、Twitter）上部署时的泛化能力与公平性保障。其核心挑战在于：尽管模型在源平台（如Kaggle数据集）上表现出高精度（AUC 0.983–0.987），但在目标平台上的判别性能（AUC下降30.3%–39.5%）、校准能力（ECE从0.056–0.060升至0.499–0.542）、预测公平性（原始差异指数DI高达0.17，调整后达0.11–0.29）及特征归因稳定性（前10个关键词的Jaccard相似度为0）均出现严重退化。解决方案的关键在于提出“跨平台公平性评估框架”（Cross-Platform Fairness Evaluation, CPFE），涵盖判别性能、校准性、统计显著性、预测公平性与归因稳定性五个维度，系统揭示了模型在异构平台环境下的多维失败模式，并通过平台特异性温度缩放验证了校准失败与判别性能分离的可解耦性；进一步实验表明，使用目标平台标注数据进行微调比仅用于校准更有效，凸显了目标域标签作为训练信号的重要性。该研究主张将CPFE五轴审计作为心理健康自然语言处理系统在异构环境中部署的标准化评估流程。
链接: https://arxiv.org/abs/2608.26138
作者: Rajveer Singh Pall,Sameer Yadav
机构: Gyan Ganga Institute of Technology and Sciences(甘加工程技术与科学学院); Department of Computer Science and Business Systems(计算机科学与商业系统系)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 15 pages, 24 figures, 10 tables. Submitted to IEEE Transactions on Neural Networks and Learning Systems
Abstract:We introduce the Cross-Platform Fairness Evaluation (CPFE) framework – a five-axis audit protocol covering discriminative performance, calibration, statistical significance, prediction equity, and attribution stability – and apply it to four transformer models (BERT, RoBERTa, Emotion-DistilRoBERTa, GoEmotions-RoBERTa) trained on a Kaggle mental health corpus (n=35,556) and evaluated on Reddit (n=6,257) and Twitter (n=2,883) test sets with emotion labels mapped to clinical proxies. All three independently evaluated models exhibit consistent and substantial cross-platform AUC degradation (30.3-35.4% on Reddit, 37.9-39.5% on Twitter) relative to within-platform performance (AUC 0.983-0.987), confirmed across five independent training seeds. Calibration failure is concurrent and severe: ECE rises from 0.056-0.060 in-domain to 0.196-0.229 on Reddit and 0.499-0.542 on Twitter. Platform-specific temperature scaling reduces mean ECE by 88.0% without altering discriminative performance (mean |delta AUC|0.01), confirming separable failure modes. Prediction equity analysis reveals large cross-platform disparities (raw DI 0.17; prior-shift-adjusted DI: 0.11-0.29 on Reddit), with equalized odds differences of 0.753-0.830 for mental health proxy classes on Reddit and 0.755-0.831 for anxiety on Twitter. Attribution stability analysis shows near-complete vocabulary divergence across platforms (Jaccard J=0 in 14/16 model-class pairs at K=10). These findings support treating cross-platform validation across all five CPFE axes as a standard requirement for mental health NLP systems in heterogeneous environments. In a single-seed fine-tuning experiment, mean AUC improved by 0.216, suggesting target-platform labels provide greater benefit as training signal than as calibration signal.
[NLP-121] Interpretable Fairly Evaluated Automated L2 Speaking Assessment that Beats the Single-Human Ceiling and Why Pause Encoding Does Not Change LLM Fluency Scores
【速读】： 该论文旨在解决第二语言（L2）英语学习者在缺乏真实对话伙伴的情况下难以进行有效口语练习，且口语表达常伴随高度焦虑情绪的问题。当前自动化口语训练与评分系统虽快速发展，但其评分结果的可信度受限于准确性、可解释性、公平性及与人类评判标准的基准一致性。为此，研究提出一种基于可解释特征与大语言模型（LLM）融合的混合框架，用于自发性L2对话的自动评分。其核心创新在于：采用确定性的De-Jong语速时序复合指标作为连续性特征基础，结合单一文本-LLM提供的粗粒度流利度判断，实现对130段可用音频的L2演讲评估。该混合模型在与共识金标准（consensus gold）对比中达到Spearman等级相关系数ρ=0.818，优于81%的80名专业评分员，接近最优评分员水平，并达到可靠性校正后最大一致性（κ_max=0.99）的约83%。关键发现表明，流利度信号主要来源于实际测量的语速时序特征，而非提示中暂停位置的编码方式；对停顿的书写方式（如内联标注或中句语义准则）进行控制实验后，均未显著提升相关性（Δρ < ±0.1），证实了模型性能依赖于客观语音特征而非提示工程。所有结论均通过双独立学习者隔离方法、配对自举置信区间、独白负控实验、经典测量指标复现及按母语背景（L1）的公平性审计等多重验证手段支持，确保了结果的稳健性与可解释性。
链接: https://arxiv.org/abs/2608.26137
作者: Eichi Uehara
机构: Aflo Technologies Inc.(Aflo科技公司)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG); Sound (cs.SD)
备注: 17 pages, 3 figures, 5 tables
Abstract:Second-language (L2) English learners can rarely rehearse speaking with a partner. Speaking is also the most anxiety-laden skill. These gaps drive a fast-growing market for automated speaking practice and scoring. But an automated score is trustworthy only if it is accurate, interpretable, fair, and benchmarked against the right human bar. We build an interpretable feature-plus-LLM hybrid for spontaneous L2 dialogue. We evaluate it without ever fitting to the human labels, against the ICNALE Global Rating Archive: 140 speeches rated by ~80 trained raters on 10 analytic criteria. We score the 130 L2 speeches with usable audio. A deterministic De-Jong speech-timing composite reaches rho=0.764. Blended with a single text-LLM fluency judgment, it reaches Spearman rho=0.818 against the consensus gold. This agrees with the consensus better than 81% of the 80 individual trained raters: above the median rater (rho=0.73) and near the best, and at ~83% of the reliability-corrected maximum (kappa_max=0.99). The blend improves on the composite alone by +0.054 (paired-bootstrap 95% CI [0.017, 0.108], excludes 0); the LLM adds a coarse fluency ranking that the continuous composite refines. We also report a controlled null on pause encoding, bounded to effects below about +/-0.1 rho at this sample size. Holding the LLM and learner words fixed and varying only how pauses are written into the prompt, inline pause locations do not beat aggregate pause statistics (-0.069, CI [-0.15, +0.08]), and a grounded mid-clause criterion gives no reliable gain. The fluency signal comes from the measured speech-timing features, not from how pauses are written for the LLM. We back every claim with two agreeing learner-isolation methods, paired-bootstrap CIs, a monologue negative control, per-feature reproduction of classical measurements, and a per-L1 fairness audit.
[NLP-122] Reward-Informed Sparse Autoencoders and the Solution-Completeness Confound
【速读】： 该论文旨在解决如何利用奖励信号（reward signal）提升语言模型激活空间中稀疏自编码器（Sparse Autoencoders, SAE）的可解释性，特别是使其能够捕捉与推理质量相关的关键特征。其核心问题是：现有基于强化学习（Reinforcement Learning, RL）生成的高奖励轨迹是否能通过SAE揭示出真正反映推理能力的本质特征，而非仅体现任务完成形式上的差异。解决方案的关键在于构建一种受奖励信息指导的稀疏自编码器（Reward-Informed SAE, RI-SAE），即通过将GRPO（Generalized Reward Policy Optimization）生成的推理延续分为高奖励（“好”）与低奖励（“坏”）两类，训练标准的JumpReLU SAE以学习这些激活的稀疏分解，并分析所得特征在区分两类样本时的表现。研究发现，在Llama-3.1-8B模型中，仅有少数（16,384个特征中的稀疏子集）特征实现了显著的类间分离（轮廓系数0.79），但对照实验表明这种分离主要反映的是解题完成度（solution completeness），而非真正的推理质量——例如，仅依靠文本长度、封闭式推理结构和答案框定等结构性线索即可达到AUC 0.70，且99%的“好”完成结果被框定，而“坏”完成中仅有69%被框定。相比之下，未接触过奖励信号的通用SAE无法实现任何有效分离（轮廓系数0.01）。因此，该方法虽提供了一种低成本、无需标签的利用RL信号进行可解释性分析的途径，但其揭示的大部分内容实为输出形式而非深层推理机制；仅少数两个特征（符号数学、程序化与评价性语言）具有可读性，可作为推理过程的示意性表现，而非独立的推理能力表征。
链接: https://arxiv.org/abs/2608.26136
作者: Tanvi Nagilla,Alexander Jameson,Daniel Manta,Shayaan Uddin
机构: 未知
类目: Computation and Language (cs.CL)
备注: 6 pages, NEMI
Abstract:Sparse autoencoders (SAEs) decompose language-model activations into sparse, interpretable features, and an appealing way to aim them at reasoning is to curate their data with a signal reinforcement learning already produces: the reward. We build such a reward-informed SAE (RI-SAE): we split GRPO trajectories into high-reward (“good”) and low-reward (“bad”) reasoning continuations, train a standard JumpReLU SAE on their activations, and then ask what the resulting good/bad separation actually measures. On Llama-3.1-8B a sparse subset of the 16,384 features does separate the classes (silhouette 0.79 on the selected features versus 0.005 for the full code), but a control battery shows the separation is largely solution completeness rather than reasoning quality: a TF-IDF text classifier already splits the classes (AUC 0.75–0.83), and three structural cues alone (length, a closed reasoning block, and a boxed answer) reach AUC 0.70 (99% of good versus 69% of bad completions are boxed). A generic SAE that never saw the reward does not separate the classes at all (silhouette 0.01, no discriminative features), so the 0.79 is in-sample fitting of this curated signal rather than structure that a reward-blind dictionary recovers. We therefore present the recipe and its control battery together: reward filtering is a cheap, label-free way to reuse RL signals for interpretability, but most of what it surfaces is completion form. Two discriminative features are still readable (symbolic mathematics; procedural and evaluative language), which we take as illustrative rather than as isolated reasoning.
[NLP-123] Data Science Approaches to Evaluating Honours Candidates
【速读】： 该论文旨在解决从碎片化、非结构化的开源情报（OSINT）中准确估算公众对特定个体情感态度的难题，尤其在高风险决策支持场景下实现透明、可复现且可审计的情感评估。其核心挑战在于如何将异构的网络信息转化为可靠、可解释的个人层面情感分布，并区分正面、中性与负面情感信号。解决方案的关键在于构建一个模块化自然语言处理（NLP）数据科学流水线，系统整合网络搜索、文本提取、相关性过滤、分词、共指消解及情感分析等步骤，形成端到端的自动化处理流程。尤为关键的是引入了MINOS——一种基于领域知识的情感分析算法，专门用于识别与声誉风险、不当行为及积极公共贡献相关的语言模式，相较于AFINN和VADER，MINOS在已知声誉结果的公众人物案例中展现出更清晰的情感类别分离能力。该方法通过结合人类监督（human-in-the-loop）机制，实现了高可信度、可追溯的情感评估，已在英国荣誉体系（UK Honours system）的应用中验证其有效性，为要求具备高标准公共行为的荣誉授予提供量化支持。
链接: https://arxiv.org/abs/2608.26135
作者: Francesca von Braun-Bates,Sunreeta Sen,Indraayudh Talukdar,Anirban Lahiri
机构: Ministry of Justice(英国司法部); Joint Counter-Terrorism Prisons and Probation Hub(联合反恐监狱与假释中心); Arndit Ltd.(阿尼迪特有限公司); Indian Institute of Technology Delhi(印度理工学院德里分校); Kainos(凯诺斯)
类目: Computation and Language (cs.CL)
备注: 13 pages, 6 figures, corrects typographical errors from published version and includes full-colour figures
Abstract:We present a modular data-science pipeline for estimating public sentiment towards individuals from fragmented, unstructured open-source intelligence (OSINT). The method chains web search, text extraction, relevance filtering, tokenisation, co-reference resolution, and sentiment analysis to convert heterogeneous web material into auditable person-level sentiment distributions. We compare AFINN and VADER with MINOS, a domain-informed sentiment algorithm designed to detect language associated with reputational risk, misconduct, and positive public contribution. Applied to public figures with known reputational outcomes, MINOS gives the clearest separation between positive, ambiguous, and negative cases. The results show that chained NLP and OSINT methods can support transparent, reproducible, human-in-the-loop sentiment assessment for high-stakes decision support. We demonstrate the approach on the UK Honours system, where individuals are required to display high standards of public conduct to maintain an Honour.
[NLP-124] Agent Seer: Synthesizing Scenarios from Specification Understanding
【速读】： 该论文旨在解决生成式 AI (Generative AI) 代理在使用外部工具时，缺乏可扩展、动态且真实可信的评估场景的问题。传统手动构建测试场景的方法依赖深厚的领域知识，难以跨工具生态系统扩展，并且生成的基准静态化，无法追踪不断演进的 API 变化。其核心解决方案在于利用工具规范（如函数名、自然语言描述和类型化参数模式）中蕴含的语义信息，无需人工标注、无需实时工具调用或领域特定调优，即可自动生成高质量的多轮对话评估场景。关键创新点在于提出 Agent Seer 框架，通过增强原始参数模式、合成带分级复杂度的工具输出，并将其转化为基于模拟数据的多轮对话，从而实现高精度的工具调用准确性和对话连贯性。实验表明，该方法在涵盖不同领域与工具集规模的七组 Model Context Protocol (MCP) 规范上均表现优异，尤其在中小型工具集上实现完全工具覆盖；分析揭示参数模式复杂度是影响评估质量的最强预测因子，而参数值准确性不足是主要失败模式，这一细粒度问题在仅依赖名称匹配的粗粒度指标下难以察觉。
链接: https://arxiv.org/abs/2608.26133
作者: Harish Karumuri,Mahesh Vemula,David Lopes Pegna
机构: Apple(苹果)
类目: Computation and Language (cs.CL)
备注:
Abstract:Evaluating AI agents that use external tools requires realistic test scenarios that capture how practitioners compose tools and iterate across conversation turns. Constructing such scenarios by hand demands deep domain expertise, does not scale across tool ecosystems, and produces static benchmarks that cannot track evolving APIs. We observe that tool specifications – function names, natural-language descriptions, and typed parameter schemas – already encode sufficient semantic information to synthesize realistic evaluation scenarios without manual curation or live tool execution. Agent Seer builds off this latent information: from a single Model Context Protocol (MCP) specification, with no examples, no live tool access, and no domain-specific tuning. This pipeline enriches raw schemas, generates graded scenarios with synthetic tool outputs, and expands them into mock-data-grounded multi-turn dialogues that exhibit strong tool-calling correctness and conversational coherence. Evaluation quality is measured by applying this pipeline on seven MCP specifications spanning diverse domains and tool-suite sizes and measuring the tool-calling correctness and conversational coherence. The pipeline achieves strong quality across all domains, with complete tool coverage on small and medium specifications. Two findings emerge within this analysis: parameter schema complexity is the strongest correlate of quality variation – tool-suite size plays a smaller, orthogonal role – and argument value accuracy is the dominant failure mode among imperfect scenarios, a sub-dimension invisible to coarse-grained name-match metrics. Subjects: Computation and Language (cs.CL) Cite as: arXiv:2608.26133 [cs.CL] (or arXiv:2608.26133v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2608.26133 Focus to learn more arXiv-issued DOI via DataCite
[NLP-125] SLM-Conditioned Hierarchical Relation Routing for Labeled Property Graph Learning
【速读】： 该论文旨在解决传统图神经网络（Graph Neural Network, GNN）在处理带标签的属性图时，因将节点和边的异质性文本与类别属性简单编码为静态特征向量而导致语义证据选择能力受限的问题。其核心挑战在于难以动态识别并筛选对特定预测目标具有相关性的上下文信息，从而影响消息传递的准确性与可解释性。解决方案的关键是提出一种基于小语言模型（Small Language Model, SLM）条件化的分层关系路由机制（SLM-Conditioned Hierarchical Relation Routing）：通过引入一个参数高效的SLM，将图结构的软令牌（soft tokens）作为输入，生成针对目标节点的条件化路由查询；该查询首先在每种关系类型内部筛选相关消息，再在关系级摘要间进行信息路由，最终生成对拓扑GNN提供的结构表示的有界残差更新。该机制在保持原始结构证据稳定的同时，允许上下文语义信息动态调节预测结果，实现了对邻居及关系类型层面的可解释分析，并提供了一种通用框架以整合语言模型衍生的语义先验至富属性图学习中。
链接: https://arxiv.org/abs/2608.26132
作者: Michal Podstawski
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Labeled property graphs combine relational structure with heterogeneous textual and categorical properties attached to both nodes and relationships. Conventional graph neural networks typically represent these properties as static feature vectors, limiting their ability to determine which semantic evidence should influence message propagation for a particular prediction target. We propose SLM-Conditioned Hierarchical Relation Routing, an architecture that integrates a small language model directly into graph message selection. A topology GNN provides a stable structural representation and prediction anchor. For each target node, incident messages combine the neighbor’s structural state, node-property encoding, relationship-property encoding, and relationship type. A parameter-efficient SLM processes structured graph soft tokens and produces a target-conditioned routing query. This query first selects relevant messages within each relationship type and subsequently routes information across relation-level summaries. The resulting representation provides a bounded residual update to the topology anchor, preserving structural evidence while allowing contextual semantic information to modify the prediction. The architecture supports interpretable analysis at both the neighbor and relationship-type levels and provides a general mechanism for integrating language-derived semantics into property-rich graph learning.
[NLP-126] Evaluating Language Models in Realistic Conversational Contexts
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在开放域、多轮对话场景下，缺乏可靠且可扩展的人类尺度对话质量评估体系的问题。现有评估框架多基于摘要、翻译或短文本问答任务设计，难以有效衡量长周期、高复杂度对话中的连贯性与一致性，且其评价指标的构建往往依赖合成数据而非真实人类反馈。为填补这一空白，论文提出UPHELD（UPwork Human-Scale Evaluated Long Dialogues）基准，这是一个大规模、带参考答案的对话评估数据集，包含由专业剧本作者创作的数百个完整人机对话，涵盖超过3万条专家生成的对话回合，并配有3.6万余条每轮对话的人类标注。基于UPHELD，研究系统评估了传统自动评估指标及基于大模型作为评判者（LLM-as-a-judge）的无参考方法，发现这些方法与专家人类判断的相关性较低。针对此问题，研究进一步提出“多重评判者”（Mixture-of-Judges）框架，通过融合多种评估信号，使评估结果与人类判断的相关性提升约30%。该解决方案的关键在于构建一个以真实人类注释为基础、具备高生态效度的大规模对话评估基准，并在此基础上发展出更鲁棒、更贴近人类认知的多源评估机制，从而为人类尺度对话智能的评测提供了坚实、可信的基线支持。
链接: https://arxiv.org/abs/2608.26131
作者: Ilija Subasic,Andrew Rabinovich,Zhao Chen
机构: 未知
类目: Computation and Language (cs.CL)
备注: International Conference on Machine Learning 2026
Abstract:As Large Language Models (LLMs) are increasingly deployed to serve open-ended, multi-turn interactions, evaluating conversational quality at human scale has become a central challenge. Existing evaluation frameworks built for summarization, translation, or short-form QA tasks fall short of adequately measuring the consistency of human-scale dialogue, especially when derivation and validation of these metrics themselves often rely on synthetic rather than human sources. We fill the gap by introducing UPHELD (UPwork Human-Scale Evaluated Long Dialogues), a large, reference-full benchmark for evaluating human-scale conversational ability beyond factual correctness. UPHELD consists of hundreds of complete human-to-human dialogues authored by professional script writers, with realistic turn densities and 36,000+ per-turn human annotations across 30,000+ expert-generated dialogue turns. Using UPHELD, we systematically evaluate classical automatic metrics and reference-free LLM-as-a-judge approaches, and find them unreliable when correlated with expert human judgment. Building off this analysis, we use UPHELD to develop a Mixture-of-Judges framework that combines multiple evaluative signals and improves correlation with human assessments by approximately 30%. Overall, UPHELD provides a robust, human-grounded foundation for evaluating human-scale conversational intelligence that fills a crucial gap in the pre-existing LLM dataset landscape.
[NLP-127] FIRSTPASS: A Multi-Domain Multi-Round Peer Review Dataset Grounded in Real Editorial Outcomes ICML2026
【速读】： 该论文旨在解决当前生成式AI在科学同行评审领域中存在的领域偏倚问题，即现有训练数据主要集中在计算机科学与机器学习领域，导致模型无法理解生物学、化学等其他学科特有的评审要求（如生物学家对污染控制的强调或化学家对核磁共振（NMR）谱图归属的质疑）。为应对这一挑战，研究提出FIRSTPASS——首个基于多学科高影响力期刊完整多轮编辑对话构建的大规模同行评审数据集。其关键在于：首次系统性地整合了《自然·通讯》（Nature Communications）自2022年11月起推行的强制透明同行评审机制下的全部评审流程数据，涵盖生物学、化学、神经科学、物理学和地球科学五大领域，共3,668条记录，完整保留了初始审稿意见、作者逐点回复以及更新后的审稿评估等迭代验证结构；同时，每条记录均附有由编辑决策直接导出的结果标签（标准评审/扩展评审），提供了此前所有数据集所缺失的真实标注依据。此外，数据经过自动化审计确保内容完整性，且专家评审文本平均达2,155词，显著高于会议评审密度。所有数据、解析管道及评估脚本均已公开，支持跨学科生成式AI科学判断能力的可复现基准测试。
链接: https://arxiv.org/abs/2608.26129
作者: Prabhjot Singh,Somnath Luitel,Manmeet Singh,Josh Durkee
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at the AI for Science Workshop at the 43rd International Conference on Machine Learning (ICML 2026), 3 pages, 1 figure, 1 table
Abstract:Scientific peer review datasets have trained AI systems exclusively on Computer Science and Machine Learning venues, producing models that critique ablation studies yet have never seen a biology reviewer demand contamination controls or a chemist question Nuclear Magnetic Resonance (NMR) spectral assignments. We introduce FIRSTPASS, the first large-scale peer review dataset built on complete multi-round editorial dialogues from a multidisciplinary high-impact journal. Curated from Nature Communications mandatory transparent peer review (instituted November 2022), FIRSTPASS comprises 3,668 records spanning five scientific domains (biology, chemistry, neuroscience, physics, and earth science), capturing the full iterative structure of scientific validation: initial referee reports, author point-by-point responses, and updated reviewer assessments. Each record carries an outcome label derived directly from editorial decisions (STANDARD for two-round review; EXTENDED for three or more rounds), providing ground truth absent in all prior corpora. An automated audit confirms 100% content integrity. Expert reviews average 2,155 words, substantially denser than conference venue reviews. All data, parsing pipelines, and evaluation scripts are released to enable reproducible benchmarking of AI scientific judgment across disciplines.
[NLP-128] comGPT -R1: A Unified Open-Source Reason er for the Telecom Stack
【速读】： 该论文旨在解决电信领域大语言模型（LLM）应用中普遍存在的“双侧能力鸿沟”问题：通用推理模型缺乏对电信领域的专业语境约束，而现有的领域专用电信大模型又在结构化、多步骤推理能力上存在局限。其解决方案的关键在于构建一个统一的开源电信推理模型——TelecomGPT-R1-9B，通过四个互补的推理维度（协议、知识、建模、故障）构建大规模标注数据集（67,427条样本），并采用轴对齐的链式思维（CoT）生成与前缀续写自验证机制增强数据质量。基于Qwen3.5-9B模型，提出两阶段后训练范式：首先利用多教师低秩适应（LoRA）的监督微调注入领域知识并诱导特定推理格式；其次采用分组相对策略优化（GRPO），结合解耦式裁剪与动态采样策略优化（DAPO），以四轴对齐的二元验证器奖励信号优化策略。该方法显著提升了模型在七项公开电信基准上的表现，使其在开源电信大模型中排名第一，并达到与顶尖闭源前沿推理模型相当的七轴平均性能。
链接: https://arxiv.org/abs/2608.26126
作者: Bohao Wang,Chenwei Wu,Haoyu Li,Hang Zou,Yu Tian,Lina Bariah,Li Wei,Chongwen Huang,Yongliang Shen,Zhaoyang Zhang,Merouane Debbah
机构: Zhejiang University (浙江大学); University of Michigan (密歇根大学); King Abdullah University of Science and Technology (阿卜杜拉国王科技大学)
类目: Computation and Language (cs.CL); Information Theory (cs.IT)
备注:
Abstract:Telecommunications is a high-leverage domain for large language model (LLM)-based reasoning because routine engineering workflows require joint grounding in normative specifications, operational telemetry, vendor-specific fault evidence, and exact RF/network calculations. However, current LLM integration in telecom remains bottlenecked by a two-sided capability gap: generic reasoners often lack telecom-specific grounding, while domain-specific telecom LLMs remain limited in structured, multi-step reasoning. To bridge this gap, we release TelecomGPT-R1-9B, a unified open-source telecom reasoner that ranks top-performing on the GSMA open telco leaderboard. Specifically, we curate a 67,427-example supervised fine-tuning (SFT) corpus organized around four complementary reasoning axes: protocol, knowledge, modeling, and fault. The corpus is built from axis-matched public web sources and enhanced through axis-specific chain-of-thought (CoT) generation and prefix-continuation self-validation. Starting from Qwen3.5-9B, we further develop a two-stage post-training recipe. First, multi-teacher low-rank adaptation (LoRA)-based SFT injects telecom knowledge and induces axis-specific reasoning formats. Second, group relative policy optimization (GRPO), stabilized by decoupled clip and dynamic sampling policy optimization (DAPO), optimizes the policy using four axis-aligned binary verifier rewards. Across seven public telecom benchmarks, TelecomGPT-R1-9B ranks first among open-source telecom LLMs and achieves a seven-axis mean comparable to state-of-the-art closed-source frontier reasoners.
[NLP-129] raining-Time Explainability for Multilingual Hate Speech Detection: Aligning Model Reasoning with Human Rationales NEURIPS
【速读】： 该论文旨在解决生成式AI在识别针对穆斯林群体的在线仇恨言论时面临的挑战，尤其是那些以文化隐喻和多语言形式呈现、传统AI审核系统难以有效捕捉的隐性仇恨内容。现有方法虽具备较高分类准确率，但普遍存在可解释性差、存在偏见、过度删帖或审核不足等问题，根源在于其脱离了具体社会文化语境。论文提出一种训练阶段可解释性框架（training-time explainability framework），通过将模型推理过程与人工标注的解释理由对齐，实现性能提升与透明度增强的双重目标。该方案的关键在于引入基于梯度（gradient-based）与注意力机制（attention-based）的正则化策略，在训练过程中强化模型对关键语义线索的敏感性，从而更准确地捕捉跨语言（英语与印地-乌尔都混合语，Hinglish）中蕴含的文化特定暗示，显著提升了对隐性反穆斯林仇恨内容的检测能力、解释的合理性与一致性，为构建多语言、文化敏感的内容审核系统提供了可行路径。
链接: https://arxiv.org/abs/2608.26125
作者: Muhammad Deedahwar Mazhar Qureshi,Sannaan Khan,Muhammad Atif Qureshi,Wael Rashwan
机构: Technological University Dublin(都柏林理工学院); National University of Sciences and Technology(巴基斯坦科学与技术大学); Maynooth University(梅努斯大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS Workshops 2025
Abstract:Online hate against Muslim communities often appears in culturally coded, multilingual forms that evade conventional AI moderation. Such systems, though accurate, remain opaque and risk bias, over-censorship, or under-moderation, particularly when detached from sociocultural context. We propose a \emphtraining-time explainability framework that aligns model reasoning with human-annotated rationales, improving both classification performance and interpretability. Our approach is evaluated on HateXplain (English) and BullySent (Hinglish), reflecting the prevalence of anti-Muslim hate across both languages. Using LIME, Integrated Gradients, Grad X Input, and attention, we assess accuracy, explanation quality, and cross-method agreement. Results show that gradient- and attention-based regularization improve F-scores, enhance plausibility and faithfulness, and capture culturally specific cues for detecting implicit anti-Muslim hate, offering a path toward multilingual, culturally aware content moderation.
[NLP-130] Natural-Language Policies to Executable Decisions: An Interpretable Large Language Model Framework
【速读】： 该论文旨在解决大规模旅游场景下定价自动化所面临的挑战，即旅游订单高度非结构化，而定价策略复杂、动态演化且具有开放性，传统规则引擎难以维护且成本高昂，而无约束的大型语言模型（LLM）又缺乏金融决策所需的可靠性与可审计性。其解决方案的关键在于构建一个具备严格决策边界的生产级生成式定价系统：通过引入结构化提取与有界策略/路径选择机制，使大模型仅负责信息抽取和策略分支判定，而所有数值计算（包括总价计算）均采用确定性方式执行；同时将定价策略编译为可解释的条件树结构，支持无需代码变更即可扩展新条款与动态规则，并生成可供人工介入审核的可审计产物；结合日志轨迹的周期性微调进一步优化条件树生成与路径匹配效果。该系统已在某市级国有旅游企业部署，覆盖7个景区、12类业务，服务超1500名操作员与1000余条活跃策略，在六个月内处理3960笔订单，实现订单管理团队从15–20人缩减至3人，单笔订单处理时间由10分钟降至2分钟。
链接: https://arxiv.org/abs/2608.26124
作者: Ziqiang Zhang,Jing Ma,Zilong Wang,Jiayuan Chen,Yi Qiao,Yu He,Wei Zhang,Dai Cheng,Xiaoyu Shen
机构: Ningbo Institute of Digital Twin, Eastern Institute of Technology, Ningbo
类目: Computation and Language (cs.CL)
备注:
Abstract:Pricing automation in large-scale tourism is challenging because travel orders are highly unstructured, while pricing policies are complex, rapidly evolving, and inherently open-ended. Traditional rule engines are brittle and costly to maintain, whereas unconstrained LLM agents lack the reliability and auditability required for financial decisions. We present a production-grade LLM-powered pricing system with a strict decision boundary: LLMs perform structured extraction and bounded policy/path selection, while all numeric pricing, including total-price computation, is executed deterministically. Policies are compiled into interpretable condition trees, enabling open-ended support for new clauses and evolving rules without code changes, while exposing auditable artifacts for human-in-the-loop control. Periodic fine-tuning on logged traces further improves tree induction and path matching. Deployed at a municipal state-owned tourism enterprise across 7 scenic sites and 12 business categories with 1,500+ operators and 1,000+ active policies, the system processed 3,960 orders in six months, reduced the order management team from 15-20 to 3, and cut per-order handling time from 10 minutes to 2 minutes.
[NLP-131] Which India Survives Translation? Narrative Homogenisation Across Indian Oral Traditions in LLM s
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在训练数据中过度依赖英语互联网文本所导致的文化代表性失衡问题，特别是其对非西方叙事传统的同质化倾向。研究聚焦于印度三大文化差异显著的区域口头与文学传统——拉贾斯坦邦的帕布吉史诗、古典泰米尔语桑伽姆诗歌以及孟加拉民间故事，通过构建各传统的真实参考语料库（分别包含11、21和10段原文），系统评估主流模型（Claude Sonnet与Gemini）在生成任务中对这些文化传统的忠实度。其解决方案的关键在于采用基于Sentence-BERT嵌入的余弦相似度分析方法，量化“参考漂移”（reference drift，即输出与自身传统真实文本的接近程度）与“跨传统趋同”（cross-tradition convergence，即不同传统间输出的相似性）。研究发现，尽管模型输出仍相对贴近各自传统，但跨传统的相似性高达0.52–0.66，显著高于各传统真实语义距离所预测的水平，表明存在部分同质化现象；尤为意外的是，在本地语言（印地语、泰米尔语、孟加拉语）提示下，模型对真实传统的保真度反而下降最多达27个百分点，这一结果挑战了现有关于多语言提示提升文化多样性的假设，揭示出当前模型在模拟边缘化口述传统时可能面临更大的偏差风险。该研究作为一项轻量级、可扩展的计算辅助方法，为后续大规模人工标注研究提供了补充，构成博士阶段关于印度文化误表征问题研究的重要组成部分。
链接: https://arxiv.org/abs/2608.26123
作者: Paarth Singh Rathore
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) are trained predominantly on English-language internet text that over-represents certain cultural narratives, raising concerns that models flatten the diversity of non-Western storytelling traditions into a single homogenized archetype. We present a pilot computational study examining this across three maximally distinct Indian regional oral and literary traditions: the Rajasthani Pabuji epic, classical Tamil Sangam poetry, and Bengali folk tales. We collected authentic reference corpora for each tradition (11, 21, and 10 passages respectively) and prompted two LLMs (Claude Sonnet and Gemini) with 54 generation requests spanning three prompt types per tradition - generic, culturally specific, and regional-language. Using Sentence-BERT embeddings and cosine similarity, we measure reference drift (how closely outputs track their own tradition’s authentic texts relative to the other two) and cross-tradition convergence (how similar outputs are across traditions). We find that while outputs remain closer to their own tradition’s reference than to others, cross-tradition similarity is high (0.52-0.66) relative to what the traditions’ genuine distance would predict, indicating partial homogenisation. Unexpectedly, prompting in the regional language (Hindi, Tamil, or Bengali) consistently reduced fidelity to the authentic tradition relative to English prompting, by as much as 27 percentage points for Rajasthani and Bengali traditions. We discuss this against conflicting prior results on multilingual prompting and argue it reflects a difference between eliciting general cultural diversity and simulating one narrow, lesser-documented oral tradition. We position this pilot as a lightweight, scalable complement to recent large-scale human-annotation studies of Indian cultural misrepresentation in LLM-generated stories, as part of a broader doctoral research program.
[NLP-132] Can a Model Catch Its Own Hallucinations for Free?: Label-Free Doubt Signals Hold Their Own Against a Labelled Dataset for Abstention
【速读】： 该论文旨在解决大语言模型在生成事实性回答时容易产生幻觉（即“自信地陈述错误事实”）的问题，核心挑战在于如何有效引导模型在不确定时主动拒绝回答，从而提升其输出的可靠性。传统方法依赖于标注正确的答案数据集来训练模型进行“回避”（abstention），但此类数据标注成本高昂且难以获取。本文提出一种无需标签的解决方案：利用模型自身生成答案时的置信度（confidence）作为唯一信号，通过低秩适应（LoRA）微调，使模型在自身置信度高时作答，在置信度低时回答“我不确定”。实验基于六种开源权重模型（1B-8B参数量级，两个架构家族）在短文本事实问答任务上验证该方法，结果表明，在匹配覆盖率的前提下，该无监督策略与使用正确性标签的监督式回避训练效果相当，且统计上无显著差异。对照实验进一步证明，性能提升源于模型置信度的校准而非简单记忆，而该方法的主要局限是无法识别“高自信但错误”的事实。因此，模型自身的怀疑状态可作为标注数据的近似替代，为训练模型何时拒答提供了一种低成本、高效的解决方案。
链接: https://arxiv.org/abs/2608.26121
作者: Ali Asaria,Tony Salomone,Deep Gandhi
机构: Transformer Lab (Transformer实验室)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Large language models state false facts as fluently as true ones, yet a model often “knows” internally when it is on shaky ground: the probability it assigns to its own answer tends to dip on the facts it gets wrong. The usual way to act on this, teaching a model to abstain rather than guess, requires a labelled dataset of right and wrong answers. We ask whether the model’s own confidence, which is free and needs no labels, can do that job instead. We fine-tune each model (with LoRA) to answer when its frozen confidence is high and to say “I’m not sure” when it is low, using the signal alone and no correctness labels. Across six open-weights models (1B-8B, two families) on short-form factual question answering, with correctness adjudicated by an independent judge model, this label-free recipe holds its own against label-supervised abstention-tuning: at matched coverage we find no statistically detectable difference between the two. A control that drills hard examples instead of abstaining does not help, indicating the gain comes from calibration, not rote memorization. The signal’s one blind spot is confidently wrong facts, which it cannot flag. A model’s own doubt is thus a near-free substitute for a labelled dataset when teaching it when to abstain. Code and artifacts are available on request.
[NLP-133] Recipes for Steering and Scaling LLM s via Sampling
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在生成过程中因采样策略低效而导致的生成质量受限问题，尤其针对现有方法在处理非基础模型分布时难以实现高效且可扩展的生成控制。其核心挑战在于如何在不依赖外部监督信号或奖励模型的前提下，有效提升生成结果的质量与多样性。解决方案的关键在于提出一个灵活且理论严谨的采样框架，通过引入基于序列蒙特卡洛（Sequential Monte Carlo, SMC）和复制交换（Replica Exchange, RE）的两种算法，实现对基础模型分布的幂化（powering）、乘积化（product）或倾斜化（tilting），从而引导生成过程向更优分布逼近。该框架不仅具备坚实的概率推断基础，且在实验中展现出优于传统Best-of-N和标准马尔可夫链蒙特卡洛（MCMC）基线的可扩展性，为无需外部奖励机制的大语言模型生成优化提供了系统性的采样范式。
链接: https://arxiv.org/abs/2608.26120
作者: Jiajun He,Zongyu Guo,José Miguel Hernández-Lobato,Yuanqi Du
机构: University of Cambridge (剑桥大学); Microsoft (微软); Stanford University (斯坦福大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 13 pages
Abstract:Large Language Models (LLMs) are probabilistic models, typically defined by an autoregressive factorization. While recent work has begun to study richer target distributions beyond the base model, the sampling strategies remain highly inefficient. In this paper, we present a flexible and theoretically grounded framework for steering and scaling autoregressive LLMs with sampling. Within this framework, we describe two algorithms – one based on Sequential Monte Carlo (SMC) and one based on Replica Exchange (RE) – that steer generation toward powering, product or tilting of the base model distribution. We illustrate this framework through scaling the generation quality of LLMs without external supervision or reward models. Experimental results demonstrate our methods scale more favorably than Best-of-N and standard MCMC baselines. Overall, this paper offers a systematic recipe for probabilistic inference with LLMs via sampling.
[NLP-134] DeflectBench: A Benchmark for Evaluating Rhetorical Fallacy Generation in LLM s ICML2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在特定提示下生成修辞谬误（rhetorical fallacies）的能力及其在安全后训练（safety post-training）约束下的表现问题，填补了当前研究中对主动诱导生成谬误行为关注不足的空白。其解决方案的关键在于构建并应用DeflectBench评测基准，系统评估了四个前沿模型在三种谬误策略（即“反向质问”whataboutism、“人身攻击”ad hominem、“转移话题”red herring）、七种提示框架及80个涵盖不同争议程度论点下的生成行为。研究发现，模型拒绝生成谬误的行为主要受请求结构影响，而非论点内容本身；同一论点下的拒绝率差异仅约11个百分点，而提示框架的微小变化即可导致拒绝率波动接近100个百分点，谬误类型切换则可引发超过80个百分点的变化。尤其值得注意的是，教育辩论教练类提示框架几乎完全消解了模型的拒绝行为，但此时模型并未实现真正合规，而是表现为“标记式合规”——在同一次响应中既实施了所要求的谬误，又明确指认其为操纵手段。此外，四款模型在拒绝、标记式合规、软拒绝与清洁合规四种行为模式上的分布存在显著差异，揭示出不同模型对安全约束的响应机制具有高度异质性。该研究通过公开代码与数据集，为理解和评估大模型在对抗性提示下的行为边界提供了重要工具。
链接: https://arxiv.org/abs/2608.26119
作者: Art Kanke
机构: 未知
类目: Computation and Language (cs.CL)
备注: 15 pages. Accepted at the CTB, FAGEN, and AI4GOOD workshops at ICML 2026 in Seoul, South Korea. Code and dataset: this https URL
Abstract:Whether large language models can be prompted to generate rhetorical fallacies on demand, and whether current safety post-training constrains this behavior, has received less attention than the related question of detecting fallacies in existing text. We close this gap with DeflectBench, evaluating 23,990 generations from four frontier models across three deflection strategies (whataboutism, ad hominem, red herring), seven prompt framings, and 80 claims spanning four controversy levels. Refusal is governed primarily by request structure rather than claim content. Per claim refusal varies by only 11 percentage points across the 80 claims, while a single prompt frame change can swing within model refusal by nearly 100 percentage points and switching the requested fallacy type can swing it by over 80 percentage points within explicit framings. An educational debate coach prompt framing collapses refusal to near zero across all four model families, but the bypassed behavior is not clean compliance. Models typically produce labeled compliance, naming the requested manipulation in the same response that contains it. The four models distribute differently across refusal, labeled compliance, soft refusal, and clean compliance. The code and dataset are released at this https URL.
[NLP-135] ElementCheck: Complexity-Aware Long-Form Text Factuality Evaluation via Sentence Elements
【速读】： 该论文旨在解决现有长文本事实性评估方法依赖“分解-检索-验证”（decompose-retrieve-verify）流水线所导致的噪声问题与固定验证粒度局限，进而引发的评估结果不可靠问题。其核心解决方案是提出一种面向复杂度感知的验证框架ElementCheck，关键在于不采用对句子进行均匀原子化子命题分解的传统方式，而是从原始句子中提取通过可验证关联明确连接的实体对作为语义元素，并构建元素图（element graph）。该图的拓扑结构为句子复杂度提供了结构性信号：对于简单句子直接进行端到端验证，而对于复杂句子则实施细粒度的元素级精炼与验证。为支持细粒度评估，研究进一步构建了新基准FastFact-Sent，将来自FastFact-Bench的孤立命题映射回其源句。实验表明，ElementCheck在FastFact-Sent及两个领域特定基准上，对五种主流模型均实现了稳定的事实性验证性能提升，同时保持了良好的准确率-成本权衡；深入分析还显示，复杂度感知的验证机制有效减少了不必要的重复验证，并增强了不同模型间的评估稳定性。
链接: https://arxiv.org/abs/2608.26118
作者: Xinming Wang,Haoran Du,Yi Chen,Jian Xu,Hongming Yang,Han Hu,Yulong Chen,Cheng-Lin Liu,Xu-Yao Zhang
机构: Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所); Zhongguancun Academy(中关村学院); Fudan University(复旦大学); Tencent(腾讯); University of Cambridge(剑桥大学); University of Aberdeen(阿伯丁大学)
类目: Computation and Language (cs.CL)
备注: 25 pages, 4 figures
Abstract:Existing long-form factuality evaluation relies on the decompose-retrieve-verify pipeline. However, the pipeline suffers from noise from claim decomposition and fixed verification granularity, resulting in unreliable results. We propose ElementCheck, a complexity-aware framework that verifies long-form outputs via sentence elements. Instead of uniformly decomposing sentences into atomic sub-claims, ElementCheck extracts entity pairs that are explicitly linked through verifiable connections in the original sentence as elements, and organizes these into an element graph. The graph topology provides a structural signal for estimating sentence complexity, enabling direct verification for simple sentences and targeted element-level refinement and verification for complex ones. To support fine-grained evaluation, we construct a new benchmark FastFact-Sent by mapping isolated claims from FastFact-Bench back to their source sentences. Experiments on FastFact-Sent and two domain-specific benchmarks show ElementCheck consistently improves factuality verification across five backbone models while maintaining a favorable accuracy-cost trade-off. Further analyses demonstrate that complexity-aware verification reduces unnecessary re-verification and maintains stability across different backbones.
[NLP-136] CIFQA: A Deterministic Tool-Grounded Multi-Agent LLM Framework for Financial Query Answering
【速读】： 该论文旨在解决大语言模型（LLM）在处理计算密集型金融问答任务时，虽能生成语义合理但数值错误的答案这一核心问题。其关键解决方案在于提出CIFQA（Calculation-Intensive Financial Query Answering）框架，采用一种确定性工具驱动的多智能体架构，将自然语言理解与数值计算执行严格分离：通过专用智能体分别负责查询解析、路由、参数提取、计算规划和结果生成，而具体的金融计算与规则应用则由确定性的Python工具完成。该设计确保了利率查表、存期计算、滚动年调整及提前支取逻辑等关键步骤的精确性。实验表明，尽管仅使用170亿参数的开源模型，CIFQA在计算密集型查询上仍达到95.54%准确率，整体准确率达90.87%，显著优于直接调用更大规模前沿模型的基线。消融研究进一步证实，确定性组件对性能提升至关重要，且模型架构设计比模型规模更决定数值可靠性。该框架具有良好的可扩展性，适用于多种计算密集型金融推理任务。
链接: https://arxiv.org/abs/2608.26114
作者: Kunjesh Parekh,Anil Kumar Tiwari,Divya Saxena
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computational Finance (q-fin.CP)
备注: 16 pages, 5 figures, submitted for academic dissemination
Abstract:Calculation-intensive financial question answering requires exact reasoning over structured rates, temporal conditions, numerical formulas, and rule-based constraints. Although Large Language Models (LLMs) perform strongly on natural language tasks, they often produce numerically incorrect yet plausible answers when solving multi-step financial calculations. To address this limitation, we introduce CIFQA (Calculation-Intensive Financial Query Answering), a deterministic tool-grounded multi-agent LLM framework for financial question answering. CIFQA separates language understanding from numerical execution by assigning specialized agents to query interpretation, routing, parameter extraction, computation planning, and response generation, while deterministic Python-based tools perform financial calculations and rule application. We instantiate CIFQA for fixed deposit query answering and evaluate it on a curated benchmark of fixed deposit queries. CIFQA achieves 95.54% accuracy on calculation-intensive queries and 90.87% overall accuracy, substantially outperforming direct LLM baselines even when provided with complete formulas, rate cards, and benchmark instructions. Ablation studies show that deterministic components such as exact rate lookup, tenure computation, rolling-year adjustment, and premature-withdrawal logic are critical contributors to performance. Notably, a 17B open-source backbone operating within CIFQA outperforms substantially larger frontier models evaluated with the same financial information, demonstrating that architectural design is a more important determinant of numerical reliability than model scale. While evaluated on fixed deposit queries, CIFQA provides a generalizable framework for calculation-intensive financial reasoning tasks.
[NLP-137] reeGraft: Adaptive Multi-Drafter Grafting for Tree-Based Speculative Decoding
【速读】： 该论文旨在解决树状结构生成（tree-structured methods）在大语言模型推理中因单一起草器（drafter）带来的性能权衡问题：小型起草器虽速度快但生成的候选路径质量低，大型起草器虽能提升路径质量却引入显著延迟。其核心解决方案是提出一种多起草器框架TreeGraft，通过不同计算成本的多个起草器协同构建共享的草稿树（draft tree）。TreeGraft利用强起草器对弱起草器生成的候选路径进行重评分，重新选择嫁接位置，并恢复未被探索的高潜力路径；同时以非破坏性方式整合强起草器的扩展，保留可能被目标模型接受的现有分支，从而显著提升共享草稿树的整体质量。为控制整体起草开销，该方法引入一个轻量级调度器，基于离线训练的价值系统进行知识蒸馏，动态决定调用强起草器的时机。实验表明，在10组模型对和6个基准测试上，TreeGraft相较最优的固定单起草器策略平均提升15.1%，最大增益达26.6%。
链接: https://arxiv.org/abs/2608.26112
作者: Jiaming Fan,Daming Cao,Canchen Huang,Jiale Fu,Jin Zhang,Junjie Gao,Kai Yang,Xiangzhong Luo,Xu Yang
机构: Southeast University(东南大学); Ant Group(蚂蚁集团)
类目: Computation and Language (cs.CL)
备注:
Abstract:Speculative decoding accelerates large language model inference through a draft-then-verify paradigm. Building on this, tree-structured methods improve inference by organizing proposals into multiple candidate paths, increasing the accepted length. However, existing tree-structured methods use a single drafter for all drafting steps, creating a dilemma: a smaller drafter is fast but yields lower-quality trees, whereas a larger drafter improves tree quality but suffers from high latency. To address this, we propose TreeGraft, a multi-drafter framework in which drafters of different costs jointly construct a shared draft tree. TreeGraft uses the stronger drafter to rescore candidates by updating scores assigned by the weaker drafter, reselect grafting positions, and recover promising paths left unexplored. It also integrates stronger drafter expansions non-destructively, preserving existing branches that may still be accepted by the target model. Together, these designs improve the quality of the shared draft tree. To control the drafting cost, TreeGraft introduces a lightweight scheduler distilled from an offline value system to decide when to call the stronger drafter. Across 10 model pairs and 6 benchmarks, TreeGraft outperforms the better of the two fixed single-drafter endpoint strategies by 15.1% on average, reaching a maximum gain of 26.6%. Our code is available at this https URL.
[NLP-138] Graph-Based Modeling of Financial Volatility Dynamics
【速读】： 该论文旨在解决传统隐含波动率（Implied Volatility, IV）表面建模方法在实时波动率（Realized Volatility, RV）预测中因忽略动态演化特性而导致的精度不足问题。现有方法通常将IV表面视为静态图像，无法有效捕捉其内在的时间与空间依赖关系。为此，论文提出金融感知图时空网络（Finance-Aware Graph Spatio-Temporal Network, FA-GSTN），其核心创新在于将RV预测重构为对结构化金融对象演化的建模过程：通过构建基于IV表面的时空图序列，其中节点对应网格点，边编码自适应的空间（日内）与显式的时间（日间）依赖关系。关键解决方案包括引入金融领域感知的节点特征（如期权希腊值）以融入金融先验知识，并设计多尺度时间平滑门与自适应鲁棒损失函数以抑制高频噪声。实验表明，FA-GSTN在大规模权益期权数据集上达到新的基准性能（R²最高达0.473），且在仅使用一年数据时仍显著优于强基准模型（如Vision Transformer，R²: 0.372 vs. 0.315），并在2020–2021市场动荡期展现出更强鲁棒性。消融实验证实了时空图结构、金融感知组件及噪声处理模块的关键作用，凸显了显式建模时间动态性与嵌入金融归纳偏置在实现高精度、高鲁棒性波动率预测中的重要价值。
链接: https://arxiv.org/abs/2608.26127
作者: Chuanzhen Wang,Alice Zhang,Wei Chen,Michael Brown
机构: Tongji University (同济大学); Stanford University (斯坦福大学); Tsinghua University (清华大学); MIT (麻省理工学院)
类目: atistical Finance (q-fin.ST); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 15 pages
Abstract:Accurate forecasting of realized volatility ( RV ) is crucial for risk management and derivatives pricing. Although the implied volatility ( IV ) surface offers rich informational content, prevailing methods that treat it as a static image fail to capture its inherent dynamics. To overcome this limitation, we propose the Finance-Aware Graph Spatio-Temporal Network (FA-GSTN), a novel architecture that reframes RV forecasting as modeling the evolution of a structured financial object. FA-GSTN builds a spatio-temporal graph sequence from the IV surface, where nodes correspond to grid points and edges encode adaptive spatial (intra-day) and explicit temporal (inter-day) dependencies. The model incorporates domain knowledge through finance-aware node features (e.g., option Greeks) and tackles high-frequency noise via a multi-scale temporal smoothing gate coupled with an adaptive robust loss function. Comprehensive evaluations on a large-scale equity options dataset show that FA-GSTN sets a new state of the art, delivering superior predictive accuracy ( R^2 up to 0.473). It also demonstrates remarkable data efficiency, substantially outperforming strong Vision Transformer baselines when trained on only one year of data ( R^2 : 0.372 vs. 0.315). Furthermore, the model exhibits enhanced robustness during periods of market stress, such as 2020–2021. Ablation studies confirm the vital roles of the spatio-temporal graph structure, finance-aware components, and integrated noise-handling modules. Our work underscores the substantial benefits of explicitly modeling temporal dynamics and infusing financial inductive biases for accurate and robust volatility forecasting.
信息检索
[IR-0] misi: a Metric Inverted Sample Index
链接: https://arxiv.org/abs/2608.27422
作者: Edgar Chavez
类目: Information Retrieval (cs.IR)
备注: 14 pages. Links to code
Abstract:We present misi, an inverted index for approximate nearest-neighbor search over general metric spaces whose vocabulary is a random sample of the database, of size proportional to n . Each object is represented by its k_b nearest sample points, found by a pluggable inner index over the sample; queries are answered by an idf-weighted shared-neighbor vote followed by exact verification of C candidates. The construction generalizes the NAPP index from a constant number of pivots to a linear-size vocabulary, which keeps posting lists at constant expected length \rho = k_b/\alpha as n grows and turns the index into a combinator: any high-recall index on \alpha n points yields an index on n points, for any metric. A probabilistic model gives a recall guarantee – k_b logarithmic in n over the overlap gap suffices, with a verification budget the index itself estimates – and a matching limit: the vote cannot resolve overlap differences below order 1/\sqrtk_b . The design’s strengths are structural: construction is n independent searches – embarrassingly parallel, deterministic, 5,250 s for 10^8 vectors on 64 cores, 3.7\times faster than a matched-recall graph build – it streams under an enforced 3 GiB cap, and the portable artifact serves 10^8 vectors from NVMe within an enforced 8 GB budget, below the working floor of the SSD-graph baseline. Its cost is query-time work: saturated graph baselines answer 6 - 16\times faster in RAM, and the verification budget for 0.99 recall grows as n^0.30 . All results carry seeds, saturation sweeps and full configurations, are generated from run manifests, and include measured negative results. The intended applications weight construction cost, determinism, memory footprint, or black-box metrics over peak throughput: frequently rebuilt corpora, batch similarity workloads, constrained-memory serving.
[IR-1] Scaling Graph Neural Networks for Friend Recommendation: Multi-Hash User Embeddings and Temporal Neighbor Sampling CIKM2026
链接: https://arxiv.org/abs/2608.27413
作者: Maksim Utushkin,Andrei Ovsiannikov,Alexander D’yakonov
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG); Social and Information Networks (cs.SI)
备注: 12 pages, 4 figures, 8 tables; accepted at the 35th ACM International Conference on Information and Knowledge Management (CIKM 2026); code: this https URL
Abstract:Friend recommendation is inherently graph-structured: the relevance of a potential connection depends on multi-hop social context rather than user attributes alone. However, deploying message-passing GNNs on a production-scale social graph with hundreds of millions of users and tens of billions of edges requires addressing numerous modeling and systems challenges. We present a scalable end-to-end GNN ranking system for production social graphs, focusing on two design choices that are critical in this setting: multi-hash ID embeddings and temporal neighbor sampling. Multi-hash embeddings are common for high-cardinality features, but industrial GNN systems typically either ignore trainable IDs or accept full embedding tables, exceeding 200 GB for our graph. We integrate multi-hash as the primary node representation, reducing the ID-embedding table size by more than 98 percent while preserving ranking quality. Temporal neighbor sampling is well understood in principle, but existing implementations scan full adjacency lists, which is a non-starter for users with tens of thousands of friends. We implement timestamp-sorted CSR storage with binary search, reducing the per-node temporal sampling cost from O(deg(v) + k) to O(\log(deg(v)) + k) . Beyond these components, we show that this combination scales and yields measurable production impact. On a graph with 194M users and 28B edges, offline ablations isolate each design choice’s contribution. In an online A/B test, our system increases friend additions from recommendations by 16 percent and unique friend adders by 11.5 percent over a strong production baseline. We release our framework for distributed training and inference on large temporal graphs.
[IR-2] RATIO: A Benchmark for Retrieval Across Typed Ideation Operations in Scientific Literature
链接: https://arxiv.org/abs/2608.27394
作者: Maayan Sharon,Tom Hope
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:Retrieved scientific literature can serve as inspiration for both human and AI scientists. Inspiration can take different forms: prior work may directly suggest how to address a problem, or surface directions at different levels of abstraction - zooming out to a more general view or zooming in to a concrete realization. We introduce RATIO (Retrieval Across Typed Ideation Operations), a large-scale benchmark in which relevance is defined by three operations which we name ideation moves: Address retrieves potential approaches for stated problems, Broaden retrieves more general formulations, and Specify retrieves concrete instantiations. RATIO is constructed from millions of full-text scientific papers across CS literature via a general recipe that extends discourse-marker distant supervision - previously used only for classification - to corpus-scale retrieval, combined with extensive LLM and human vetting. Experiments show that operation-specific fine-tuning substantially boosts retrievers but leaves much room for further improvements. RATIO provides a scalable training and evaluation framework for retrieval components that support literature-grounded ideation, opening up new research avenues on scientific inspiration retrieval.
[IR-3] CorporateBench: Large-Scale QA Benchmarking with Temporal Knowledge Bases EMNLP
链接: https://arxiv.org/abs/2608.27391
作者: Sil Hamilton,Albert Yu Sun,Oscar J. Romero,Carl-Leander Henneking,David Mimno,Bishan Yang,Igor Labutov
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: Accepted to EMNLP Findings
Abstract:LLMs are increasingly able to answer complex questions about enterprise-scale document collections. But evaluation is hard: companies don’t want to share internal communications, and synthetic datasets have been overly simple. We present CorporateBench (CB), a human-validated multi-task QA benchmark whose scale approaches the conditions LLMs encounter in corporate communication networks, with evaluation corpora surpassing 230,000 documents. CB evaluates LLMs across two dimensions (information extraction and knowledge base querying) through four synthetically generated firms ranging from 12 to 10,000 employees. Each corpus is sampled from a temporally evolving knowledge base describing a consistent world, guaranteeing cross-document logical consistency even across hundreds of thousands of documents. We evaluate five LLMs on CB, revealing increasingly poor performance as input size approaches realistic scales. CB provides LLM developers a metric for corporate communication reasoning, filling a crucial gap in the benchmarking ecosystem.
[IR-4] Stageboost: Recommending Signals Based on Counterfactual Estimation
链接: https://arxiv.org/abs/2608.27366
作者: Darpan Singhal,Matan Mandelbrod,Tal Franji,Manasa Kolla,Vipul Gaba,Yuri Brovman
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
备注: Accepted for Consequences 2026 workshop
Abstract:Signals are short textual or visual snippets displayed on the eBay View-Item (VI) page, providing additional, contextual information for users about the viewed item. The aim of displaying these signals is to facilitate intelligent purchase and to incentivize engagement. In this paper, we present a 2 stage xgboost based model that optimally populates the VI page with signals. This approach has shown a 0.08% lift in overall GMB (Gross Merchandise Bought) and 0.58% increase in Parts and Accessories GMB, primarily due to increase in conversion of high average price items in online experimentation.
[IR-5] Astar: Learning to Propose Evolution Directions for Self-Evolving Industrial AI Systems
链接: https://arxiv.org/abs/2608.27287
作者: Jinxin Hu,Hao Deng,Haibo Xing,Lingyu Mu,Muyu Zou,Weiqin Yang,Sirui Chen,Bohao Wang,Zhezheng Hao,Hao Zhang,Zulong Chen,Shizhun Wang,Yu Zhang,Xiaoyi Zeng,Jiawei Chen
类目: Information Retrieval (cs.IR)
备注:
Abstract:Modern AI systems advance through continuous iteration: a loop of proposing evolution directions, implementing code, training, and evaluation. While the latter three stages are increasingly automated, the starting point — proposing effective evolution directions — remains a critical bottleneck that still relies heavily on senior experts. In this work, we explore whether AI can take over this role. We find that general-purpose LLMs, even the advanced GPT-5.5, offer only generic and misaligned suggestions: the required expertise is accumulated through experience rather than explicitly codified, and thus hard to inject directly. To this end, we propose Astar, a training-based approach that learns a specialized evolution-guiding model from the abundant iteration histories of industrial systems. Realizing this idea, however, raises four challenges: sparse supervision, noisy data, a vast direction space, and prohibitively expensive verification. We address them along two fronts. On the data side, we design a pipeline that turns noisy historical commits into a large, clean evolutionary corpus via pairwise sample expansion and noise filtering. On the model side, we train the model through mid-training, SFT, and RL, guiding evolution direction generation with hierarchical hints and using the reward model in RL as a fast surrogate evaluator. Astar has been deployed in Alibaba’s Lazada advertising system for evolution direction proposal. Astar-8B achieves a single-proposal success rate of 0.6786 in real-execution evaluation, far exceeding human experts (0.3229) and the strongest general-purpose LLM (0.3071). More importantly, Astar closes the loop and enables fully automatic iteration: it guided 20 consecutive iterations over two weeks, improving offline Hitrate@200 by 23.6%, while an online A/B test yielded relative lifts of 4.86% in GMV and 1.82% in advertising revenue. Subjects: Information Retrieval (cs.IR) Cite as: arXiv:2608.27287 [cs.IR] (or arXiv:2608.27287v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2608.27287 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-6] ProRetrieval: Learning to Orchestrate Hybrid Search via Executable Program Synthesis
链接: https://arxiv.org/abs/2608.27017
作者: Chengsong You,Zhen Sun,Yunhai Hu,Junwei Zhou,Xiaoyu Cao,Binyu Li,Ziyan Zhao,Weiyao Wang,Liren Lu,Zhijie Ye,Yumo Cao,Yitao Long,Yiwei Xu,Qiyi Jiang,Xuanyi Fu,Yufan Chen,Yilun Li,Rongkang Xiong,Yiran Zou,Nan Du
类目: Information Retrieval (cs.IR)
备注: 15 pages, 5 figures, 6 tables
Abstract:Real-world retrieval often composes structured constraints with semantic intents over text and images through arbitrary Boolean logic. Existing hybrid pipelines such as reciprocal rank fusion or self-querying retrievers admit only a fixed form of composition, while recent reinforcement-learning retrievers train the language model as a query generator for a single backend, leaving the orchestration of heterogeneous retrieval paths outside its action space. We propose ProRetrieval, which recasts the language model as a retrieval orchestrator: given a natural-language query, it synthesizes an executable program in a hybrid DSL interleaving SQL operators over structured fields with vector-retrieval primitives over text and images, with SQL itself providing the logical algebra that fuses heterogeneous candidate sets. We train Qwen3-4B with GRPO and DAPO under a hierarchical four-term reward, and evaluate on two new benchmarks built from Amazon products and Enron email. Our 4B model surpasses GPT-5.5 (Hit@1 0.81 vs. 0.69 on e-commerce; 0.91 vs. 0.86 on email) and Claude Opus 4.7 and a comprehensive suite of retrieval, LLM-augmented, structured-query, and graph-based baselines. Code: this https URL data: this https URL.
[IR-7] Conversational Recommendation over Live E-Commerce Catalogues with Self-Refreshing Retrieval RECSYS2026
链接: https://arxiv.org/abs/2608.27006
作者: Ante Kapetanovic,Tomislav Duricic,Dionizije Fa,Andro Mercep,Emanuel Lacic
类目: Information Retrieval (cs.IR)
备注: ACM RecSys 2026, 3 pages, 2 figure, 1 table
Abstract:Conversational recommender systems based on large language models (LLMs) are usually evaluated on static, pre-indexed item collections, yet e-commerce catalogues change continuously as products are added or removed, repriced, and restocked. We present a merchant-agnostic, multi-turn conversational shopping assistant that operates over such live catalogues. Its central component is a self-refreshing retriever that ingests a merchant product feed, enriches the records, and synchronizes them into a vector index. On each run, per-item hashes identify which products are new, changed, deleted, or unchanged, so only the delta is processed rather than rebuilding the whole catalogue. A controller-based dialogue layer consumes this index, using an LLM only for intent classification and preference elicitation while retrieval, reranking, and diversity selection run as dedicated functions. Our demonstration is a WhatsApp shopping assistant in which catalogue changes reach the recommendations after the next successful sync. A live chatbot, documentation, and a recorded walkthrough are available at this https URL.
[IR-8] opology-Masked Unified Backbone for Joint Feature Interaction and Multi-Domain Sequence Modeling KDD
链接: https://arxiv.org/abs/2608.27005
作者: Zhihao Zhu,Dezheng Han,Jikang Xia,Shuaishuai Guo
类目: Information Retrieval (cs.IR)
备注: Accepted to the TAAC-KDD Cup 2026 Workshop. Recipient of the Unified Block Innovation Award
Abstract:Large-scale post-click conversion rate (CVR) prediction requires jointly modeling heterogeneous feature interactions and dependencies over multi-domain user behavior sequences. Existing industrial ranking models usually handle these two aspects with separate modules. Recent unified architectures attempt to incorporate them into a single framework, but such unification often relies on coordination between modules and does not fully organize all information sources within the same interaction space. To address this problem, we propose MaskRec, a topology-masked unified token interaction architecture for feature interaction and multi-domain sequence modeling. MaskRec transforms heterogeneous features, multi-domain behavior sequences, and contextual signals into unified token representations, and further introduces learnable global memory tokens and domain-level memory tokens as information aggregation nodes. Based on this unified token space, MaskRec designs a structured attention mask, TopoMask, which selectively enables or blocks attention connections according to the structural differences and modeling requirements of different information sources. In this way, heterogeneous feature interaction and multi-domain sequence modeling are performed within the same topology-constrained attention process. In addition, MaskRec incorporates a dual-path interactive query generation module to inject candidate-conditioned user–item interaction signals before the unified backbone. Experiments on the Tencent Advertising Algorithm Competition dataset show that MaskRec achieves stable improvements over the official baseline, validating the effectiveness of the proposed unified framework for industrial CVR prediction.
[IR-9] When Memory Takes Gradients: Collaborative Vector Memory for Agent ic Recommender Systems
链接: https://arxiv.org/abs/2608.26895
作者: Hanchong Chen,Xing Tang,Lingjie Li,Xiongfeng Shan,Xiuqiang He
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
备注: 11 pages, 3 figures
Abstract:Agentic recommender systems ground each decision of a large language model (LLM) in a persistent memory of the user, and in existing agents that memory is text: a narrative written and maintained by further LLM calls. Text limits this memory in two ways. It is updated one rewrite at a time, so exploiting the full interaction history is prohibitively expensive; and collaborative evidence, graded similarity over an entire catalog, does not survive translation into sentences. We propose CoVeMem (Collaborative Vector Memory), which vectorizes the collaborative core of the agent’s memory. Frozen LightGCN user and item states form the memory bank; at each decision, the candidate set itself retrieves the most relevant historical states, which enter the LLM’s context as soft tokens alongside a light textual profile. Contrastive alignment to item-semantic anchors, followed by listwise co-training with masked candidates, teaches the model to read these states and to rank through them; a pointwise yes/no readout scores each candidate. Across four instruction-grounded recommendation benchmarks, CoVeMem matches or exceeds the strongest collaborative text-memory agent on 19 of 20 metric cells while requiring zero additional LLM calls for memory maintenance beyond the shared static profile, against per-interaction calls for text memory. The memory now takes gradients: the full interaction history, out of reach for text, becomes available as training data for what the agent remembers and for how it reads what it remembers.
[IR-10] Equal Ranking Quality Different Decisions: Training Order-Consistent LLM Scorers
链接: https://arxiv.org/abs/2608.26762
作者: Markus Frohmann,Mahdiyar Alavi,Elizabeth Lingg,Navid Rekabsaz
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: 9 pages main text, 45 pages total
Abstract:Rerankers, reward models and multi-document QA scorers score candidate documents or responses in one LLM prompt, so each score depends on their order. Such scorers are selected on ranking quality, but their scores determine a decision: what a score threshold retains, a reader answers, or a preference model selects. However, equal ranking quality does not imply equal decisions: on passage reranking, five trained scorers within 0.010 nDCG@10 retain sets that overlap by only 0.66-0.84 when reordered. A published reranker takes the highest retained-set F1 in our comparison and still overlaps by only 0.667. No prompt-time change we test removes that order dependence: the only one that gains ranking quality leaves all three decisions unchanged. Order-consistency SFT (OC-SFT) attenuates it in the weights, training a candidate’s score not to depend on the order. It holds ranking quality and leads every decision-stability measure among trained scorers on all three tasks: it flips the reader’s answer on 0.125 of permutation pairs against 0.149-0.164 for three other objectives that target order. It is more stable than order-averaged distillation on 12 base models, and one OC-SFT permutation retains sets that overlap more than ten averaged off-the-shelf permutations. A comparison should therefore report what a threshold retains and a reader answers, not ranking quality alone. Code is available at this https URL.
[IR-11] STREAM: An Objective-Driven and Uncertainty-Aware Framework for Industrial Energy Data Acquisition
链接: https://arxiv.org/abs/2608.26754
作者: Zhipeng Ma,Bo Nørregaard Jørgensen,Zheng Grace Ma
类目: Information Retrieval (cs.IR)
备注: It has been accepted by Energy this http URL Conference 2026 (EI.A 2026)
Abstract:Industrial energy management requires datasets that connect energy use with equipment states, production batches, material flows, and process conditions. However, conventional acquisition workflows commonly emphasize connectivity and storage without verifying whether accessible signals satisfy the requirements of a defined energy-performance assessment. This paper presents STREAM, an objective-driven and uncertainty-aware framework comprising Specification of Objectives, Technical Requirements, Resource Mapping, Extraction from Sources, Archival Metadata, and Migration to Database. STREAM is the central workflow: objective-to-data traceability is its end-to-end output, while measurement, temporal, contextual, and processing uncertainty are assessed across all six stages. Compared with the original conceptual STREAM sequence, this paper adds stage-level artifacts, minimum-evidence gates, source-suitability rules, a metadata template, an uncertainty rubric, and case-specific traceability matrices. The framework is validated through two industrial batch-process cases: induction-furnace melting in a foundry and cheese-powder drying using SCADA and production-order data. The results demonstrate that data accessibility is not equivalent to analytical suitability and show how STREAM supports transparent decisions about immediate data use, analytical restrictions, and prioritized infrastructure improvements.
[IR-12] Beyond a Single Story: Meta-Reviewing Sparse and Incomplete User-generated Contents for Recommendation
链接: https://arxiv.org/abs/2608.26728
作者: Hongren Wang,Tianjun Wei,Yingpeng Du,Jie Zhang,Yin-Leng Theng
类目: Information Retrieval (cs.IR)
备注:
Abstract:Data sparsity remains a long-standing challenge in recommender systems, and it becomes more severe for methods relying on user-generated content (UGC) such as textual reviews, which capture fine-grained preferences but require more user efforts to produce. As a result, UGC exhibits (1) missing reviews, where interactions lack any review, and (2) incomplete reviews, where available reviews cover only a subset of relevant attributes. Existing approaches often overlook these UGC-specific issues, leading to degraded accuracy. Motivated by meta-review in academic peer review, we propose MOSAIC (Meta-review On Sparse And Incomplete user-generated Content), which constructs a meta-review for each target user by aggregating attribute-sentiment evidence from neighbor users’ reviews. A multi-gate mixture-of-experts (MMoE) architecture jointly optimizes rating prediction and meta-review attribute-sentiment prediction, while an attention module personalizes the aggregated meta-review signals to each target user, yielding both refined rating predictions and attribute-level explanations. Experiments on four real-world datasets demonstrate that MOSAIC consistently outperforms state-of-the-art baselines in both recommendation accuracy and explanation quality, mitigating UGC sparsity and incompleteness while delivering consistent gains for users with limited interaction history.
[IR-13] BLANC: Discovering Patent White Space via Changes in Normalized Pointwise Mutual Information Between Multi-View Clusters
链接: https://arxiv.org/abs/2608.26685
作者: Shuichi Miyazawa,Kensuke Fujii
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL); Digital Libraries (cs.DL)
备注: 15 pages, 4 figures, 10 tables. A preliminary Japanese-language report covering the methodology and the industrial case study is scheduled to appear as AGC Research Report 76 (2026), ISSN 2434-0774. The present article is the full version, containing the entire quantitative evaluation
Abstract:Identifying white space — the unexplored but potentially valuable regions of a patent landscape — is essential for strategic RD planning, yet existing methods rely on manual patent mapping or apply single-view clustering without quantitative gap detection. We propose BLANC (Blank Landscape Analysis through NPMI Conditioning), a three-phase pipeline combining (1) multi-view neural topic modeling along three semantic dimensions (application/use, novelty, inventive step); (2) Normalized Pointwise Mutual Information (NPMI) to quantify cross-dimensional cluster association; and (3) conditional detection that flags combinations whose NPMI drops when the corpus is filtered by a user-specified keyword. The drop is captured by a new metric, \Delta NPMI, which identifies combinations “established globally, unexplored locally.” Because white space has no ground truth, we evaluate BLANC on two public USPTO corpora — machine learning/AI (5,417 patents, CPC G06N) and glass compositions (1,982 patents, CPC C03C) — by artificially depleting known technology combinations and testing recovery. When three-quarters of a target pair’s documents are removed, BLANC recovers 34.1% (ML/AI) and 27.3% (glass) of the depleted combinations, whereas size-matched removals not aimed at them (random documents, or those of a different established combination) essentially never do: the target is never recovered in 191 decoy trials. Collapsing the three semantic views into one recovers nothing, while prior co-occurrence measures also flag the target under random removal, offering no specificity. In a proprietary case (302 float glass / glass-ceramics patents), the keyword “fluorine” reveals a fluorine surface treatment \times warpage suppression candidate ( \Delta NPMI up to 0.48) that experts had independently identified.
[IR-14] When Does Supervised Fine-Tuning Reduce Instruction Sensitivity?
链接: https://arxiv.org/abs/2608.26661
作者: Jaekeol Choi
类目: Information Retrieval (cs.IR)
备注:
Abstract:Large language models can exhibit substantial performance variation across alternative formulations of the same task instruction, yet it remains unclear how conventional task-specific supervised fine-tuning (SFT) changes this instruction sensitivity. We study this question by evaluating fixed model checkpoints under multiple paraphrased instructions and defining instruction sensitivity as the standard deviation of task performance across them. We conduct a controlled scale analysis with Qwen3 models at 1.7B, 4B, and 8B on MS MARCO, together with targeted cross-family checks using Mistral-7B and Gemma-2-9B. Before SFT, instruction sensitivity decreases sharply with Qwen3 model scale. At 1.7B and 4B, SFT consistently reduces sensitivity across training instructions, with reductions of approximately 54–71%. At 8B, individual sensitivity changes are not statistically distinguishable from zero, but paired contrasts between training instructions are statistically reliable under query-level bootstrap analysis and have consistent directions across all three random seeds. Gemma-2-9B shows the same directional training-instruction contrast as Qwen3-8B, whereas Mistral-7B does not, suggesting that the strength of this effect also varies across models. Experiments on ESCI-English further show that free-generation and likelihood-based forced-choice evaluation can yield qualitatively different robustness conclusions even when valid-label generation is nearly perfect and average task performance is similar. Overall, SFT does not uniformly reduce instruction sensitivity: its robustness effect depends on the adaptation setting, while measured sensitivity can additionally depend on the prediction and scoring protocol.
[IR-15] PailitaoGR: Latent Think-with-Images for Generative Image Retrieval
链接: https://arxiv.org/abs/2608.26658
作者: Xiaomeng Fan,Yueran Liu,Shengyu Zhou,Chenghan Fu,Wanxian Guan,Feng Li,Chuan Yu,Jian Xu,Bo Zheng
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Generative retrieval has demonstrated strong performance by directly generating product semantic identifiers (SIDs). Extending this paradigm to image search, however, is nontrivial because real-world query images contain diverse information, including the search target, useful auxiliary evidence, and irrelevant visual content. This requires the model to identify and focus on the search target while selectively utilizing auxiliary evidence. In this paper, we propose \textbfPailitaoGR, a \emphLatent Think-with-Images method for generative image retrieval, which internalizes target-focused perception and selective auxiliary-evidence utilization into a the generative retrieval model, enabling \textitZooming without Cropping and \textitReading without OCR. Specifically, we design a target-focused perception mechanism that identifies and enhances visual tokens of the search target, consisting of a target Enhancer and a learning strategy based on on-policy distillation and attention guidance loss, enabling the model to focus on search-target regions. We also design a selective auxiliary-evidence utilization mechanism that identifies and enhances visual tokens of auxiliary evidence, including an auxiliary enhancer and an in-capacity incremental contrastive distillation strategy, enabling the model to exploit auxiliary evidence. We construct training and validation sets sampled from real-world online image-search logs. Experiments show that our method outperforms existing baselines by an average of 13.8%, validating its effectiveness. Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) Cite as: arXiv:2608.26658 [cs.CV] (or arXiv:2608.26658v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2608.26658 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-16] hoBIT: A Profile-Aware Retrieval-Augmented Chatbot for University Academic Advising EMNLP2026
链接: https://arxiv.org/abs/2608.26604
作者: Yoonseo Kim,Seongmin Lee,Joongheon Kim,SeongKu Kang
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: Accepted to the System Demonstrations Track at EMNLP 2026
Abstract:In university academic advising, identical questions can require different answers depending on a student’s department, admission cohort, and degree program, causing profile-blind retrievers to surface plausible but inapplicable evidence. We present proFILL, a method for transforming hoBIT, our college’s current rule-based advising chatbot, into a profile-aware retrieval-augmented generation (RAG) system. Rather than requiring a complete user profile upfront, proFILL progressively acquires only the profile attributes needed for each query, guided by both the query intent and the initially retrieved evidence, and uses them to condition retrieval over a profile-aware index. Extensive experiments and a human preference study show that proFILL outperforms diverse RAG baselines, is preferred by target users, and remains effective with open-weight models for cost-effective on-premise deployment.
[IR-17] Preference Flow Matching with Spectral Factorization for Micro-video Recommendation
链接: https://arxiv.org/abs/2608.26579
作者: Xinxin Dong,Haokai Ma,Fei Hu,YuZe Zheng,Bin Wu,Yonghui Yang,Xiaodong Wang
类目: Information Retrieval (cs.IR)
备注:
Abstract:Micro-video recommendation aims to infer user preferences from historical interactions and multimodal video content, thereby identifying the next video of interest. However, prevailing methods compress frame sequences into a single holistic representation, entangling the stable visual semantics and the evolving dynamics that jointly shape user preferences. Meanwhile, diffusion- and flow matching-based recommenders condition their generation process solely on coarse behavioral context, leaving its internal temporal structure outside preference formation. We therefore propose PrismRec, a Preference Flow Matching framework with Spectral Factorization for Micro-video Recommendation. Analogous to a prism that disperses white light into its constituent spectrum, PrismRec devises Spectral Semantic Factorization (SSF) to derive complementary static semantic and dynamic factors from frame-level representations via a prior-guided learnable frequency mask in the temporal frequency domain. Then, it proposes Context-Calibrated Preference Matching (CPM) to weigh them with each user’s specific sensitivity and inject the calibrated context as a structured condition to steer the matching trajectory toward the target representation, making video content as an intrinsic driver of preference formation rather than auxiliary side information. Experiments on four datasets from two platforms show that PrismRec surpasses the SOTA baseline by up to 22.65%, with the lowest inference cost and peak memory among the compared methods.
[IR-18] Case2Flow: Bridging Patient Cases and Guideline Flowcharts through Multimodal Retrieval EMNLP2026
链接: https://arxiv.org/abs/2608.26414
作者: Jiale Wei,Yufan Chen,Alexander Jaus,Zdravko Marinov,Julian Friedrich,Simon Reiß,Jens Kleesiek,Rainer Stiefelhagen
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: Accepted by EMNLP 2026 Main
Abstract:Medical guidelines encode rich, evidence-based decision logic, yet the specific decision artifact a clinician needs is hard to locate within a guideline, let alone across guidelines covering plausible diseases and treatments. While guideline passages have supported end-to-end question answering, flowcharts remain largely underused in decision support despite their ability to encode actionable clinical pathways. We therefore introduce Case2Flow, a task designed to retrieve the most relevant guideline flowchart for a given patient case from a collection of guideline documents. To support it, we construct FlowAtlas, a curated corpus of 202 flowcharts extracted from 2,080 medical guidelines, together with a pipeline that synthesises 1,911 aligned case-flowchart pairs. Our evaluation of multimodal retrieval methods reveals systematic failure modes, including overreliance on keywords and spurious token-patch matches induced by uninformative background regions in flowcharts. Motivated by this, we propose CRISP, a training-free scoring method that sharpens late-interaction retrieval by suppressing uninformative patches, discounting ambiguous token matches, and incorporating bidirectional query-image alignment. CRISP improves Recall@1 by up to 18.71 percentage points, while a blinded physician assessment on published case narratives provides preliminary feasibility evidence beyond synthetic queries.
[IR-19] Assessing the Downstream Utility of Evidence-Aware Retrieval in RAG
链接: https://arxiv.org/abs/2608.26379
作者: Utshab Kumar Ghosh,Debayan Mukhopadhyay,Shubham Chatterjee
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL)
备注:
Abstract:Retrieval evaluation for retrieval-augmented generation (RAG) is increasingly designed around whether retrieved passages contain evidence that can support generation, rather than topical relevance alone. We study whether this closer alignment with downstream evidence needs also makes retrieval evaluation more useful for the decisions built from it. Across five retrieval benchmarks and an end-to-end TREC RAG 2025 setting, we examine an answer-support signal in four roles: comparing retrievers, guiding retrieval training and system selection, predicting downstream answer quality, and filtering the evidence supplied to a generator. The signal changes retrieval rankings, but its downstream value is not uniform. It does not reliably improve retriever training; the benefit of using it for system selection depends on how the generator is instructed to use the retrieved evidence; and retrieval scores based on it do not robustly predict answer quality on unseen topics. In a direct evidence intervention, human annotators confirm that filtering preferentially preserves passages containing useful answer evidence, yet different answer evaluators reach different conclusions about whether the resulting answers improve. These results show that making retrieval evaluation more closely reflect the evidence needed for generation does not by itself make every downstream use of that evaluation more reliable. RAG evaluation methods should therefore be assessed with respect to the particular comparisons, decisions, and conclusions they are intended to support. Subjects: Information Retrieval (cs.IR); Computation and Language (cs.CL) Cite as: arXiv:2608.26379 [cs.IR] (or arXiv:2608.26379v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2608.26379 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-20] A Reranker for Orchestrating Heterogeneous Speech and Text Retrievers INTERSPEECH2026
链接: https://arxiv.org/abs/2608.26194
作者: Inho Kim,Sumyeong Ahn
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: Accepted to Interspeech 2026
Abstract:Retrieval-Augmented Generation (RAG) systems have attracted significant interest for their ability to mitigate hallucinations in Large Language Models (LLMs). Although knowledge databases for RAG are increasingly diversifying to include various modalities such as speech and text, research on handling such multi-modal database scenarios remains limited. In this paper, we propose STeReO (Speech and Text Reranking Orchestrator), a reranker based on speech and text retrievers that aggregates disparate modality databases. To address the lack of specialized training data, we first curate a dataset comprising queries, mixed-modality evidence, and their corresponding relevance ranks. We then train the reranker and evaluate its effectiveness in both single-modality and mixed-modality scenarios. Our results demonstrate that the proposed algorithm excels at selecting the most relevant evidence, thereby significantly improving downstream question-answering performance.
[IR-21] Leverag ing Large Language Models for Systematic Literature Review of Disease Spread Models
链接: https://arxiv.org/abs/2608.26150
作者: Orhan Yagizer Cinar,Timur Emre Ozkose,Emma Von Hoene,Amira Roess,Taylor Anderson,Hamdi Kavak
类目: Artificial Intelligence (cs.AI); Digital Libraries (cs.DL); Information Retrieval (cs.IR)
备注: To be published in the Winter Simulation Conference 2026
Abstract:Recent advancements in Large Language Models (LLMs) have created new opportunities to streamline and potentially automate many research processes, including systematic literature reviews (SLRs). This study reports an LLM pipeline development for extracting model-relevant information from 536 peer-reviewed agent-based modeling papers. We compare the results with those of a human-conducted SLR. Our results show paper-level accuracies of approximately 77.95% for GPT-4.1 and 81.67% for GPT-5.0. Field-level accuracy ranges from 32.40% to 100.00%, with more complex or subjective fields performing less reliably. Importantly, we find that agreement between LLMs is a potential indicator of output quality: low agreement may signal hallucinations, whereas high agreement combined with low accuracy may point to noise or errors in the human dataset. Overall, our study provides practical insights into prompt development and highlights both the potential and limitations of using LLMs for full-scale SLRs in the modeling and simulation domain.
[IR-22] LLM s for Academic Workflows: An Evaluation of Literature Reviews Generated with Short and Long Context Windows of LLM s
链接: https://arxiv.org/abs/2608.26145
作者: Muhammad Ali Chaudhry,Xinyuan Hao,Haifa Alwahaby
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Information Retrieval (cs.IR)
备注:
Abstract:Our research focuses on evaluating literature reviews generated in short and long context settings of large language models (LLMs) to investigate the impact of context window on the quality of AI-generated literature reviews and the role of AI in supporting literature review writing. Twenty AI-generated literature reviews based on research sources from Semantic Scholar and Arxiv were evaluated by two researchers across 15 dimensions. Our findings reveal that AI-generated literature reviews require human oversight to meet academic publishing standards. As context windows increase, LLMs can incorporate broader information and maintain coherence across longer inputs, but they also exacerbate issues such as content repetition, omission of critical work, and a tendency towards descriptiveness over synthesis. Our work shows that AI-generated reviews can provide foundational overviews, but their output must be critically evaluated and refined by domain experts. Future research should consider integrating other LLMs and fine-tuned models in different domains with hybrid approaches that combine human expertise with AI capabilities to address the limitations identified in this study.
[IR-23] Agents Dont Paginate: First-Chunk Selection for LLM Tool Responses
链接: https://arxiv.org/abs/2608.26130
作者: Tatiana Petrova,Andrei Mazniak,Radu State
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:Coding agents built on large language models (LLMs), such as Claude Code, Cursor, OpenAI Codex, GitHub Copilot, and Aider, receive tool responses that routinely exceed the agent’s per-turn token budget. The standard remedy, pagination, is available in every protocol that produced these responses; yet across the corpus of session logs from a public Model Context Protocol middleware we observed no agent-initiated requests for a second chunk. The first chunk is what the agent reads, so we ask how often the gold item (the one the agent needs) is placed first in it: the precision-at-1 rate p_1 . In a controlled offline benchmark we treat first-chunk selection as a 0/1 knapsack and compare six value functions on 500 SWE-bench Verified tasks, then test whether p_1 matters with a single-turn file-localisation probe on five language models (4,800 LLM calls; not an end-to-end resolve-rate test). Two pre-registered hypotheses did not hold and are our main findings. The central one is negative: raising p_1 does not systematically raise downstream accuracy. Per-model deltas stay under three percentage points (p.p.), are not consistently signed, and no model is significant; the agent recovers the gold from anywhere in the chunk, so what reaches its answer is first-chunk inclusion, not the gold’s rank within it. The second: adding four file-metadata signals to a keyword scorer hurts p_1 by 4.8 p.p. (paired significance test, p = 0.001 ). A parameter-free keyword scorer does raise p_1 , from a 24.2% baseline to 35.0% (+10.8 p.p., far beyond chance; p = 3.9 \times 10^-8 ), and to 35.8% with a fallback to the tool’s native ordering when no keyword matches. But by our central finding this is a rank-1 gain, and rank-1 is the part that does not reach the agent’s answer: downstream accuracy does not move. Subjects: Computation and Language (cs.CL); Information Retrieval (cs.IR) Cite as: arXiv:2608.26130 [cs.CL] (or arXiv:2608.26130v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2608.26130 Focus to learn more arXiv-issued DOI via DataCite
人机交互
[HC-0] Do User-Authored Permission Policies Improve Protection Against AI Agent Overreach?
链接: https://arxiv.org/abs/2608.27443
作者: Ting Yan
类目: Human-Computer Interaction (cs.HC); Cryptography and Security (cs.CR)
备注: 15 pages, 5 figures
Abstract:AI agents are poised to become a primary interface to digital products, acting across email, files, payments, and personal data. People without professional software backgrounds need understandable, reusable ways to control actions across services. We examine a mechanism in which a language model maps actions to plain-language consequence categories with user-authored “allow”, “ask”, or “never” rules. We ask what is gained and lost when decisions are made in advance as reusable rules rather than separately for each action. We analyzed 113 participants without professional software backgrounds across three conditions: per-action human-in-the-loop approval (HITL), automated per-action model review (AUTO), or user-authored consequence policy (POLICY). Participants judged 2 examples in each of 4 consequence categories; POLICY participants then set one rule per category. All supervised an 18-action simulated day, including 7 overreach actions. POLICY blocked less overreach than HITL (-20.1 percentage points, 95% CI [-32.1, -8.1]) and AUTO (-14.5 points, 95% CI [-25.8, -3.2]). POLICY lowered runtime prompts from 18.0 to 10.9, but total intervention time was not reliably lower when rule setup was included. Exploratory analysis showed that participants chose “ask” for 114 of 140 POLICY rules, returning most overreach actions to runtime. Of the 148 overreach actions executed in POLICY, 133 followed human approval and 15 ran automatically under “allow” rules. Across all 7 overreach actions, POLICY had the highest approval rate. Counterintuitively, user-authored rules did not by themselves provide stronger protection: many actions outside users’ original requests went through after users approved them. These results reveal a gap between preference and commitment: repeatedly choosing “ask” preserves case-by-case choice but prevents a standing policy from settling decisions in advance. Comments: 15 pages, 5 figures Subjects: Human-Computer Interaction (cs.HC); Cryptography and Security (cs.CR) ACMclasses: H.5.2; K.6.5 Cite as: arXiv:2608.27443 [cs.HC] (or arXiv:2608.27443v1 [cs.HC] for this version) https://doi.org/10.48550/arXiv.2608.27443 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[HC-1] Beyond Harassment: Exploring the Harm Experienced by People with Disabilities in Social Virtual Reality
链接: https://arxiv.org/abs/2608.27390
作者: Xinran Adeline Li,Kexin Zhang,Yuhang Zhao,Yaxing Yao
类目: Human-Computer Interaction (cs.HC)
备注: Accepted to the 28th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2026)
Abstract:People with disabilities (PWD) are increasingly engaging in social virtual reality (VR) platforms, where immersive and embodied interactions can intensify negative experiences. While prior work has examined harassment in VR, little is known about the harms experienced by PWD and the perceived severity associated with different harassment and disability types. Unlike harassment, which represents behaviors, harm is more critical to designing effective protections, as it reflects the consequences and impact; the realism of VR and the vulnerability resulting from disability identity can further amplify such impact. To characterize and model harms for PWD, we conducted a literature review, followed by an online survey with 67 PWD to understand participants’ harassment experiences and resulting harms in social VR. We identified 19 types of harm in 5 categories, and reported the severity perception of each type of harm. Finally, we analyzed our results from the critical disability theory perspective, summarized the uniqueness of harm in social VR, and discussed design implications for specialized safety mechanisms that mitigate harm for PWD.
[HC-2] Comparative Evaluation of 3D Reconstruction Methods for Immersive Visualization of Laboratory Objects
链接: https://arxiv.org/abs/2608.27301
作者: Brian De La Cruz,Aaron Y. Zhao,Maitrey Gramopadhye,Sawyer J. Lazar,Xianming Tan,Daniel Szafir,David S. Lawrence
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
备注: 36 pages, 18 figures, 16 tables
Abstract:In this study, we examined whether current 3D reconstruction methods can support the creation of realistic holographic representations of laboratory objects for educational use. In this regard, we compared four approaches: photogrammetry, a neural radiance field (NeRF)-based method, Gaussian splatting, and LiDAR. These methods were used to generate holographic models of common laboratory items and their fidelity was evaluated by graduate students. Participants assessed the models for shape, color, texture, and visual defects using a repeated-measures design. Across objects, the NeRF-based method produced the most consistently high-fidelity representations, particularly for transparent, reflective, or low-texture items that were difficult to capture with other approaches. Shape and color were generally reproduced more successfully than texture, suggesting that some visual properties remain more challenging to represent accurately in educational holograms. Beyond identifying the strengths and limitations of each reconstruction method, the study demonstrates a practical workflow for creating immersive learning objects that may support pre-laboratory preparation, spatial reasoning, and student engagement in AR/MR-based educational environments. These findings offer design-relevant insights for educators and researchers developing immersive digital learning experiences.
[HC-3] BrailleBench: Investigating Multi-Criteria Braille Comprehension in Large Language Models
链接: https://arxiv.org/abs/2608.27268
作者: Jinghan Zhang,Fengran Mo,Zhiyu Chen,Xiaoyan Han,Kunpeng Liu,Chang-Tien Lu
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注:
Abstract:Although Large language models (LLMs) mediate access to knowledge and computational assistance, their capabilities should benefit vulnerable groups in the same way. However, it is unclear whether existing AI systems are inclusive enough for blind and deafblind users to access the same functionality through Braille, whose indicators, contractions, and digital representations introduce distinct requirements for model comprehension. To this end, we introduce BrailleBench, a benchmark for evaluating LLMs in Braille comprehension from different Criteria. BrailleBench aligns 5,570 instances from five datasets, including mathematics, commonsense, and multi-hop question answering across English and Braille Grades 1 and 2. Different configurations are designed to understand whether the systems can comprehend Braille-authored content, express answers in Braille, and complete end-to-end Braille interaction. To ensure the quality and prevent evaluation bias, the benchmark is built through a deterministic, expert-reviewed pipeline via a self-created Braille Toolkit without using any data instances generated by LLMs. We evaluate six representative LLMs from various aspects. The results reveal a persistent gap between print-English capability and Braille accessibility. Braille understanding and expression are asymmetric, where Grade 2 is especially fragile on the input side compared to Grade 1, and fully Braille requests further reduce performance. The experimental observations provide valuable guidance for the development of future Braille AI systems. All related resources in BrailleBench are publicly available for future research.
[HC-4] A Point-of-Prescription Safety-Check System for Adverse Drug Reactions in Rural Bangladeshi Hospitals: A Feasibility Study
链接: https://arxiv.org/abs/2608.27239
作者: Shahir Abdullah
类目: Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注: This paper is just a pre print, still in working stage, 4 pages
Abstract:Adverse drug reactions (ADRs) are a major, largely preventable source of patient harm. In high-income settings, electronic health records store a patient’s allergy history and warn prescribers when a contraindicated drug is ordered; in rural Bangladeshi public hospitals no such record exists for outgoing patients, a single physician may see on the order of one patient per minute, and a patient’s history of severe reactions does not survive between visits. This paper proposes and outlines the evaluation of a lightweight, smartphone-based safety-check system for this setting. At registration a soft identifier (a phone number) is recorded; after the physician writes a prescription, its image is captured, the brand names are resolved to active ingredients using national drug references, and the ingredients are matched against the patient’s recorded severe reaction history. The system is retrieval-based rather than predictive, and is silent by default, raising a flag only for high-risk matches a design grounded in the alert-fatigue literature. We frame the work as a feasibility study: we describe the proposed framework and an evaluation plan measuring workflow fit under high volume, usability, identity-resolution reliability, and retrospective detection of known reaction cases. We explicitly do not claim a clinical-outcome effect, which the low base rate of severe events places beyond a single-site feasibility study.
[HC-5] Surrounded by Friends: Design and Evaluation of Immersive Layouts of Egocentric Network for Visual Analytics
链接: https://arxiv.org/abs/2608.27194
作者: Kentaro Takahira,Takanori Fujiwara,Wong Kam-Kwai,Kento Shigyo,Leni Yang,Hiroaki Natsukawa,Yalong Yang,Huamin Qu
类目: Human-Computer Interaction (cs.HC); Emerging Technologies (cs.ET); Graphics (cs.GR)
备注:
Abstract:This paper explores design considerations for egocentric network layouts in immersive environments, providing fresh empirical insights that enhance egocentric network analysis. An egocentric network focuses on the topological and semantic relationships around a focal node (ego) and its neighboring nodes (alters), targeting local sub-networks rather than the whole network. Traditional desktop environments, limited by display constraints, often face visual clutter as node numbers grow. Building on recent findings that immersive environments enhance network analysis, we explore layouts tailored for these spaces. We begin by identifying essential design properties and dimensions for egocentric network layouts, taking into account the unique features of immersive environments. Based on these, we design four layouts-Cube, Cylindrical, Radial, and Spherical-that vary across design dimensions. We evaluate these layouts in a user study with 24 participants completing egocentric analysis tasks. Our study suggests that Cube performed well for tasks focused on ego-alter connection strength. In contrast, Spherical was more effective for understanding alter topology, minimizing occlusion, and efficiently utilizing 3D space. These findings inform design implications for future immersive egocentric network layouts.
[HC-6] Soft Active Electromyography Interface for Machine Learning-Enabled Silent Speech Recognition
链接: https://arxiv.org/abs/2608.27048
作者: Yuta Kurotaki,Shusuke Yamakoshi,Reitaro Yoshida,Yutaka Isoda,Tamami Takano,Yuji Isano,Yusuke Miyake,Kentaro Kuribayashi,Hiroki Ota
类目: Machine Learning (cs.LG); Human-Computer Interaction (cs.HC); Sound (cs.SD)
备注: 17 pages, 5 figures, supplementary information
Abstract:Silent speech recognition (SSR) provides an alternative communication pathway in the absence of audible speech. However, conventional approaches are limited by the need for constant facial attachment, privacy concerns, and unstable signal acquisition. Here, we propose a soft, active electromyography (EMG) interface that enables word-level SSR using machine learning. Worn on the hand, the device uses a fingertip electrode that can be positioned near the lips to acquire EMG signals only when needed. The interface integrates liquid metal (LM) interconnects, transparent flexible printed circuit (FPC) electrodes, and elastomer encapsulation to ensure high mechanical stability during finger motion. A deep neural network trained on these stable signals achieved a mean accuracy of 97.2 \pm 1.3% across three subjects in classifying a 30-word vocabulary, demonstrating robust linguistic discrimination. Furthermore, real-time drone control validates the practicality of this approach in noisy and privacy-sensitive environments where conventional voice recognition fails. This study highlights the potential of soft, wearable EMG systems as secure and intuitive human-machine interfaces.
[HC-7] Exploring Normativity in Stable Diffusion: Insights for XAI in the Arts
链接: https://arxiv.org/abs/2608.26980
作者: Michelle Dutoit,Baptiste Caramiaux
类目: Human-Computer Interaction (cs.HC)
备注: In Proceedings of Explainable AI for the Arts Workshop 2026 (XAIxArts 2026) arXiv:2607.20131
Abstract:Generative text-to-image (T2I) systems are increasingly adopted in creative practice, yet their normative behaviors remain underexplored from the perspective of creative practitioners. In this workshop paper, we present a within-subject study with 14 creative practitioners using Stable Diffusion to create illustration from two tasks of differing specificity. We investigate whether and how practitioners perceive normative behavior in a T2I system and how it impacts their creative process depending on task specificity. Our findings show that participants perceived normative behavior through invariant patterns, stereotypical output, and unsolicited omission or addition of details. These experiences led to feelings of disempowerment and creative compromise. We discuss implications for XAIxArts, including prompt transparency and artist empowerment in creative contexts.
[HC-8] Evaluating human and LLM screening workflows in a conceptually complex scoping review: Recall–workload trade-offs and run-to-run consistency
链接: https://arxiv.org/abs/2608.26885
作者: Nikol Figalová,Lynn Huestegge,Anne Böckler-Raettig
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Software Engineering (cs.SE)
备注:
Abstract:Background. Large language models (LLMs) are increasingly used for screening in evidence synthesis, where false negatives can remove relevant studies before full-text assessment. We compared human and LLM title-and-abstract screening workflows in a preregistered study embedded in a conceptually complex scoping review. Methods. After a conservative title-only screen, 1,131 records were screened by one review lead, four trained assistants screening non-overlapping subsets, and seven complete LLM runs using different models and processing configurations, including a nominally identical repeat run. We compared retained workload, operational recall against 316 verified eligible records, agreement, run-to-run consistency, and procedural burden. Because eligibility was verified only for records advanced and assessed in the parent review, recall estimates were operational. Results. No workflow recovered all verified eligible records. The human workflows and two GPT-5.4 file-batch runs retained 42.2-45.0% of records while achieving 82.3-82.9% recall. Gemini 3.1 file batches achieved the highest recall (83.9%) but retained 56.7% of records. All-at-once configurations recovered fewer eligible records than corresponding file-batch configurations. Two nominally identical GPT-5.4 file-batch runs agreed on 91.7% of records but differed on 94 records, including 29 verified eligible records retained by only one run. Discussion. LLM screening performance depended on the implemented workflow, not model identity alone. Processing configuration, workload, record-level variation, and human-LLM decision integration are therefore substantive properties of deployed systems. For high-recall tasks, LLMs are better suited to validated, auditable, human-supervised workflows than autonomous exclusion. Subjects: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Software Engineering (cs.SE) Cite as: arXiv:2608.26885 [cs.AI] (or arXiv:2608.26885v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.26885 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Nikol Figalová [view email] [v1] Thu, 27 Aug 2026 09:42:39 UTC (70 KB)
[HC-9] Dynamic Tree Colors: Adaptive Discriminable Hierarchies with Minimum Instability
链接: https://arxiv.org/abs/2608.26734
作者: Tobias Mertz(1 and 2),Steven Lamarr Reynolds(1),Jörn Kohlhammer(1 and 2) ((1) Fraunhofer IGD, (2) TU Darmstadt)
类目: Human-Computer Interaction (cs.HC); Graphics (cs.GR)
备注: 11 pages, 12 figures, supplemental materials available at: this https URL
Abstract:Hierarchical color maps can support users in the analysis of hierarchical data. For large hierarchies, dynamic color maps can improve discriminability upon user interactions, but the incremental color changes may cause users to lose their orientation in the data set. To address this challenge, we present Dynamic Tree Colors, a dynamic hierarchical color map that can be configured to a suitable tradeoff between discriminability and color stability. We also define quality metrics for both criteria and investigate our algorithm’s performance with respect to these metrics as well as a user study with 18 participants. Our results indicate that Dynamic Tree Colors yields good results in a wide range of application scenarios, but it does not achieve the performance of the state-of-the-art algorithm Cuttlefish in the specific scenario that algorithm was designed for.
[HC-10] RegulAR: Graph-Grounded Error Recognition and Assistance for Procedural Tasks in AR
链接: https://arxiv.org/abs/2608.26715
作者: Yi-Lin Ye,Jindu Wang,Hiu Tung Wong,Shuchang Xu,Huamin Qu,Wong Kam-Kwai
类目: Human-Computer Interaction (cs.HC)
备注: 16 pages, 7 figures. Accepted to UIST 2026
Abstract:Errors are inevitable in procedural tasks, yet most AR guidance systems focus on step-by-step instruction delivery rather than helping users recognize and recover from mistakes. We present RegulAR, an AR task assistant for procedural error recognition and recovery. RegulAR models task instructions as a hierarchical dependency graph and combines this structure with a Multimodal Large Language Model (MLLM) to interpret egocentric observations during execution. This enables RegulAR to track progress, identify deviations by error type, estimate their impact on later steps, and deliver appropriately salient interventions through an in-situ head-up display that visualizes task state and recovery guidance. By making procedural structure explicit, RegulAR supports not only next-step guidance, but also reasoning about what went wrong, why it matters, and how users can get back on track. In a within-subject study (N=12), participants reported better task-structure understanding and recovery support with RegulAR than the MLLM-only baseline.
[HC-11] Processing/p5 Defined through Practice and Learning
链接: https://arxiv.org/abs/2608.26614
作者: Kit Kuksenok,Lee Tusman
类目: oftware Engineering (cs.SE); Human-Computer Interaction (cs.HC)
备注: 10 pages
Abstract:Processing/p5 libraries across different programming languages enact consistent priorities for creative coding as a designed experience. While different programming language ecosystems, like Java and JavaScript, are each associated with their own affordances, community norms, and patterns of use, Processing/p5 sketches across these languages share similarities. Based on case studies of building an implementation of Processing/p5 in two host languages, JavaScript and Lua, we propose a list of software decision-making guiding aspects that constitute Processing/p5, regardless of host language. We discuss this framework in the context of decisions in other exploratory and creative tools that demonstrate how each of the guiding aspects can be operationalized differently than in the case studies. The proposed list highlights opportunities for learning, research, and artistic practice through creation of new Processing/p5 libraries for creative coding and algorithmic art.
[HC-12] EmoSay: Artificial Intelligence-Driven Text-to-Emotional-Speech System for Affective Communication in Extended Reality
链接: https://arxiv.org/abs/2608.26566
作者: Sikiru Ademola Adewale,Sunday D. Ubur,Nikitha Donekal Chandrashekar,Onyeka Emebo,Denis Gračanin
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:While contemporary neural text-to-speech (TTS) systems have achieved high levels of intelligibility, they frequently lack the emotional nuance required for authentic affective communication. This limitation is particularly critical in Extended Reality (XR), where the absence of emotionally expressive audio can diminish user presence and spatial immersion. We present EmoSay, an Artificial Intelligence-driven Text-to-Emotional-Speech (TTES) system designed to bridge the semantic-affective gap in immersive environments. EmoSay modulates a neural synthesis pipeline using discrete emotional prompts, delivering the output through a Unity-based interface featuring high-fidelity spatialized audio. The system was evaluated through a comprehensive user study focusing on perception, engagement, and the subjective sense of empathy. Our results demonstrate that EmoSay significantly enhances the immersive experience, achieving a System Usability Scale (SUS) score of 74.76, indicating strong usability and seamless integration within the XR workflow. Subjective assessments reveal a high degree of perceived naturalness and a strong positive correlation between emotional expressiveness and user engagement. Regression analysis identifies vocal naturalness as the strongest of the tested predictors of user satisfaction, suggesting that EmoSay’s affective prosody helps meet the heightened expectations for realism in immersive settings. This work contributes a scalable, affect-aware framework for inclusive XR design and demonstrates the role synthetic emotion can play in fostering human-computer rapport through voice-first interaction.
[HC-13] Direct Manipulation and Natural Language Programming Together at Last?
链接: https://arxiv.org/abs/2608.26359
作者: Parker Ziegler,David Minh-Duy Cao,Justin Lubin,Sarah E. Chasins
类目: Programming Languages (cs.PL); Human-Computer Interaction (cs.HC)
备注: 32 pages, 5 figures, OOPSLA2 2026
Abstract:Decades of programming languages research has contributed novel approaches to program editing that go beyond modifying text, including direct manipulation programming, structure editing, and automated refactoring tools. However, the rapid growth of natural language programming largely reinforces a view of programs as text and program editing as (unstructured) text transformation. How can we develop unified programming systems that bridge the gap between these approaches, supporting multiple editing paradigms in concert? We take a first step toward answering these questions by introducing a framework that enables program editing via both direct manipulation and natural language, and instantiate this framework in a variant of the \textttcartokit direct manipulation programming system. Our key insight is to treat programs as sequences of structured edits and to use an edit language as a shared interface for both direct manipulation and natural language interactions, leveraging constrained decoding to support the latter. Using our instantiation, we conducted a within-subjects study ( N =18) to understand how the combination of direct manipulation and natural language as editing modalities changes the programming process compared to each modality alone. Perhaps surprisingly, we found that study participants overwhelmingly chose to edit via direct manipulation when both modalities were available, performing just 6.14% of edits via natural language. Our thematic analysis of study sessions revealed that direct manipulation aided task decomposition, encouraged incremental editing, and helped mitigate known challenges in natural language programming related to understanding model capabilities and model-generated code. Our edit-based framework and study findings lay out a possible pathway for future research on programming systems that blend natural language with alternative editing modalities.
[HC-14] Decolonial Discourse in Postcolonial Contexts: How YouTubers Negotiate Audience Tensions Platform Governance and State Influence
链接: https://arxiv.org/abs/2608.26351
作者: Dipto Das,Bryan Semaan
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:Decolonial discourse on online platforms is often framed in terms of creator motivations and expressive possibilities. In this paper, we examine what it takes to sustain such discourse under layered sociotechnical constraints. Drawing on semi-structured interviews with YouTubers engaging in Bengali decolonial discourse, we analyze how audience publics, platform governance, and state influences shape what becomes sayable, visible, and viable. We show how fragmented postcolonial identities among audiences produce legitimacy policing, harassment, and coordinated backlash, requiring ongoing relational labor from creators. At the platform level, differential monetization, opaque moderation, and copyright regimes reorganize which publics are economically viable and reinforce existing hierarchies. Further, intermediaries such as multi-channel networks mediate regulatory pressure, introducing political risks and constraints on participation. In response, content creators engage in strategies of negotiation, including boundary work, infrastructural improvisation, and multi-platform distribution. Overall, our findings highlight the layered dynamics of decolonial discourse in postcolonial contexts and the continuous work required to sustain it in platformed environments.
[HC-15] Kale: A Transformation-Safe Spreadsheet System
链接: https://arxiv.org/abs/2608.26345
作者: Michael Coblenz,Jacob Yim,Ajinkya Bokade,Mounika Padala,Julia Epshtein,Priyanka Bhatia,Piyush Chauhan,Simran Gill,Aniket Gupta,Grishma Gurbani,Vaibhav Khetan,Arushi Munjal,Jeffery Tung,Joanna Yang
类目: Human-Computer Interaction (cs.HC); Software Engineering (cs.SE)
备注:
Abstract:Spreadsheet formulas can refer to rectangular ranges of arbitrary size. When a user changes the structure of a referenced table, the spreadsheet system updates the references to refer to a new range. Unfortunately, this new range may differ from the user’s expectations, introducing bugs in spreadsheets. We describe a user study showing that standard reference semantics are error-prone, resulting in significant risk to users. We introduce Kale, a prototype system that eliminates the risk of inserting these kinds of bugs by restricting the kinds of references that can be expressed. We show that Kale can be used effectively by users to complete tasks that are error-prone in traditional spreadsheet systems. Finally, we describe a corpus study that evaluates the extent to which the reference restrictions in Kale might have implications on users.
[HC-16] Calibration-Free Cuffless Blood Pressure Estimation Using Multimodal ECG-PPG Fusion on a Google Pixel Watch
链接: https://arxiv.org/abs/2608.26325
作者: Jathushan Kaetheeswaran,Boyi Ma,Ali Abedi,Shehroz S. Khan,Milad Lankarany
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Inadequate blood pressure (BP) monitoring and management outside of clinical settings can worsen major cardiovascular risk factors such as hypertension. While cuff-based devices are commonly used for at-home monitoring, these devices can be inconvenient for daily use due to their sensitivity to body positions, upper-arm constrictions, and limited portability. A promising alternative is emerging in the form of consumer-grade smartwatches, where physiological signals related to cardiac activity can be used to estimate BP non-invasively and continuously across daily living conditions. In this work, we use data collected from a Google Pixel Watch in 40 participants to develop and compare several algorithm approaches for BP estimation. We found that our proposed deep learning model achieved the strongest overall performance, and that fusing smartwatch signals with demographic information improved model generalizability to unseen individuals. However, we also identified that model accuracy was not consistent across participant subgroups, with obese individuals yielding higher estimation errors than others. This study highlights the feasibility of consumer-grade smartwatches as accessible platforms for deploying robust BP estimation algorithms, though clinical reliability will require larger, more diverse populations and additional sensing modalities.
[HC-17] “A Second Set of Eyes”: The Process and Challenges of Software Documentation Review
链接: https://arxiv.org/abs/2608.26232
作者: Avinash Bhat,Ian Arawjo,Disha Shrivastava,Jin L.C. Guo
类目: oftware Engineering (cs.SE); Human-Computer Interaction (cs.HC)
备注: Accepted at GROUP 2027, 30 pages
Abstract:Organizations assign documentation work to technical writers, yet the knowledge required to produce it is distributed across developers, managers, and other practitioners. Prior work has established quality criteria for judging “good” documentation, but it has not examined how practitioners bring that expertise to improve documentation quality or the challenges they face in doing so. Through semi-structured interviews with experienced technical writers ( n=31 ) from different organizations, our work reveals the individual and collaborative effort required to maintain documentation quality. We identify five distinct stages of the documentation review process: self review, technical review, editorial review, play testing, and post-publication feedback. Each stage draws on practitioners with distinct expertise to address quality across content, presentation, and user experience. Our findings surface organizational and technical challenges writers face in recruiting expert reviewers, navigating development timelines, and contending with tools not specifically designed for documentation workflows. Our work positions documentation review as a crucial yet understudied site of collaborative work and opens new research and design directions for process improvement and tool support.
[HC-18] Can You Say This for Me? Speaking Up by Proxy in Co-Located Discussion
链接: https://arxiv.org/abs/2608.26185
作者: Yue Shen,Rehema Abulikemu,Ryan P. McMahan,Yan Chen
类目: Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET); Human-Computer Interaction (cs.HC)
备注: Accepted to ACM UIST 2026, Detroit, MI, USA
Abstract:Equal participation in co-located discussion is important for effective collaboration, yet people often hold back when they anticipate negative interpersonal or professional consequences, especially when raising a point requires voicing it themselves. We present SecondVoice, a mixed-reality system that enables people to speak up through an embodied virtual proxy. By separating what is said from who says it, SecondVoice brings hesitant points into the live spoken discussion without putting the speaker on the spot. Using a private overlay, users specify their intent through a structured specification process rather than composing a full utterance. The system reformulates the input and voices it into the conversation through the proxy. We characterize a design space of participation channels under social risk. In a preliminary within-subject study (N = 16), we compare the complete SecondVoice system with an anonymous text-board channel across two group discussion tasks. Half of participants reported using SecondVoice for a point they did not say aloud, compared with 18.8% for the text board. Proxy-delivered points entered the spoken floor and were followed by multi-turn group engagement, which we did not observe after text-board posts. Participants described the channel as situationally valuable but identified tradeoffs around timing, ownership, and trust in reformulation.
[HC-19] utorTrace: A Dataset and Taxonomy for Classifying Learner Behavioral States during AI-Assisted Programming Education
链接: https://arxiv.org/abs/2608.26184
作者: David Barron,Xiaohang Tang,Rezky Dwisantika,Minsun Kim,David H. Smith IV,Jiaming Cui,Yan Chen
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: Accepted to ACM UIST 2026, Detroit, MI, USA, 14 pages, 5 figures Dataset: this https URL
Abstract:AI programming tutors provide scalable support, yet lack the behavioral context human tutors rely on to adapt support to learners’ needs. We present TutorTrace, a dataset and behavioral abstraction pipeline that makes learners’ behavioral context visible and computable in real time from low-level IDE telemetry. Across four deployments in two introductory Python courses (N=480), TutorTrace captures approximately 180K telemetry events, 13,633 behavioral segments, and 27 continuously computed metrics. From this foundation, we derive a taxonomy of learner activity before the first AI query, between consecutive queries, and across the full session, enabling systems to respond not just to what learners say, but to what they have done leading up to the help-seeking moment. In a preliminary classroom evaluation, behavior-aware prompts were associated with a decrease in intervals between queries with no independent work from 50.0% to 20.7%. As an additional demonstration of downstream utility, we evaluate TutorTrace on two held-out prediction tasks: whether a learner will query within the next 60 seconds (AUROC=.726) and whether an upcoming query reflects guided or dependent help-seeking (AUROC=.717). Together, these findings show how behavioral context can enable adaptive AI tutoring at scale.
[HC-20] Why did My Robot Just Change Personality? Prompting Guidelines for a Grounded Robot Persona in LLM -Based HRI
链接: https://arxiv.org/abs/2608.26182
作者: Ashita Ashok,Franziska Babel,Patrick Holthaus,Rucha Khot,Karla Bransky,Fethiye Irmak Dogan,Karsten Berns,Silvia Rossi,Minha Lee,Guy Laban
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Robotics (cs.RO)
备注: Accepted for publication at the 35th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN 2026)
Abstract:Large language models (LLMs) are increasingly used for verbal interaction in social robots, yet prompt design in human-robot interaction (HRI) remains underspecified. As a result, robots may present hallucinated capabilities, unclear behavioural boundaries, and misleading personas. This paper develops a framework for prompt design in LLM-based robots and introduces a structured prompt template comprising eight functional components through which robot behaviour can be specified, bounded, and adapted. The framework is grounded in a review of prior LLM-based HRI work and complemented by survey and discussion data from HRI experts gathered at the Robo-Identity workshop at IEEE RO-MAN 2025 (N=27). The qualitative findings highlight limited legibility of robot personality, the need for user adaptation, and strong ethical concerns about safety, deception, and governance. Based on these findings, we present prompting guidelines accompanied by proof-of-concept template as a structured design and reporting aid for HRI research. We argue that prompt design should be treated as a socio-technical problem rather than a minor implementation detail, requiring explicit capability boundaries, transparent behavioural assumptions, and context-sensitive safeguards to support reliable and interpretable HRI.
[HC-21] Mitigating Fabrication in Multi-Stage LLM Pipelines for Hiring: An Empirical Evaluation of Prompt Guardrails and Human-in-the-Loop Checkpoints
链接: https://arxiv.org/abs/2608.26171
作者: Hiroko Takano
类目: Computers and Society (cs.CY); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注: 13 pages, 2 figures
Abstract:Multi-stage LLM hiring pipelines (resume improvement, interview question generation, answer feedback) can fabricate credentials, inflate qualifiers, and invent experience. We evaluate two mitigations, prompt guardrails and human-in-the-loop (HITL) checkpoints, against a fully automated baseline. In a controlled experiment (10 synthetic resumes x 2 job descriptions x 3 repetitions x 3 conditions; 180 runs), the baseline (C1) produced at least one unsupported claim in 96.7% of outputs (mean 6.80 findings/output). Prompt guardrails (C2) reduced finding density by 86% (6.80 to 0.92/output), but 50.0% of outputs still contained a fabrication, showing prompt-level mitigation alone is insufficient. A human checkpoint after resume improvement (C3) eliminated all identity fabrications, reduced finding density by 59% (6.88 to 2.82/output), reduced item-level fabrication from 96.7% to 75.0% (p=.022), and cut capture of JD-embedded trap requirements from 47% to 2% (vs. 5% under the guardrail). An exploratory analysis of multi-specialty resumes shows contamination rising monotonically with domain distance between specialties, suggesting career changers are especially exposed. The reviewer in this study caught all flagrant fabrications, but subtle qualifier drops and plausible new claims survived review roughly half the time (54.5% removal). Neither mitigation degraded the deliverable: claim retention exceeded 99% under both. The interventions are complementary: the guardrail eliminates unprompted additions and qualifier inflation cheaply, while the checkpoint gives near-categorical guarantees against the most severe failures, invented identities and JD-baited claims. These results support a layered architecture combining guardrails with a human checkpoint. A supplementary run with a newer-generation model (90.0% baseline fabrication rate) suggests the problem is not resolved by model progress alone.
[HC-22] Improving LLM Interpretability with User-Centric Chain-of-Thought Reasoning
链接: https://arxiv.org/abs/2608.26166
作者: Philipp Schröppel
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:Advancing reasoning capabilities allow large language models (LLMs) to tackle increasingly complex problems, while reasoning traces - intermediate steps toward solutions - open up high-stakes applications by enabling human inspection of AI decision-making. However, current approaches prioritize model performance over human interpretability, limiting effective human-AI collaboration. In this study, we design and evaluate a human-centered approach that structures reasoning traces based on self-contained, verifiable steps, enabling users to independently assess and correct AI reasoning. Our approach uses XML-like tags to encode reasoning content and metadata, facilitating targeted feedback. Evaluation on mathematical reasoning tasks shows our approach maintains equivalent performance to standard Chain-of-Thought reasoning while enhancing interpretability. User studies demonstrate significant improvements in perceived usefulness and ease of use. This work advances understanding of how user-centric design of LLM outputs can better serve human collaboration needs in high-stakes AI deployments.
[HC-23] Artificial Intelligence Models Can Predict and Collaboratively Modulate Human Memory Search
链接: https://arxiv.org/abs/2608.26152
作者: Eric Lacosse,Mariana Duarte,Graham Todd,Peter M. Todd,Daniel C. McNamee
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: 18 pages, 5 figures; includes Supplementary Information
Abstract:Large language models (LLMs) exhibit unprecedented natural language generation and many text-based problem-solving capabilities. Indeed, in many language-based tasks, for example routine coding, these artificial intelligence models have reduced, or even eliminated, the need for human input. But rather than replacing human cognitive effort, LLMs may instead serve as cognitive tools to extend human abilities, particularly when they are engaged in a task requiring open-ended conceptual exploration and creative ideation. However, we are yet to understand how these models may enhance such generative human cognitive abilities in human–AI interactions. In this study, we explore and evaluate the ability of LLMs to follow and enhance human mental trajectories during semantic memory search. To test this, we use the semantic fluency task (SFT), a classic cognitive paradigm requiring generative semantic memory retrieval that has long served to characterize convergent and divergent thinking in humans. We demonstrate that an LLM’s abilities to track and predict human memory trajectories in this task exceed those of other humans.
计算机视觉
[CV-0] UrbanGround: From Local Perception to Spatial Agency in a Real-Scale City
链接: https://arxiv.org/abs/2608.27456
作者: Tianjie Ju,Zheng Wu,Yueqing Sun,Yuhan Cui,Bobo Li,Shengqiong Wu,Pengzhou Cheng,Haodong Zhao,Zongru Wu,Xinbei Ma,Doris Zhang,Kunling Li,Mong-Li Lee,Wynne Hsu,Hao Fei,Qi Gu,Gongshen Liu,Zhuosheng Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 35 pages, 11 figures, 7 tables. Project Page: this https URL , Code Repository: this https URL
Abstract:Multimodal large language models (MLLMs) can interpret a street view, but urban agency depends on whether such local evidence remains useful after the agent starts to move. In this paper, we investigate how far current MLLM agents can turn local urban perception into reliable action in a complicated real-scale city. We propose UrbanGround, the first sandbox to make this question testable in a physically constrained replica of Hong Kong built from territory-wide 3D geospatial data. UrbanGround supports closed-loop interaction from a first-person view and provides an interactive map for navigation. Agents can directly enter the 3D city and explore from a first-person view. Our analysis follows the growth of the spatial problem through three research questions. We first test whether an agent can ground a local scene well enough to answer spatial questions after active observation. Then we ask whether that grounding supports navigation as destinations become farther away and less explicit. Finally, we examine whether the resulting behavior survives changes in route availability and pedestrian motion. Contemporary MLLM agents usually show useful atomic abilities in visual recognition and short-range spatial reasoning, while orientation and pedestrian-aware movement remain unreliable. Their central failure emerges over extended exploration, where local abilities do not compose into sustained goal-directed behavior and errors accumulate without effective correction. We hope UrbanGround will support broader study of how far current MLLM agents can explore reliably in complex, open-ended urban environments.
[CV-1] Retrieval Heads Meet Vision: Uncovering How VLMs Locate and Extract Visual Information
链接: https://arxiv.org/abs/2608.27417
作者: Chanho Park,Daehyeon Choi,Jihyun Lee,Minhyuk Sung
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-language models (VLMs) can locate an image region referred to by a text prompt and route the corresponding visual evidence to the output, yet the internal mechanism behind this behavior is not understood. Inspired by retrieval heads in large language models, we ask whether VLMs contain an analogous mechanism for visual retrieval. We answer affirmatively by introducing Visual Retrieval Heads (VRHs), a small subset of attention heads (about 1.7-2.6%) that are causally responsible for grounding text descriptions to image regions. To find them, we recast existing head-scoring methods under a unified design space over query tokens, key aggregation, and cross-sample aggregation. We then show that scoring attention from output prediction tokens with a sum over the ground-truth referent region most reliably identifies causal heads. Across eleven VLMs and five referring-expression benchmarks, masking only the top 20 VRHs reduces grounding accuracy by up to 80 percentage points, while masking the same number of random heads has little effect. Beyond replicating the causal-sparse-universal triad established for text retrieval heads, VRHs exhibit several properties not previously reported: they generalize across visual reference tasks, remaining causal on attribute, spatial, counting, and visual-math benchmarks despite being discovered through bounding-box prediction; they are functionally specific, preserving output format while corrupting localization; and they are architecturally shared, transferring causally across VLMs that share an LLM backbone but differ in vision encoder, projector, and instruction tuning.
[CV-2] Reconstructing Humans and Objects in Interaction using Large Reconstruction Models ECCV2026
链接: https://arxiv.org/abs/2608.27407
作者: Agniv Chatterjee,Georgios Pavlakos
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at ECCV 2026. Project Page: this https URL
Abstract:Estimation of Human-Object Interactions in 3D (3D HOI) is a fundamental problem in 3D computer vision with applications in AR/VR, robotics, and embodied AI. However, reconstructing these interactions in 3D remains challenging due to depth ambiguities, occlusions, and object shape variability. Existing approaches are primarily concerned with reprojection and contact constraints, fitting parametric human models and object templates to 2D images. In this paper, we explore a different avenue. We present MILO, a framework that leverages the visual capabilities of Large Reconstruction Models (LRMs) to recover detailed 3D human-object interactions from a single image. Our key observation is that LRMs provide a powerful geometric scaffold that preserves relative human-object arrangement and proximity cues. This significantly simplifies the reconstruction procedure, reframing the problem as interpreting the LRM mesh: we segment it into human and object components, fit a parametric body model to the human part, and optionally align an object template to the object part (if such a template is available). MILO achieves strong reconstruction accuracy and outperforms existing baselines across multiple benchmarks and interaction scenarios. Our code is available at this https URL.
[CV-3] CLAP: Cross-Embodiment Video World Models are Zero-Shot Physical Simulators
链接: https://arxiv.org/abs/2608.27406
作者: Kechen Liu,Ola Shorinwa
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:State-of-the-art action-conditioned video models are typically restricted to a single robot embodiment, preventing them from leveraging the vast corpus of heterogeneous video data that contains rich signals for learning generalizable physics. To bridge this gap, we introduce CLAP, a framework for cross-embodiment action-conditioned video generation capable of being trained on diverse, internet-scale videos across human and robotic agents. CLAP is grounded in the insight that universal physical laws govern spatiotemporal dynamics regardless of the actor. However, cross-embodiment learning is non-trivial because action representations vary sharply across robot platforms and are typically absent in human videos. CLAP addresses this fundamental challenge through the following core contributions. First, CLAP reconciles disparate action spaces using end-effector poses, language instructions, and latent actions. Second, to resolve their individual limitations, CLAP introduces a curriculum-based cross-embodiment learning recipe that first learns foundational physical priors across unlabeled video data using latent actions and subsequently grounds them in end-effector action spaces for zero-shot deployment to real-world tasks. Crucially, CLAP approaches or surpasses state-of-the-art single-embodiment video models in challenging environments like DROID. These performance advantages compound via few-shot adaptation to establish a novel paradigm for training single-embodiment video world models. Ultimately, CLAP delivers the most comprehensive suite of action-conditioned video world models to date - spanning diverse action-conditioning spaces (end-effector, language, and latent) and robot morphologies (including cross-embodiment, DROID, Bridge, bimanual YAM robots, and G1 humanoids). We open-source all code and models. Project Website at this https URL .
[CV-4] LeVJEPA: Efficient Scalable Video Pretraining without the Heuristics
链接: https://arxiv.org/abs/2608.27395
作者: Lukas Kuhn,Lucas Maes,Giuseppe Serra,Quentin Le Lidec,Yann LeCun,Randall Balestriero,Florian Buettner
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Video carries the temporal structure of the physical world, yet learning representations from it has remained computationally expensive: prevailing self-supervised methods either prevent representation collapse through architectural asymmetries, coupling an exponential-moving-average target encoder, a stop-gradient, and a capacity-limited predictor, or circumvent it by reconstructing masked content in pixel space. We introduce LeVJEPA, the first video encoder trained under LeJEPA’s collapse-free objective, which dispenses with both. A single encoder is trained with an invariance loss over global and local views of a clip, regularized by SIGReg, which excludes collapse with a provable guarantee. The architecture reduces to an encoder and a projector, and the objective to a single hyperparameter. This formulation admits two properties. First, the cost of pretraining is governed by the number of tokens the encoder observes; uniform random token dropping renders this number small while simultaneously improving downstream accuracy. At matched epochs on identical data, LeVJEPA matches or surpasses V-JEPA 2 across ViT-S/B/L at 5.6 to 20.8x less pretraining compute, and at matched total FLOPs it exceeds the strongest video baseline by 7.6 points on ImageNet-1K while remaining competitive on motion-centric benchmarks. Second, since no asymmetry between branches is required, the encoder can be trained with block-causal attention at no measurable accuracy cost: temporal ordering becomes a property of the encoder itself. Against a compute-matched DINOv2 trained on frames of the same videos, LeVJEPA approaches the image-pretrained encoder on appearance-centric evaluation while nearly doubling its motion-centric accuracy. These results indicate that, once its computational overhead is removed, video becomes a viable and in several respects preferable substrate for general-purpose visual pretraining.
[CV-5] Successive Capacity Growth: Task-Complexity-Driven Width and Depth Expansion for Vision Transformer Encoders in JEPA World Models
链接: https://arxiv.org/abs/2608.27367
作者: Frederik Berenz
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 12 pages, 2 figures, 6 tables
Abstract:Joint-Embedding Predictive Architectures (JEPAs) for world modeling typically employ fixed-size Vision Transformer encoders that are over-provisioned for simple tasks and under-provisioned for complex ones, with significant redundancy across attention heads. We propose Successive Capacity Growth (SCG), a method that starts from a minimal encoder (1 head, 2 layers, 283K parameters) and grows incrementally in width (adding attention heads for low-level semantic capacity) or depth (adding transformer blocks for higher-order semantic abstraction), driven by a task-agnostic test-and-verify mechanism that exploits function-preserving expansion to safely trial architectural changes and roll back if they do not improve prediction loss. The Sketched Isotropic Gaussian Regularizer (SIGReg) ensures that all learned semantic dimensions remain statistically independent and aligned with the predictive objective, preventing collapse even as the architecture grows. On a 60-dimensional multi-object dynamics task, SCG naturally triggers depth expansion, improving prediction loss by 20.3% over the fixed small baseline with 56 times greater parameter efficiency than scaling to the fixed large model; on a 2D navigation task, a single width expansion yields even an 23% improvement over the fixed large model. Across all three tested environments of increasing complexity, the adaptive encoder matches or exceeds the fixed small baseline, with zero false-positive expansions and bit-exact function preservation (ratio = 1.0, absolute difference = 0.0). The take-away is that JEPA world model encoders need not be pre-allocated at maximum capacity - they can grow successively as the task demands, achieving significant compute and data efficiency while maintaining representation quality.
[CV-6] KnockGS:interaction-Grounded Calibrationof Physical Gaussian Representations
链接: https://arxiv.org/abs/2608.27365
作者: Chenchen Ge,Hanwen Shen,Bowen Jing,Jiyuan Cai,Xiaofeng Wang,Hongsen Lei,Weitao Zhou,Dandan Zhang,Haibao Yu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Physics-integrated 3D Gaussian representations now allow reconstructed deformable objects to be simulated and rendered under explicit material models. Existing pipelines, however, assume that material parameters are known or manually specified, limiting their applicability when these parameters must be inferred from observed object dynamics. We propose KnockGS, an interaction-response PhysicalGS framework that estimates the elasticity and density scales of a 3D Gaussian object from its dynamics under a known applied force. Rather than treating physical simulation only as a forward process, we turn the force-induced response into a calibration signal: temporal response features are xtracted from the observed dynamics, the two material scales are estimated from those features, and the estimate is then frozen and written back into the same simulator so that it can be tested on an interaction it was never fitted this http URL evaluate the framework on both parameter recovery and response-level fidelity. The estimated scales are compared against hidden ground truth, and the re-simulated object is measured against the target using 3D particle trajectories, response-curve statistics, and rendered-frame quality. Across five held-out material targets, our method recovers the scales substantially more accurately than response retrieval, global regression, or a fixed default material, and the frozen estimate remains predictive under interactions that differ in direction and in magnitude. Interaction response therefore carries enough information to calibrate material scales in physically grounded 3D Gaussian this http URL study is a first step toward interactive PhysicalGS systems that calibrate a Gaussian asset whose rendered appearance and simulated response are consistent.
[CV-7] PAWBench: How Far Are We from Probabilistically Aligned World Modeling?
链接: https://arxiv.org/abs/2608.27345
作者: Yuandong Pu,Le Zhuo,Sayak Paul,Gabriel Jorge Menezes,Avram Đorđević,Shiyang Li,Yifan Zhou,Bin Fu,Wenlong Zhang,Junjun He,Yu Qiao,Yihao Liu,Jingbo Xing,Xi Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent video generation models are increasingly framed as world models. Many physical processes can unfold in more than one valid way. Therefore, a world model should reproduce not only a plausible trajectory, but also the distribution of possible behaviors under the same initial observation and action. We call this distribution-level requirement probabilistic alignment. However, existing evaluations largely assess individual-video plausibility and do not test whether repeated generations recover the correct distribution. This raises a central question: how far are current video generators from probabilistically aligned world modeling? To answer it, we formalize probabilistic alignment as a distributional criterion for world models and introduce PAWBench, a benchmark for evaluating video generators as stochastic samplers of world dynamics. We further introduce PAWEval, an outcome-level protocol that converts repeated video rollouts into empirical distributions over possible physical behaviors. Across 50 scenarios and eleven current systems, no model consistently matches the reference probabilities while recovering the range of valid behaviors. Having established this gap, we test whether language prompts, initial noise sampling, or model training can reshape the model’s predictive distribution. We believe our work can serve as a foundation for future efforts to move towards probabilistically aligned world modeling.
[CV-8] R2M-Bench: Evaluating Revisit Memory via Relative Consistency in Interactive Video World Models
链接: https://arxiv.org/abs/2608.27328
作者: Qiwen Gu,Bingjie Gao,Rui Chen,Geng Li,Jifan Li,Qishuai Wen,Li Niu,Jing Tang,Xiangxiang Chu,Junqiao Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Code: this https URL
Abstract:High similarity between first-visit and return frames does not necessarily show that a video world model remembered the scene; the intervening rollout may simply have changed very little. This ambiguity makes absolute revisit scores sensitive to rendering stability, repetitive content, and failed motion. We introduce \emphR2M-Bench (\textbfRelative \textbfRevisit \textbfMemory Benchmark), a benchmark of observable revisit-selective consistency. For every detected return, R2M-Bench compares the revisit pair with two controls from the same rollout: a gap-matched non-revisit pair that measures generic temporal stability and a short-range pair that estimates short-horizon consistency. These comparisons produce \emphMemoryGain (MG), the revisit advantage over the temporal baseline, and the \emphNormalized Memory Ratio (NMR), which normalizes this advantage by the short-to-baseline dynamic range. R2M-Bench combines 100 reference scenes with three leave-and-return trajectories to form 300 instances and evaluates appearance fidelity, scene and object identity, local geometry, and persistent state. Across seven action-conditioned video world models, Overall NMR correlates with human consistency judgments at Spearman’s \rho=0.547 (95% CI [0.45,0.63] ). Its within-model correlation magnitude with generated motion is 0.072 , compared with 0.207 for raw revisit similarity, indicating that relative calibration substantially reduces the slow-motion shortcut. DreamX-World-Memo achieves the highest Overall NMR among the evaluated video models. Together, these results support same-rollout relative calibration as a practical way to distinguish revisit-specific consistency from generic temporal stability.
[CV-9] Detection of Christmas tree plantations from high-resolution aerial imagery. A case study in the French Morvan
链接: https://arxiv.org/abs/2608.27290
作者: Francesca Razzano,Emanuele Dalsasso,Adrien Baysse-Lainé,Silvia Liberata Ullo,Gilda Schirinzi,Jocelyn Chanussot
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 9 figures. Submitted to IEEE JSTARS; currently under revision
Abstract:Christmas tree plantations are economically relevant, yet a largely unexplored application domain in Remote Sensing (RS). Their delineation is challenging because of high planting density, short rotation cycles, visual confusion with surrounding vegetation, the availability of dense labels for one reference year only, and severe class imbalance at the landscape scale. Although Deep Learning (DL) methods have shown strong potential for vegetation mapping, existing approaches are typically designed for forests, generic plantation systems, or orchards, and do not explicitly address the structural specificity and hard-negative confusion that characterize Christmas tree plantations. In response to these challenges, this work makes three main contributions: (i) it frames Christmas tree plantation mapping as a distinct rare-target semantic segmentation problem; (ii) it introduces a Hard Negative Mining (HNM) strategy to improve discrimination against confusing background patterns; and (iii) it evaluates the proposed framework across complementary levels, including supervised testing, temporal transfer, and large-scale validation. On the 2020 test set held out, the best model, DeepLabV3 with a ResNet-34 encoder, achieves an IoU of 0.733 and an F1-score of 0.846. HNM substantially improves precision-recall behavior, increasing the area under the precision-recall curve from 0.204 to 0.913. Temporal inference further shows meaningful transferability, reaching IoU/F1 values of 0.751/0.858 on 2017/2018 and 0.691/0.817 on 2023. Large-scale validation further highlights the intrinsic difficulty of the task, as Christmas tree plantations occupied only a very small fraction of the extent of the common evaluation, corresponding to 1,498.4 ha (1.72%) in 2017/2018 and 1,782.2 ha (2.04%) in 2023 out of 87,309.4 ha in total.
[CV-10] ADP: Task-Aware Deformable Prediction for Single-Stage 3D Object Detection
链接: https://arxiv.org/abs/2608.27282
作者: Su Wang,Yaochen Li,Min Yang,Jiaohao Nie,Chang Liu,Yuehu Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO); Systems and Control (eess.SY)
备注: Accepted to the 2023 IEEE Intelligent Vehicles Symposium (IV 2023)
Abstract:Most single-stage 3D object detectors complete different tasks with the same extracted features. Nevertheless, it is impossible to project features into a common space that is adaptive for all the tasks. We present a novel task-aware deformable prediction (TADP) method for single-stage 3D object detection to solve this problem. Firstly, a triple feature refinement aggregation module is designed to extract three-level features adaptively. Additionally, we design the multi-scale feature aggregation block to fuse multi-scale features in a scale-aware manner. Finally, the prediction of each task is deformed with the designed plug-and-play task-aware deformation head. It can percept the emphasis and interaction of each task. We also designed three different deformation modules. The experimental results demonstrate that the proposed deformation head shows good results on other detection methods. The experimental results on the KITTI dataset demonstrate that the car mAP is 80.91%, surpassing many state-of-the-art methods on the KITTI benchmark.
[CV-11] Sidecar: Training-Free Semantic Reuse for Character-Consistent Free-form Visual Storytelling
链接: https://arxiv.org/abs/2608.27280
作者: Sibo Dong,Sarah Adel Bargal
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Visual storytelling requires generating images that follow a narrative while preserving consistent character identities across frames. In free-form story generation, a character is fully described only when first introduced and is later referred to by a type-level mention or pronoun. Although this setting better reflects natural storytelling, later prompts may omit important identity-related semantics, making character consistency more difficult to maintain. We propose \textbfSidecar, a plug-and-play semantic augmentation module that preserves entity-level information from the initial description and injects the missing semantics into later prompt embeddings. Sidecar requires no additional training and does not modify the architecture of the base diffusion model. Experiments on FreeStoryBench show that Sidecar consistently improves prompt-image alignment and character consistency across multiple SDXL- and FLUX-based baselines, with negligible computational overhead.
[CV-12] Decoupled I/O-Dominant Pipelines for Large-Scale Whole-Slide Image Embedding Extraction
链接: https://arxiv.org/abs/2608.27278
作者: Mayanka Chandrashekar,Xi Zhang,Ethan Seefried,Tirthankar Ghosal,John Gounley,Heidi Hanson
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Whole-slide images (WSIs) are central to computational pathology but are prohibitively large, making patch-based processing the practical unit for foundation model inference. At scale, however, generating and handling massive numbers of patches on quickly introduces significant I/O and orchestration overhead, often dominating end-to-end performance. We present a decoupled, I/O-aware pipeline for large-scale WSI embedding extraction that decomposes the workflow into three stages: (1) patch generation and staging, (2) embarrassingly parallel embedding inference, and (3) sharded vector database ingestion. This design isolates data movement from compute, enabling efficient patch delivery, scalable multi-node inference with minimal communication. The resulting system produces a distributed vector database where embeddings are persistently coupled with rich metadata (e.g., patient, slide, and patch attributes), enabling efficient filtering, retrieval, and downstream reuse. This representation database is compact and reusable for tasks such as retrieval, classification, and few-shot learning, particularly benefiting low-resource environments. We show that decoupling I/O, computation, and ingestion enables high-throughput WSI embedding extraction at scale. By characterizing the scaling envelope, we demonstrate that storage dominates beyond moderate concurrency, reframing WSI embedding extraction as a data-centric systems problem rather than a purely compute-bound workload.
[CV-13] UniFLM: United Segmentation and Measurement on Fetal Limb Ultrasonic Image
链接: https://arxiv.org/abs/2608.27240
作者: Zeen Zhou,Qiuhua Chen,Xiaojun Cao,Changmao Chen,Chao Sun,Bo Du
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Prenatal ultrasound examination is crucial for assessing fetal limb development and detecting congenital anomalies. However, existing artificial intelligence models often overlook fetal lethal skeletal dysplasias due to the lack of high-quality annotated data and a unified framework for multiple long bones. Moreover, generic segmentation models struggle with the inherent noise and semantic gaps in ultrasound images. To address these challenges, we construct the Fetal Limb Bones (FLB) dataset, comprising high-quality annotations for the humerus, femur, tibia-fibula, and radius-ulna. Furthermore, we propose UniFLM, a unified framework for automatic cross-plane segmentation and measurement. UniFLM incorporates a Semantic-Aware Skip Connection module to bridge the semantic gap between encoder and decoder features, and a Positive Sampling strategy to adaptively filter noise and extract essential semantic information. Finally, a Point Regression Mapping module is introduced to learn clinician annotation patterns for precise bone length measurement. Extensive experiments conducted on the FLB dataset demonstrate that the proposed UniFLM achieves superior accuracy and enhanced generalization capabilities in fetal long bone assessment compared to current state-of-the-art models.
[CV-14] DINOcular: Self-Supervised Visuospatial Representations
链接: https://arxiv.org/abs/2608.27226
作者: Farkhat Almukhamedov,Sami Azirar,Hermann Blum
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We introduce a self-supervised framework for learning joint visuospatial representations from RGB-D observations. While modern vision foundation models are trained almost exclusively on RGB images, many embodied systems have access to explicit depth sensing, which provides geometric information that monocular inputs cannot recover. Our method integrates depth-derived geometric priors with a visual backbone through inter-patch and intra-patch fusion, enabling the model to encode both appearance and spatial structure efficiently. The resulting representation shows promising improvements on 3D awareness while preserving semantic transfer: it outperforms prior methods of comparable scale on multiple 3D geometry benchmarks, and remains competitive when probed for standard RGB-D semantic segmentation tasks.
[CV-15] CODE: Cross-Modal Calibration and Dynamic Suppression for Open World Object Detection
链接: https://arxiv.org/abs/2608.27214
作者: Hao Xu,Zhaoning Shi,Hehe Jin,Bo Ma
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by ACM Multimedia 2026 (MM '26)
Abstract:Open World Object Detection (OWOD) built on multimodal foundation models often suffers from semantic ambiguity caused by unidirectional text-to-vision matching, while rigid outlier penalties may over-suppress unknown objects near known-class decision boundaries. We propose CODE (Cross-Modal Calibration and Dynamic Suppression), a unified inference-time framework with three complementary components. Cross-Modal Joint Confidence Calibration injects global visual prototypes to calibrate text-driven known-class predictions. Uncertainty-Guided Universal Objectness Enhancement measures classification hesitation from local visual responses to strengthen potential unknown objects. Dynamic Outlier Suppression via Confidence Margin replaces rigid suppression with a margin-aware adjustment that preserves ambiguous out-of-distribution instances. Experiments on the Real-World Detection benchmark demonstrate that, with the OWL-ViT L/14 backbone, CODE achieves 21.7 U-mAP and 40.8 K-mAP in Task 1, surpassing the previous state of the art by 2.6 and 2.3 points, respectively.
[CV-16] PACE: A Unified Condense-and-Extract Paradigm for Fast VLM Inference EMNLP2026
链接: https://arxiv.org/abs/2608.27206
作者: Junjie Liu,Shengyuan Ye,Xu Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 22 pages, 9 figures, 13 tables. Accepted to Findings of EMNLP 2026
Abstract:Vision-Language Models (VLMs) demonstrate exceptional visual reasoning capabilities, yet their inference costs escalate rapidly with the proliferation of visual tokens. Existing visual token pruning methods exhibit two fundamental limitations. First, most approaches operate exclusively post-vision encoder, leaving the substantial latency of the visual encoding phase unoptimized. Second, under strict token budgets, these methods often fail to jointly preserve holistic visual contexts and fine-grained details, leading to performance degradation. To address these bottlenecks, we propose PACE (Pixel-Adaptive Condense and Extract), a training-free inference framework that accelerates both the vision encoder and the Large Language Model (LLM) via a unified Condense-and-Extract paradigm. During the Condense stage, an Adaptive Pixel Compressor (APC) evaluates visual information density prior to encoding, adaptively downsampling redundant inputs, curtailing encoder computation while preserving global context and essential visual cues. In the Extract stage, a Dynamic Dual-Attention Extractor (DDAE) selectively retains visual tokens via a fusion of internal visual signals from the encoder and semantic signals from the LLM, safeguarding task-critical details. By integrating PACE into Qwen2.5-VL-7B, the model retains 93.8% of its original performance while utilizing only 10% of the visual tokens, yielding a 3.1x speedup in time to first token (TTFT). Our code is available at this https URL.
[CV-17] Vision-centric generative AI models: A software-hardware perspective
链接: https://arxiv.org/abs/2608.27199
作者: Eleni Tselepi,Cristian Sestito,Shady Agwa,Themis Prodromakis
类目: Computer Vision and Pattern Recognition (cs.CV); Hardware Architecture (cs.AR)
备注:
Abstract:Vision generative artificial intelligence (AI) has emerged as one of the most rapidly advancing areas of deep learning. The explosion of multimodal models has made them widely associated with text-to-image applications running on large datacentres. However, vision generative models are equally needed in applications that operate under strict hardware constraints at the edge, including autonomous vehicles, agricultural sensors, and mobile devices. In this Perspective, we argue that progress in vision generative AI has been driven by output quality, with hardware evolving reactively to accommodate growing model demands. We quantify the parameter cost and energy efficiency of these models across a range of accelerator platforms, and map four generative model families against seven real-world application domains. Finally, we advocate a software-hardware co-design approach, where deployment constraints are considered from the start of the design process, ensuring that the “right model” runs on the “right hardware” to serve the “right application”, making generative AI deployment sustainable and accessible across a much broader range of platforms.
[CV-18] Knowledge Distillation Driven Semantic NOMA with GAN Refinement for 6G Robotic Vehicle Networks
链接: https://arxiv.org/abs/2608.27198
作者: Qifei Wang,Zhen Gao,Li Qiao,Ziwei Wan,De Mi,Dapeng Li,Ying Sun
类目: Information Theory (cs.IT); Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注: Presented at IEEE VTC-Spring 2026
Abstract:To achieve sustainable intelligent mobility, 6G-empowered robotic vehicles (RVs) require high-fidelity visual perception under stringent bandwidth and energy constraints. Semantic communication offers a spectral-efficient solution but suffers from severe interference in uplink non-orthogonal multiple access (NOMA) RV networks. To address this, we propose a knowledge distillation-driven and generative models-enhanced NOMA framework for robust and green RV communications, named KDG-SemNOMA. First, we develop a ConvNeXt-based deep joint source-channel coding (DeepJSCC) architecture with an enhanced attention feature (AF) module for dynamic channel adaptation. Second, to mitigate interference without inference overhead, an orthogonal transmission teacher model guides the NOMA student model via a two-stage knowledge distillation strategy. Finally, to address the over-smoothing artifacts of pixel-wise optimization, we introduce a channel-conditional GAN (cGAN). By explicitly taking the Stage-I initial reconstruction and channel states as conditional inputs, this module refines coarse outputs into high-fidelity images with realistic textures. Experiments on FFHQ-256 demonstrate that KDG-SemNOMA significantly outperforms state-of-the-art methods in both pixel-level accuracy and perceptual fidelity.
[CV-19] Unsupervised Adaptation of 3D CT Foundation Models for 3D CBCT Segmentation MICCAI MICCAI2026
链接: https://arxiv.org/abs/2608.27190
作者: Gauthier Miralles,Loic Le Folgoc,Vincent Jugnon,Pietro Gori
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the EMA4MICCAI workshop at MICCAI 2026. 11 pages, 4 figures, 1 table. Code, trained models, weights, and liver annotations are available at this https URL
Abstract:Accurate 3D segmentation of cone-beam CT (CBCT) is critical for interventional and radiation therapy applications, yet it remains limited by two compounding challenges: the scarcity of annotated CBCT data and the large domain shift from diagnostic CT. Interventional CBCT exhibits fundamental modality differences from conventional CT, driven by acquisition and physics effects as well as contrast-specific vascular content, thereby limiting effective cross-modality model transfer. We propose a novel unsupervised domain adaptation (UDA) framework based on redundancy-reducing feature alignment, enabling 3D CBCT segmentation with no target-domain annotations or inference-time adaptation. Our framework is architecture-agnostic, seamlessly adapting both CNN-based and ViT-based foundation models. We evaluate our method on two challenging CT-CBCT liver segmentation benchmarks: one for interventional vascular procedures and one for radiation therapy, demonstrating that even large-scale pretrained segmentation networks require explicit feature-space bridging to generalize across acquisition modalities, and that our approach consistently outperforms existing pretrained foundation model and UDA strategies. To support reproducibility and benchmarking, we release the liver segmentations for a public CBCT dataset, along with the code, trained models, and weights.
[CV-20] SSMB: Self-Supervised Local Feature Detection under Motion Blur
链接: https://arxiv.org/abs/2608.27181
作者: Zhenjun Zhao,Fabio Bellavia,Wenting Wang,Fan Zhu,Jiajun Wu,Suryansh Kumar,Mingqiang Wei,Haoang Li,Javier Civera
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 20 pages, 11 figures, 14 tables
Abstract:Keypoint detection under motion blur remains a significant challenge, as blur distorts local image structure and degrades the repeatability of feature localization. Existing approaches either rely on computationally expensive deblur-then-detect pipelines that may introduce restoration artifacts, or learn to regress the image positions of handcrafted keypoints extracted on sharp images, which reflects the assumptions of the handcrafted detector rather than what is truly repeatable under blur. We present SSMB, a deblur-free, self-supervised keypoint detector for motion-blurred images that requires neither handcrafted detectors nor external pseudo-labels. SSMB introduces the Local Discriminability Enhancement (LDE) module, which restores fine-grained local discriminability after global feature mixing. Training is performed in two stages. First, geometric pretraining on synthetic shapes bootstraps spatially discriminative keypoint detection without any external detector, just from the rendered geometry. Second, blur-aware training on real sharp-blur image pairs learns blur-invariant detection through a multi-component self-supervised objective that enforces cross-domain consistency, geometric alignment, and spatial coverage. Extensive evaluations on keypoint detection, image matching, relative pose estimation, and visual localization under motion blur demonstrate that SSMB establishes a new state-of-the-art among sparse keypoint detectors, consistently outperforming both supervised and self-supervised baselines across all tasks. Code, models, and datasets will be publicly available upon paper acceptance.
[CV-21] Parameter-Efficient pretrained-CT-to-MRI Transfer for Rectal Cancer Segmentation: Performance-Calibration Trade-offs MICCAI2026
链接: https://arxiv.org/abs/2608.27178
作者: Aneesh Rangnekar,Jorge Tapias Gomez,Joseph O Deasy,Harini Veeraraghavan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to the Medical Image AI in Radiation Therapy (MIART) Workshop at MICCAI 2026
Abstract:Accurate rectal cancer segmentation from magnetic resonance imaging (MRI) is essential for adaptive radiotherapy and tumor response assessment, but deployment also requires computational efficiency and informative, calibrated uncertainty estimates. We therefore introduce SWIFT, a SWin pretrained model wIth parameter-eFficient and Tumor-aware fine-tuning for rectal cancer segmentation. A Swin V2 encoder pretrained on 10,444 public 3D CT volumes using a DINOv2-style objective was adapted to T2-weighted MRI through four cumulative configurations: full fine-tuning (SWIFT), decoder compression (SWIFTe), low-rank adaptation (SWIFTe-LoRA), and a four-member LoRA-decoder ensemble (SWIFTe-LDE4). Geometric accuracy, tumor detection, radiomic agreement, and probability calibration were evaluated on a held-out 247-case test set from a single-institution cohort acquired using 1.5 or 3 Tesla GE scanners. Compared with SWIFT, SWIFTe reduced total parameters by 70.1% (from 72.8M to 21.8M) and increased tumor detection rate from 89.9% to 93.9%, while achieving a slightly lower median surface DSC (0.61 versus 0.62) and improved radiomic agreement. In a separate SWIFTe ablation, removing tumor-aware augmentation reduced detection from 93.9% to 89.9% but increased surface DSC from 0.61 to 0.64, demonstrating a detection-boundary-agreement trade-off. SWIFTe-LoRA used 14.6% of SWIFTe’s trainable parameters while retaining similar segmentation performance. SWIFTe-LDE4 achieved the lowest calibration errors among the four configurations after temperature scaling (expected calibration error, 0.217; Brier score, 0.222), although the absolute expected calibration error indicates residual miscalibration. Similar efficiency-calibration patterns were observed using the public VoCo checkpoint, supporting robustness across pretrained initializations rather than external clinical generalizability.
[CV-22] mporal Sensitivity Analysis of Tessera Embeddings
链接: https://arxiv.org/abs/2608.27175
作者: Julia Guerrero-Viu,Alex López-Cifuentes,Ignacio Pérez-Villar,Fabio Pacifici
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages
Abstract:Many Earth Observation applications need land-use/land-cover maps that are both precise and frequently updated, yet the strongest Earth Observation foundation models build their embeddings from a full year of observations. We present a controlled study of the temporal sensitivity of Tessera, one of these leading foundation models, for land-use/land-cover mapping. Keeping the encoder frozen, we recompute its embeddings over varying observation windows, from a full year down to a single day. We use them as inputs to a linear probe and a UNet segmentation head, benchmarking both of them against from-scratch networks on LUCAS, DynamicEarthNet, and PASTIS-R datasets. We show that the value of the embeddings is task-dependent. Where classes are separated by phenology, as for the crop types of PASTIS-R, they reach a mean Intersection-over-Union of 58.3 , about 46% above the best from-scratch model. Where classes are temporally stable (e.g., forests in DynamicEarthNet and LUCAS), embedding-based and from-scratch models match only under full supervision. On both datasets, Tessera embeddings remain markedly more label-efficient. Degradation under shorter temporal windows is gradual and class-dependent. Contracting the window from one year to one month costs 39% of the segmentation accuracy on PASTIS-R but only 5% on DynamicEarthNet. Single-day embeddings still classify land cover in LUCAS at 3.4 times the chance level. Our study shows that temporal coverage is therefore a tunable cost rather than a fixed prerequisite, opening regimes such as near-real-time mapping and faster land-use/land-cover refresh cycles.
[CV-23] Ancient-Bench: A Comprehensive Multi-millennial Multi-medium and Multi-script Benchmark for Ancient Chinese Artifact Text Recognition EMNLP2026
链接: https://arxiv.org/abs/2608.27169
作者: Hiuyi Cheng,Nuo Xu,Yuyi Zhang,Xuhan Zheng,Wei Pan,Jing Zhang,Dezhi Peng,Minghui Liao,Yihua Teng,Jihao Wu,Haoyu Ren,Lianwen Jin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by EMNLP 2026
Abstract:Ancient Chinese artifact text recognition is fundamental to heritage digitization, and benchmarks for ancient texts are essential for evaluating current model capabilities. However, existing benchmarks suffer from ‘‘fragmentation’’, manifested in limited temporal coverage, limited medium diversity, and incomplete script types. Therefore, we present Ancient-Bench, a comprehensive benchmark of 2,700 images for ancient Chinese artifact text recognition, featuring three dimensions: Multi-millennial (spanning 3,000 years of character evolution), Multi-medium (covering nine artifact categories), and Multi-script (encompassing seven historical script forms). To enable consistent and fair evaluation across heterogeneous media, we further define three annotation standards tailored to the medium-specific characteristics of ancient texts: symbol standardization, character standardization, and parsing standardization. Extensive experiments on Ancient-Bench covering general Vision-Language Models (VLMs) and OCR-specialist models reveal that ancient Chinese artifact text recognition remains fundamentally unsolved, with persistent challenges in variant characters, specialized symbols, and hallucination. The dataset is available at this https URL.
[CV-24] Magpie: Real-Time World Renderer for Interactive Games
链接: https://arxiv.org/abs/2608.27168
作者: Xiaoyu Zhan,Xinyu Wang,Xiaohong Zhang,Huanjie Zhu,Tengjiao Sun,Pengcheng Fang,Jiaxing Yu,Yanwen Guo,Dongjie Fu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Technical report. this https URL
Abstract:Modern game development relies heavily on conventional graphics pipelines. High-quality visual content requires modeling, material authoring, animation, lighting, effects, and runtime optimization, making asset production expensive and extending the development cycle of game prototypes. Recently, video foundation models are beginning to change film and video production, but games differ from linear media, they require not only continuous and realistic imagery, but also stable and reproducible gameplay rules, object states, and interaction outcomes. We present Magpie, a real-time generative world-rendering system for interactive games. Magpie separates gameplay execution from visual generation. Designers define scenes and rules in a game engine. At runtime, the Game Engine resolves player actions and maintains world state, while an independent Render Server generates visual output from white-box frames produced by the engine. Magpie provides a system-level implementation path for applying generative models to real-time game rendering. It preserves gameplay designability and reproducibility, and reduces the dependence of early game prototypes on complete visual assets.
[CV-25] ReViCo: Unveiling the Limitations of VLMs in Visual Text Understanding via Error Correction
链接: https://arxiv.org/abs/2608.27154
作者: Bojun Zhang,Junhong Liang,Feifei Zhai,Fengxian Ji,Yu Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision Language Models (VLMs) have shown great success in general visual tasks, yet they still struggle to deeply understand text within images. In this paper, we introduce ReViCo (Real Visual Correction), a benchmark designed to evaluate VLM text understanding through a novel task of visual text error correction. ReViCo challenges models to identify and fix text errors in real-world images, which requires a profound understanding of the interplay between visual text and its surrounding visual context. We benchmark various VLMs using two distinct paradigms: prompt-based strategy and targeted model training, both aimed at pushing the limits of current models. Our experiments reveal a striking performance gap between even the best VLMs and human, and further analysis also shows that most models struggle to accurately perceive the visual text, resulting in frequent correction errors. By highlighting these gaps, ReViCo provides a new benchmark foundation for developing more robust and text-aware VLMs.
[CV-26] ANTShapes Benchmarking Datasets for Event-Based Neuromorphic Object Classification
链接: https://arxiv.org/abs/2608.27150
作者: M. Middleton,H. Kayan,B. Sen Bhattacharya,T. Ali,E. Baikas,M. Vousden,C. Perera,O. Rhodes,E. Gheorghiu,M. A. Trefzer
类目: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: 23 pages, 11 figures
Abstract:Object classification in event-based computer vision is a task that is attracting considerable research attention. Event-based object classification is a fundamental task in the fields of security and applied computer vision, which typically use synchronous frame-based cameras and computing pipelines for operation. This approach has several practical flaws. The size, weight and power consumption of the device could prohibit deployment at the extreme edge or in covert sensing environments. Besides this, there are security concerns inherent in cloud-based or other off-device computation approaches due to the requirement of sending and receiving potentially sensitive data. Furthermore, this transmission of data introduces latency and requires consistent connectivity to the cloud infrastructure to function. The use of Spiking Neural Networks (SNNs) hosted on neuromorphic devices attempts to solve several issues present in this conventional approach. Research into event-based object classification methods are hindered by the lack of high-quality vision datasets to use. To this end, the ANTShapes simulation tool has been previously proposed to create and label event-based vision datasets. In this paper, four novel datasets of varying difficulties are created using the tool and are benchmarked against existing spiking datasets commonly used for event-based vision research (N-MNIST, CIFAR10-DVS, DVSGesture and POKER-DVS). Classification is performed using a convolutional SNN. This work simultaneously provides four datasets with rich details for future experiments to use and validates the output of the ANTShapes dataset simulation tool as being suitable for its purpose. Comments: 23 pages, 11 figures Subjects: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2608.27150 [cs.NE] (or arXiv:2608.27150v1 [cs.NE] for this version) https://doi.org/10.48550/arXiv.2608.27150 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-27] EditaLive! Unified Character Video Editing for Live Streaming
链接: https://arxiv.org/abs/2608.27123
作者: Zhiyuan Li,Chi-Man Pun,Peng-Tao Jiang,Bo Li,Xiaodong Cun
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Conventional video editing primarily focuses on scene-level content, whereas live streaming places greater emphasis on the human subject. However, directly applying existing video-editing methods to human-centric live streaming remains challenging, as they may introduce facial-expression inconsistencies and typically depend on multiple offline inference steps, making them unsuitable for real-time interaction. We propose EditaLive, a novel framework for real-time streaming character video editing. In detail, we start from a pretrained image animation model (Wan-Animate), which naturally decouples appearance from motion, and repurpose it as the base model for instruction-based human-centric video editing by reference frame editing and video reconstruction via the collected CharEdit-50K dataset. Besides, we adapt the model from offline bidirectional to causal streaming generation, and design an aligned self-rollout distillation strategy that compresses the model into a two-step sampler, where fixed RoPE and align forcing reduce training–inference discrepancies, and first-frame preserved sparse attention filters redundant historical information to mitigate appearance drift. Extensive experiments demonstrate that EditaLive delivers state-of-the-art editing performance with faithful preservation of facial expressions and low-latency real-time streaming inference.
[CV-28] How AI Experiences Art: Emergent Aesthetic Structure in a Self-Supervised Multimodal Embedding Space
链接: https://arxiv.org/abs/2608.27121
作者: Corey D.C. Heath
类目: Multimedia (cs.MM); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted at ICMI Companion '26 (Companion Publication of the 28th ACM International Conference on Multimodal Interaction), October 5-9, 2026, Napoli, Italy. 4 figures, 2 tables
Abstract:Aesthetics are an important part of the symbolism of artistic works. Although subjective, humans categorize art based on the emotion evoked regardless of modality. What remains under-explored is how AI models form their own aesthetic categorization of human-produced media without explicit labels or cross-modal supervision. We present a self-supervised framework that projects four modalities (text, audio, image and video) into a shared 256-dimensional embedding space and applies iterative clustering to discover aesthetic structure. We discuss the divergence between AI-generated cluster assignments and human affective register labels on a weakly supervised multimodal dataset. This work has applications in understanding how AI structures cross-modal similarity, organizing heterogeneous media collections for Retrieval-Augmented Generation (RAG), and automated data labeling.
[CV-29] Automated 2D and 3D Segmentation of AMD and DME Lesions in OCT
链接: https://arxiv.org/abs/2608.27095
作者: Lucia Sundberg,Zhihao Zhao,M. Ali Nasseri
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Age-related macular degeneration (AMD) and diabetic macular edema (DME) are leading causes of vision loss, and optical coherence tomography (OCT) is the standard modality for detecting and monitoring the subtle lesions that drive treatment decisions. Most deep-learning segmentation work for OCT is validated only in-domain, leaving generalization to clinical data collected under different acquisition protocols largely untested. This work develops and systematically ablates four lesion-segmentation pipelines – 2D and 3D variants for AMD and DME – reaching Dice scores of 0.76 to 0.82 with strong volumetric and surface calibration (r vol, r surf greater than or equal to 0.97 across all four pipelines) on an in-domain validation set. The ablation process establishes a full-volume, calibration-aware adoption standard that catches mechanisms an ordinary slice-level evaluation would keep, and identifies ensemble composition as the most consistent driver of improvement. To test generalization, the models are evaluated on OLIVES, an external clinical cohort with no lesion-level ground truth, using a proxy-metric framework built around biomarker AUROC, central subfield thickness (CST) correlation, and longitudinal concordance. Predictions track clinical biomarkers outside the training distribution, though less strongly than in-domain – evidence for, not validation of, automated lesion-burden tracking as a clinical tool.
[CV-30] SpatialCrafter: Single Image World Modeling with Generative 3D Proxies
链接: https://arxiv.org/abs/2608.27073
作者: Chuan Fang,Lingteng Qiu,Yixun Liang,Rui Chen,Kunming Luo,Zhaohua Zheng,Tongyuan Bai,Feipeng Tian,Zilong Dong,Zihan Zhou,Ping Tan
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 12 pages
Abstract:Explorable image-to-scene generation is essential for applications in gaming, robotics, and virtual reality. Existing methods based on video diffusion model (VDM) commonly rely on incomplete conditioning signals such as sparse point clouds or 2D panoramas, leading to stochastic hallucinations, long-term drifts and suboptimal 3D consistency. We present SpatialCrafter, a novel two-stage framework that addresses these issues by introducing a global 3D proxy for high-fidelity image-to-scene generation. Specifically, we decompose the generation process into global proxy generation and appearance refinement. For proxy generation, we propose a Point-anchored Sparse Structure~(PaSS) Flow module that predicts a spatially aligned and geometrically consistent 3D proxy. For appearance refinement, we re-frame the VDM as a Generative Deferred Refiner which synthesizes high-frequency photorealistic details upon proxy-defined scene geometry. To better integrate the proxy with the pre-trained VDM, we introduce Parallel Geometry Injection and Proxy-Aware Corruption training strategies, which improve robustness to proxy artifacts without disrupting the pretrained generative manifold. Furthermore, as no suitable dataset exists for this explorable scene generation task, we construct a new large-scale dataset of 115K scenes. To the best of our knowledge, it is the first hybrid dataset for image-to-scene generation. Extensive experiments on both synthetic and real-world datasets show that SpatialCrafter outperforms state-of-the-art methods, mitigates long-term drift, and remains robust and consistent under rapid camera motion and extreme viewpoint changes. Code, models, and the newly constructed dataset will be publicly released. See more at this https URL.
[CV-31] Beyond Classification: Task-Dependent Learnability under Privacy-Motivated Image Transformations
链接: https://arxiv.org/abs/2608.27066
作者: Leon Ranke,Wolfgang Hübner,Ronny Hug,Michael Arens,Jürgen Beyerer
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Privacy-Enhancing Technologies (PETs) in computer vision often rely on noise or image perturbations to protect visual data while securely processing it, creating a trade-off between task performance and protection. This trade-off is commonly evaluated using image classification, which primarily captures semantic separability and remains robust despite significant geometric, spatial layout or local boundary alterations. As a result, it is too simplistic as a proxy for generic vision tasks. Exhaustive downstream-task evaluation, however, is computationally expensive because models must often be trained for each PET transformation and parameter setting. We therefore propose a compute-aware multi-task protocol for evaluating PETs in model training. It combines lightweight proxy tasks that target complementary aspects of visual structure while remaining simple and fast to compute. Across irreversible privacy transformations, key-based block primitives, and learnable image encryption schemes, we demonstrate that PETs with similar classification accuracy can differ substantially on other tasks. The outcomes highlight the need for PET evaluation protocols that move beyond classification-only reporting.
[CV-32] Video-OPSD: Exploiting Privileged Visual Evidence for On-Policy Self-Distillation in Video Large Language Models
链接: https://arxiv.org/abs/2608.27065
作者: Ziyue Wang,Shiqi Huang,Weiwen Xu,Bihan Wen,Xudong Jiang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:On-policy self-distillation (OPSD) has recently emerged as an effective post-training paradigm that improves policy optimization through dense token-level supervision from a privileged self-teacher. Despite its promise, OPSD remains largely underexplored for Video Large Language Models (Video-LLMs). Existing methods typically construct privileged teachers by augmenting their context with additional information while keeping the primary input unchanged for both teacher and student. Video reasoning, however, offers a distinct source of privileged supervision within the primary input itself: long videos contain substantial temporal redundancy, and only a small subset of frames provides the evidence necessary to answer a question. Building on this observation, we present \textbfVideo-OPSD , an OPSD framework that exploits privileged visual evidence for both self-teacher construction and knowledge transfer. First, our Evidence-Grounded Self-Teacher conditions the teacher exclusively on annotated evidence frames while the student continues to reason over the complete video. This focused visual input enables the teacher to provide more informative supervision. Second, our Evidence-Guided Token Optimization adaptively weights token-level distillation according to each reasoning token’s reliance on privileged visual evidence, thereby emphasizing perceptually grounded reasoning. Experiments across video understanding and reasoning benchmarks show that \textbfVideo-OPSD consistently improves upon Standard OPSD across multiple backbones and achieves performance comparable to GRPO while requiring substantially less training time, establishing an effective and efficient post-training approach for Video-LLMs.
[CV-33] Anatomy-Guided Foundation Model Adaptation with Within-Case Prototype Supervision for Standard Plane Detection in Fetal Ultrasound Blind Sweeps
链接: https://arxiv.org/abs/2608.27051
作者: Yuzhe Zhao
类目: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注:
Abstract:Detecting the fetal abdominal circumference standard plane in low-cost obstetric blind sweeps is a highly imbalanced frame-classification problem: positive frames account for under 3% of a sequence, form short contiguous segments, and are poorly handled by off-the-shelf ultrasound and vision foundation models. We propose AnatoProto, a lightweight sequence-level framework that adapts a frozen BiomedCLIP encoder to fetal blind sweeps through four components: (i) anatomy-weighted spatial pooling that uses nnU-Net abdominal-region probabilities as a spatial prior to reweight BiomedCLIP patch tokens, so frozen semantic features are aggregated onto anatomically meaningful regions; (ii) a within-case prototype loss that pulls each frame embedding toward the mean of positive frames of the same sweep, exploiting case-level structure unavailable at the frame level; (iii) a three-stage cascade refinement (frame-segment-case-level rejecter) that lifts the prediction unit from noisy frames to structurally-constrained segments; and (iv) a hybrid prediction head that jointly models per-frame stability and inter-frame boundary transitions to suppress boundary false positives. On the ACOUSLIC-AI benchmark, AnatoProto reaches a test F1 of 67.72, outperforming the strongest foundation-model baseline (FetalCLIP + PRS, F1 = 54.52) by +13.20 F1 and the strongest video temporal-action-detection baseline (TriDet + PRS) by +15.76 F1. A synergy study, backed by embedding geometry and paired-bootstrap confidence intervals, shows that the prototype loss and anatomy-weighted pooling are not additive: applied alone the prototype loss reduces recall by 12 points, but combined with anatomy-weighted pooling it increases recall by 6.5 points – a sign-flip we trace to the accuracy of the within-case prototype.
[CV-34] Omni-Interactive Universal Embedder
链接: https://arxiv.org/abs/2608.27044
作者: Wei-Yao Wang,Kazuya Tateishi,Shuyang Cui,Christian Simon,Takashi Shibuya,Shusuke Takahashi,Yuki Mitsufuji
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Preprint
Abstract:Multimodal representation learning has been shifting from traditional two-tower architectures to large language model (LLM)-based embedders due to their strong instruction-following capabilities. Despite this progress, existing approaches primarily focus on language and image modalities, which also remain the dominant modalities for user-conditioned interactions in current embedders. In this paper, we propose the first Omni-Interactive Universal Embedder (OmniUE), which not only learns a unified embedding space across text, video, and audio by leveraging intermediate-layer representations from dedicated learnable tokens, but also supports omni-interactive querying, enabling users to provide inputs in the form of text, visual regions of interest, and audio spans. Within OmniUE, visual and audio segmenters process diverse user interactions and integrate them with an omni-LLM to produce user-conditioned any-to-any embeddings via context aggregation. To evaluate OmniUE’s omni-interactive capabilities, we introduce OmniCHOIR, benchmarking models for omni-interactive compositional audio retrieval based on the given text, video, and audio as well as unimodal or multimodal interaction prompts. OmniUE consistently surpasses state-of-the-art baselines across diverse modalities, with average improvements of 10.5% on textual-interactive video benchmarks (MMEB-v2-video), 1.1% on audio tasks (MAEB), 83.7% on visual-interactive benchmarks (SCaR), and 24.1% on our omni-interactive OmniCHOIR benchmark. We believe that jointly advancing omni-modal representation learning and omni-interactive querying paves the way toward universal embedders.
[CV-35] Multi-Person Human Motion Forecasting in Complex Scenes
链接: https://arxiv.org/abs/2608.27039
作者: Serdar Ozsoy,Lars Doorenbos,Juergen Gall
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to GCPR 2026
Abstract:Accurately forecasting the movement of people in complex scenes requires reasoning over the past and present state of the entire environment. In this context, effectively incorporating object information and social interactions into a unified framework remains particularly challenging. To address this, we propose Object-Conditioned Social Diffusion (OCSD), a conditional diffusion model that integrates motion history, multi-person interactions, and object cues into a single framework. OCSD uses an object-conditioning mechanism that modulates denoising at every timestep, enabling fine-grained human-object reasoning, and a social encoder that models the interactions between all humans in the scene. As a result, our model naturally handles varying group sizes, complex social interactions, and supports sampling multiple plausible futures. Extensive experiments show that OCSD achieves state-of-the-art results on the Humans in Kitchens (HiK) and HOI-M3 benchmarks. It reduces the two-second path error by 121.5 mm (31.3%) on HiK and 130.5 mm (33.2%) on HOI-M3 compared to prior work, and produces more realistic long-term forecasts.
[CV-36] Differentiable Jitter Correction using Deep Learning-based Image Quality Metric for Phase-Contrast Micro-CT
链接: https://arxiv.org/abs/2608.27034
作者: Junan Chen,Yiting Jia,Joscha Maier,Dominik John,Sami Wirtensohn,Imke Greving,Silja Flenner,Matthias Wieczorek,Julia Herzen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:This paper proposes a fully differentiable jitter correction method for X-ray phase-contrast micro computed tomography using a deep learning-based image quality metric that estimates and compensates per-projection rigid jitter directly from the acquired projection data, without a pre-scan motion-free reference. The approach builds on a gradient-based auto-focus strategy adapted to parallel-beam geometry. A set of candidate objective functions is benchmarked in a controlled study, and the sensitivity of the visual information fidelity (VIF) metric to the jitter artifact is verified with the target phase-contrast data. To operate without a clean reference, a compact 3D convolutional neural network is trained to predict the VIF score from a single corrupted volume. A spatially selective total variation penalty applied exclusively to the image background is introduced to penalize spurious high-frequency structures that otherwise emerge during optimization. Experiments on biological specimens acquired at different synchrotron beamlines are conducted. Evaluation uses jitter motion applied to simulated and experimentally acquired projection data. The result confirms that the integrated pipeline reliably recovers fine structural detail lost due to jitter, with generalization demonstrated across morphologically distinct samples.
[CV-37] MVC-Bench: Benchmarking Calibration of Medical Vision-Language Models EMNLP2026
链接: https://arxiv.org/abs/2608.27004
作者: Ashshak Sharifdeen,Shihab Aaqil Ahamed,Ufaq Khan,Muhammad Akhtar Munir Sujair Ibrahim,Mohamed Rafeek Mareer Ahamed,Yutong Xie,Imran Razzak,Muhammad Haris Khan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted in EMNLP 2026 Findings
Abstract:Reliable evaluation of vision-language models (VLMs) and medical vision-language models (Medical-VLMs) requires calibrated confidence, particularly under realistic clinical conditions. However, existing efforts mainly focused on improving accuracy, leaving calibration in the medical domain underexplored. To this end, we propose MVC-Bench, a calibration-centric benchmark for medical image classification with VLMs and Medical-VLMs. MVC-Bench assesses the calibration across three axes: (i) robustness to modality, backbone, and domain shift (ii) effectiveness of calibration strategies and prompt-tuning methods (iii) stability under prompt-template and random-seed variations. The benchmark covers eight different backbones, three medical modalities, including fundus imaging, histopathology, and chest X-ray under in-domain and domain shift settings. It compares post-hoc calibration, train-time calibration, and zero-shot inference methods, together with six prompt-tuning methods. Across more than 1638 controlled experiments, we report accuracy and Expected Calibration Error (ECE) as primary metrics, and further report results with complementary calibration measures, including Maximum Calibration Error (MCE) and Adaptive Calibration Error (ACE). We further investigate the underlying causes of miscalibration in VLMs and Medical-VLMs and propose a simple train-time calibration method, Multi-Class Margin (MCM) regularization, which achieves lowest ECE on 10 out of 12 settings in in-domain and remains competitive under domain shifts. Collectively, MVC-Bench provides a structured evaluation framework and actionable guidance for improving calibration in safety-critical medical workflows.
[CV-38] Virtual iEEG from Scalp EEG: Charting the Landscape of Source Imaging Intracranial Inference and Reconstruction
链接: https://arxiv.org/abs/2608.26998
作者: Dongyi He,Xiangkai Wang,Hongjie Yan,Luping Song,Wai Ting Siok,Nizhuan Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Intracranial electroencephalography (iEEG) provides temporally precise and spatially specific access to neural activity from focal and deep brain regions, but its invasiveness and restricted anatomical coverage limit routine use. These constraints have motivated scalp-to-intracranial inference, termed virtual iEEG when model outputs carry iEEG-defined event, feature, representation, or contact-level waveform semantics. This review presents a target-centred framework distinguishing event inference, feature translation, and waveform reconstruction, while separating predictability from observability, identifiability, fidelity, and utility. Evidence is evaluated according to cohort independence, anatomical and spectral coverage, train–test separation, and target-patient adaptation. Current studies support inference of selected intracranial events, low-frequency components, and task-related representations, but not unique recovery of arbitrary contact-level activity. Stronger validation requires appropriate controls, source-imaging baselines, uncertainty assessment, and incremental-utility testing. Future progress depends on independent paired datasets and prospective evidence that virtual iEEG adds value beyond scalp EEG and EEG source imaging.
[CV-39] Aphanta: Diagnosing Task-Aligned Image-Edited Intermediates for Multimodal Reasoning
链接: https://arxiv.org/abs/2608.26993
作者: Hengyuan Xu,Wei Cheng,Yumeng Ji,Xuanyang Zhang,Xianfang Zeng,Gang Yu,Xingjun Ma
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Explicit visual intermediates can help multimodal large language models (MLLMs) externalize spatial evidence and updated visual states, but their utility depends on whether an image editor can faithfully realize the required transformation. We introduce \textbfAphanta, an automated task-discovery and closed-loop diagnostic framework for the MLLM - image editor - MLLM pipeline. Aphanta evaluates three conditions—direct reasoning, reasoning with an editor-generated intermediate, and reasoning with an idealized reference intermediate—to separate potential visual headroom from the practical utility of current editors. Across 20 candidate tasks and multiple editor–MLLM combinations, we find that utility is strongly task-conditioned. Gains concentrate in visual cue injection, grounding, and counterfactual state realization, whereas intermediates requiring symbol-sensitive construction or structural extrapolation are substantially less reliable. On the selected positive-task subset, our consolidated Qwen pipeline improves the mean task score from 0.343 to 0.445 ( +10.2 points; +29.7% relative), while the full study also retains filtered and unsuccessful tasks to expose the boundary. These results position image editing as a specialized visual workspace rather than a universal reasoning mechanism, and establish Aphanta as a reusable protocol for measuring task–representation alignment, editor realization, and downstream pipeline utility.
[CV-40] mpJail: Temporal Jailbreak Attacks against Image-to-Video Generation Models ACM-MM’26
链接: https://arxiv.org/abs/2608.26971
作者: Qi Lu,Zehui Guo,David Yuanda Gan,Zijing Li,Hengda Zhang,Weijun Xu,Qiankun Zhang
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Accepted by ACM Multimedia 2026 (ACM MM '26)
Abstract:In recent years, image-to-video (I2V) generation models have made remarkable progress in subject consistency and temporal coherence, enabling high quality video synthesis. However, these advances also introduce new safety risks. Existing studies mainly focus on jailbreak attacks involving single frame violations, while largely overlooking the temporal dimension unique to video generation models. In this paper, we investigate three attack scenarios and uncover a temporal vulnerability in I2V systems: unsafe semantics may emerge not from a single frame, but from semantic composition over time. We further identify two key challenges in such attacks: temporal abstraction and semantic camouflage. To address these issues, we propose TempJail, a novel temporal jailbreak framework for I2V systems. For temporal abstraction, we decompose a target malicious caption into an initial frame visual condition and a temporal text instruction. For semantic camouflage, on the image side we model semantic injection as controlled latent perturbation in diffusion sampling and introduce gradient guidance from pretrained encoders. On the text side, we rewrite the caption into an innocuous ``subject-action-scene’’ template that bypasses safety filters while preserving temporal guidance. In the black-box inference phase, these two modalities jointly enable malicious semantics to be gradually triggered over time. Experiments on closed-source commercial models, including Kling, Seedance, Veo and PixVerse, show that TempJail improves attack success rate over prior state-of-the-art methods by 23.3% under GPT-5.2 evaluation and 22.0% under human evaluation. Our codes are available at \hrefthis https URLGitHub.
[CV-41] ClusterAttention: A training-free speedup of bidirectional attention
链接: https://arxiv.org/abs/2608.26965
作者: Kasper Nordenram,Amelie Dittmann
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 3 figures
Abstract:This paper introduces ClusterAttention, a general training-free speedup of bidirectional attention layers. Existing sparse attention methods either rely on structure in the input, such as order in language or spatial proximity in images, or use slow clustering processes amortized over several forward passes. ClusterAttention instead uses a fast recursive clustering method that adapts to the geometry of the keys and queries in each attention head to produce useful clusters. This method allows setting the size of the clusters arbitrarily. We utilize this by setting all clusters to be a fixed size that is a power of two, allowing the block-sparse attention to run at the same latency per query-key interaction as dense attention on GPUs. We also derive an expression for the output error in sparse attention, that explains the counterintuitive experimental finding that tight clusters can lead to larger errors than random clusters. We then derive the error when excluded clusters are compensated through their centroids, and show that this error shrinks with tighter clusters. We integrate this compensation into the method. On large-scale tabular data ClusterAttention speeds up TabPFN-3 arXiv:2605.13986 by two to six times, while retaining at least 99% of the dense accuracy. To our knowledge, it is the first training-free method that can be successfully applied in the setting of unstructured input and a single forward pass. For video generation with Wan 2.1-14B T2V arXiv:2503.20314 , ClusterAttention achieves output closer to dense attention and a larger speedup (1.8x versus 1.4x) compared to SVOO arXiv:2603.18636 , a leading method developed specifically for this domain, both run without offline calibration. Comments: 18 pages, 3 figures Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2608.26965 [cs.LG] (or arXiv:2608.26965v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.26965 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.5281/zenodo.22118033 Focus to learn more DOI(s) linking to related resources
[CV-42] Geo-LoRA: Geometry-Aware Subspace Evolution for Low-Rank Adaptation in Continual Learning
链接: https://arxiv.org/abs/2608.26960
作者: Yibo Feng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Rehearsal-free class-incremental learning (CIL) with LoRA adapters remains challenging because the low-rank subspaces updated across tasks evolve without geometric control, causing unstable shared representations and repetitive collapse of task-specific updates into previously occupied directions. We introduce Geo-LoRA, a geometry-aware framework that explicitly regulates how low-rank subspaces, both shared and task-specific, evolve during continual learning. For the shared branch, Subspace Projection Preservation (SPP) constrains consecutive updates to follow smooth trajectories on the Grassmann manifold, and Adaptive Core-Slack Alignment (ACSA) decomposes transitions into principal and residual components, aligning the former while modulating the latter to balance stability and plasticity. For the task-specific branch, Median-Calibrated Block Overlap (MCBO) imposes a statistical constraint via normalized projection overlap, penalizing excessive reuse to mitigate subspace crowding. These constraints jointly regulate the evolution of all LoRA subspaces across layers and tasks without introducing additional adapter types beyond standard LoRA. Geo-LoRA provides a principled geometric formulation for continual low-rank adaptation and consistently achieves state-of-the-art performance across multiple benchmark datasets and different task lengths.
[CV-43] RubricRM: Generative Reward Modeling via Dynamic Rubrics for Image Generation and Editing EMNLP2026
链接: https://arxiv.org/abs/2608.26956
作者: Zijian Kan,Wei Wang,Long Luo,Bing Zhao,Xuan Ren,Weixu Qiao,Wenbo Li,Hu Wei,Lin Qu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to EMNLP 2026 Main Conference
Abstract:Reward models play an essential role in aligning visual generative models, yet most existing visual reward models use a single scalar score or rely on fixed criteria that cannot adapt to different instructions. This limits both interpretability and task sensitivity, especially for text-to-image generation and instruction-based image editing, where different inputs require different evaluation dimensions. We propose RubricRM, a pairwise generative reward modeling framework that first produces an input-specific rubric with evaluation dimensions, weights, and scoring criteria, and then applies the rubric to score candidate images. We train dedicated RubricRM models for text-to-image generation and image editing using a two-stage training pipeline: supervised fine-tuning teaches the model the rubric-based scoring paradigm, while GRPO further improves scoring through fine-grained dimension-level rewards. Experiments on multiple generation and editing benchmarks show that RubricRM outperforms existing specialized reward models and remains competitive with strong proprietary MLLM judges despite using smaller backbones. Our models, data, and code are available at this https URL.
[CV-44] Per-View Gaussian Predictions Enable Training-Free Distractor Filtering in Feed-Forward 3DGS
链接: https://arxiv.org/abs/2608.26951
作者: Kangmin Seo,Jae-Pil Heo
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Preprint
Abstract:Feed-forward 3D Gaussian Splatting reconstructs an explicit Gaussian representation from multiple input images in one network execution, making 3D reconstruction increasingly accessible for casual captures. However, such captures frequently contain transient objects that appear in only a subset of the views. Such content can be encoded into the per-view Gaussians associated with the inputs that observe it and remain in the combined representation despite being observed by no other input. As a result, it may produce blurred, duplicated, or floating artifacts in novel views. We introduce a training-free filtering procedure that exploits this per-view prediction structure. For each input, we exclude its associated Gaussians and render the same camera using the remaining representation, revealing content that is inconsistent with the other inputs. Feature similarity forms candidate regions, and rendering-based verification retains only candidates whose removal reduces reconstruction error in the other input views. The procedure operates on a single frozen prediction without retraining or scene-specific optimization. Across three reconstruction models and two distractor benchmarks, it consistently improves novel-view quality with varying numbers of input views. On clean scenes, evaluations across four models show that the original reconstructions are largely preserved.
[CV-45] KISS-GS: 3D Gaussian Splatting Compression Kept Simple
链接: https://arxiv.org/abs/2608.26948
作者: Wieland Morgenstern,Friedrich Elias Branschke,Florian Fleischmann,Adrian Szatmari,Paul Schlack,Florian Barthel,Peter Eisert,Anna Hilsmann
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Scene reconstruction with 3D Gaussian Splatting (3DGS) has become common, however deployment remains painful as the uncompressed file sizes can be massive. Current 3DGS compression systems combine multiple strategies for file size reduction, which can obscure where gains come from and limit component reuse across training pipelines. To make the gains more transparent, we propose KISS-GS, a modular compression pipeline named after the principle of keeping things simple, designed to decouple compression entirely from training. Given a 3DGS scene reconstructed with vanilla 3DGS, we are able to reduce it through compaction by 15.7x using a combination of state-of-the-art pruning schemes. Then we encode it into an image-based format designed for simple, ubiquitous decoding. With the SOG-XT format, we propose a novel extension to Self-Organizing Gaussians with two main contributions: (i) Self-organizing 2D Codebooks and (ii) Parallel Representative Assignment Smoothing (PRAS), which leverages the symmetry of quaternion and scale parameterizations to produce 2D attribute grids more amenable to encoding. This encoding reduces scene size by 6.6x. We show that optional encoding-aware fine-tuning yields a further 2.2x. Across standard 3DGS benchmarks, our simple and modular approach thus achieves a total of 85x to 319x reductions in the size of the scene over uncompressed vanilla 3DGS, setting new benchmarks for real-world scenes and surpassing tightly integrated methods in rate-distortion. Decoding relies solely on web-native image formats, and the modular design makes each stage easy to combine with future advances in reconstruction and compaction. Code and project page: this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2608.26948 [cs.CV] (or arXiv:2608.26948v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2608.26948 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Wieland Morgenstern [view email] [v1] Thu, 27 Aug 2026 10:49:42 UTC (6,141 KB)
[CV-46] 4DSynth: Controllable Procedural World Synthesis for Dynamic Embodied Simulation
链接: https://arxiv.org/abs/2608.26947
作者: Zehao Qi,Haochen Luo,Jia-Wang Bian,Zeyu Ma,Shuyang Sun
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Embodied agents need environments that are visually diverse, physically interactive, and changing over time. Procedural simulators can generate large interactive scene collections, and recent 4D generators produce compelling visual dynamics. Combining these properties in one environment, however, still demands extensive manual effort, and the result is rarely editable or controllable enough to reuse at scale. We present 4DSynth, a controllable procedural system that turns a natural-language description, a blueprint mask, or a single photograph into an editable 4D environment with explicit geometry, animated actors, collision-free trajectories, and physics-ready simulation state. Multiple scene routes share one geometry-grounded representation, so the same pipeline handles animation, camera planning, rendering, and task generation. To validate the full pipeline, we construct 4DSynth-Nav, an interactive navigation benchmark generated entirely from 4DSynth’s procedural scenes. Two vision-language models evaluated across three difficulty tiers both fail the majority of tasks and stall after early subtasks. The same procedural controllability that produces these environments also makes each failure reproducible and each difficulty axis independently tunable. This paper presents both a controllable generation pipeline and the scalable benchmark it enables, offering a practical foundation for developing and evaluating embodied agents. Subjects: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2608.26947 [cs.RO] (or arXiv:2608.26947v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2608.26947 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-47] her the Subject Release the Scene: Query-Aware Memory Routing for Long-Horizon Autoregressive Video Generation
链接: https://arxiv.org/abs/2608.26902
作者: Chen Li,Peng Zhang,Hanyu Zhou,Jialong Zuo,Fei Wang,Daiguo Zhou,Nong Sang,Changxin Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Streaming autoregressive video models generate long videos chunk by chunk, using historical memory to maintain consistency. Existing methods typically expose subject and scene queries to history through similar policies. This stabilizes the subject, but can also lock backgrounds, viewpoints, and scene structure to previously generated states even when local motion continues. We call this failure memory-anchored scene under-progression; consistency and motion metrics alone can miss it. We introduce TetherMem, a training-free, query-aware spatiotemporal memory router for frozen video generators. TetherMem separates subject and scene queries and modulates historical access with region- and age-conditioned priors: subject queries retain identity-bearing history, while scene queries reduce reliance on subject history and stale backgrounds. Across 2,400 blinded pairwise judgments from 10 annotators, TetherMem achieves the highest estimated expected preference among eight streaming long-video baselines for overall quality (0.780) and scene progression (0.769). On complete 30-second videos, it sustains changes in background, viewpoint, and scene state while preserving subject recognizability and temporal continuity.
[CV-48] Self-OPD: On-Policy Distillation for Flow Matching Models without Teacher
链接: https://arxiv.org/abs/2608.26872
作者: Shiyi Zhang,Mushui Liu,Yunze Tong,Wanggui He,Siyu Zou,Jinlong Liu,Yunlong Yu,Jian Song,Hao Jiang,Pipei Huang,Bo Zheng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 19 pages, 10 figures
Abstract:On-policy distillation (OPD), which leverages a pre-trained, specialized teacher model to provide dense supervisory signals, has achieved significant success in Large Language Models (LLMs) and has recently been adapted to flow matching models. However, this paradigm suffers from two major issues: First, training a separate, task-specific teacher for every new objective incurs high computational costs. Second, the discrepancy between teacher and student distributions often leads to compounding errors along the generation trajectory. In this paper, we introduce \textbfSelf-OPD, a teacher-free OPD framework for flow matching models that turns the student’s own self-exploration into step-wise supervision. At each timestep, Self-OPD branches the deterministic next-state prediction into K stochastic SDE candidates, rolls them out with the ODE sampler, and compares their rewards against a deterministic self-reference baseline to obtain normalized advantages. The velocity field is optimized with an all-branch pull-push objective, where high-advantage branches attract the student and low-advantage branches repel it under direction-aware attenuation and SDE-variance normalization. For multi-objective alignment, Self-OPD fuses normalized scores at the reward level, avoiding direct gradient conflict. Experiments on single and mixed reward benchmarks show that Self-OPD outperforms prior RL and OPD methods without task-specific teachers.
[CV-49] CGS-SLAM: Collaborative Gaussian Splatting based SLAM for Multi-Agent Reconstruction
链接: https://arxiv.org/abs/2608.26868
作者: Jean-Daniel de Ambrogi,Aladine Chetouani,Vincent Nguyen,Aurélien Chateigner
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Recent advances in SLAM have leveraged 3DGS for photorealistic reconstruction and novel view synthesis. However, most methods rely on RGB-D input, which is unavailable on consumer-grade smartphones, and few integrate 3DGS within a collaborative framework. Therefore, we present CGS-SLAM, a hybrid decentralized/centralized system enabling multi-agent 3DGS SLAM using only RGB and inertial data. Each agent performs local tracking with inertial data as a motion prior and reconstructs a scaled map using a metric monocular depth estimator (Depth Pro). Keyframe encodings are shared among agents, enabling dynamic keyframing in regions of spatial overlaps with other agents, enhancing submap alignment. Afterwards, a central server aligns submaps using VGGT as a view alignment model. This bidirectional communication keeps communication cost low during mapping and global reconstruction in difficult GNSS-denied environments. Experiments on multiple datasets demonstrate competitive tracking performance, improved rendering quality over state-of-the-art methods, and accurate submap alignment.
[CV-50] Order Matters: A Chinese Multi-Panel Meme Benchmark for Vision-Language Reasoning
链接: https://arxiv.org/abs/2608.26866
作者: Haihan Li,Haihao Li,Zhenfei Xu,Jize Qian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Many multimodal tasks depend on how visual elements are ordered and composed, not only on recognizing them in isolation. Internet memes are a compact case of this problem: their punchline often depends on a constrained reading order and cross-panel visual–textual cues. While large vision-language models (LVLMs) show strong performance on single-image understanding, it remains unclear whether they can perform sequence-aware reasoning over structured meme layouts, especially in Chinese social media. We introduce CMPM, a Chinese Multi-Panel Meme benchmark with 1,214 annotated samples covering five structural types, ordering dependency, panel-order constraints, and optional comment context. We formulate a two-layer evaluation: Task1 probes structure typing and order-sensitive panel sequencing (with a context ablation setting), and Task2 evaluates Chinese meme explanation generation with human ratings on five 1-3 Likert dimensions (visual, panel, humor, context, and faithfulness). We benchmark five representative LVLMs under a unified protocol. Results indicate that canonical-display accuracy is not by itself evidence of order understanding: the primary shuffled condition produces a sharp accuracy drop, revealing a persistent gap in order-sensitive multimodal reasoning. Task2 preferences place Gemini 3.1 Pro and GPT-5.5 above the open models, while comment context yields only a small and mixed Core4 gain. Code and data will be released upon acceptance.
[CV-51] FIDA: Feature Instability-Driven Attack on Self-Supervised Facial Representation
链接: https://arxiv.org/abs/2608.26861
作者: Zhiyang Chen,Changchun Yin,Huiqin Yang,Liming Fang
类目: Computer Vision and Pattern Recognition (cs.CV); Cryptography and Security (cs.CR)
备注:
Abstract:Self-supervised learning (SSL) models are vulnerable to backdoor attacks. However, the systemic risks they pose in face representation have received little attention. The entanglement of identity features in self-supervised face learning presents unique challenges for attack stealthiness. To address this gap, we propose FIDA (Feature Instability-Driven Attack), a novel backdoor attack framework. FIDA uses subtle semantic triggers for injection, but its key innovation is a novel objective called Feature Instability Loss. It trains the encoder to increase the sensitivity of triggered features along perturbation directions sampled during attack optimization . By preventing the backdoor from exhibiting the rigid feature patterns typical of previous attacks, FIDA effectively evades the evaluated perturbation-based defenses. Experiments show that FIDA achieves a high attack success rate and generally preserves benign utility across the evaluated settings , posing a significant threat to real-world multimedia applications relying on facial analysis.
[CV-52] A Geometry-Driven Framework-Agnostic Optimization for Object Pose Estimation
链接: https://arxiv.org/abs/2608.26859
作者: Wei Chen,Tao Zhen,Zhongchen Shi,Jing Zhang,Liang Xie,Erwei Yin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to TPAMI, under review
Abstract:Current object pose estimation research remains predominantly model-centric, focusing on architectural innovations and post-processing refinements. This paper introduces a data-centric optimization by proposing a novel, physically grounded rotation representation through principal axes alignment. Our method aligns the object’s coordinate system with its inherent geometric axes, derived from inertial properties, yielding three key advantages: Inherent Stability-leveraging the energy-minimizing property of principal axes provides a robust representation that is less sensitive to noise and occlusions; Symmetry-Aware Canonicalization-explicitly resolving rotational ambiguities for symmetric objects at the data level, which fundamentally eliminates label confusion during network training; and Framework Agnosticism-the optimization is applied purely at the dataset level, ensuring plug-and-play compatibility with existing networks without any architectural modification. We validate the framework across diverse category-level and instance-level models. Extensive experiments demonstrate consistent and significant accuracy improvements, while preserving the integrity of the baseline network. This work establishes a new, geometry-driven direction for enhancing pose estimation, circumventing the need for complex network redesign.
[CV-53] From Reasoning to Pixels: Grounded Medical Multimodal LLM s for VQA and Segmentation ECCV2026
链接: https://arxiv.org/abs/2608.26856
作者: Haowen Gu,Gensheng Pei,Junzhu Mao,Qiong Wang,Mingwu Ren,Yazhou Yao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: accepted by ECCV 2026
Abstract:Although Multimodal Large Language Models (MLLMs) have demonstrated impressive performance in Medical Visual Question Answering (Med-VQA), their reliance on global image features often lacks precise pixel-level grounding, thereby limiting clinical trustworthiness. To bridge the semantic gap between high-level clinical reasoning and spatial localization, we propose \textsc\textscMedREAL (\textbfMedical \textbfREasoning-driven \textbfAnswering and \textbfLocalization), a unified framework that seamlessly aligns linguistic reasoning with spatial grounding. Specifically, \textscMedREAL introduces \textbfSeg \textbfAnchored \textbfReasoning \textbfPooling (SARP) to distill task-relevant semantic evidence directly from \texttt[SEG] tokens within the MLLM’s hidden states. Furthermore, a \textbfReasoning-to-\textbfVisual (R2V) fusion mechanism is proposed to effectively inject these reasoning-aware features into a segmentation pipeline for accurate mask decoding. To facilitate this paradigm, we construct MedRAVS-13K, a comprehensive dataset comprising 13,824 expertly validated samples across four diverse imaging modalities. Extensive experiments demonstrate that \textscMedREAL significantly outperforms state-of-the-arts, achieving 68.49% gIoU and 70.47% cIoU on benchmark evaluations. By generating evidence masks that are strictly consistent with textual diagnoses, \textscMedREAL provides a robust, interpretable framework for reasoning-driven medical image analysis.
[CV-54] MedFG-VQA: Low-Frequency Memory and Graph Attention for Lightweight Medical VQA CVPR2026
链接: https://arxiv.org/abs/2608.26848
作者: Haowen Gu,Gensheng Pei,Zeren Sun,Mingwu Ren,Xiangbo Shu,Yazhou Yao,Fumin Shen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted by CVPR 2026
Abstract:Medical Visual Question Answering (Med-VQA) holds significant promise for clinical decision support, yet faces challenges due to limited annotated data and the high computational demands of existing large vision-language models. We propose MedFG-VQA, a lightweight framework that leverages a memory bank to augment DCT-based low-frequency features and employs graph-enhanced cross-attention for effective visual-textual alignment. Specifically, our approach features two key components: Frequency-Memory Fusion (FMF), which enhances low-frequency features by retrieving from a learnable memory bank built on DCT decomposition, and Graph-Aware Cross-Attention (GACA), which aligns visual-textual features via cross-attention and refines them through graph-convolutional aggregation. To address data scarcity, we construct SynMed-VQA, a large-scale synthetic dataset comprising over 2 million question-answer pairs across 9 imaging modalities and 10 major organs, generated with GPT-4o. Extensive experiments on SynMed-VQA and three other standard biomedical VQA benchmarks demonstrate that MedFG-VQA achieves competitive or superior performance compared to much larger models while maintaining significantly lower computational costs, highlighting its efficiency and potential for clinical deployment.
[CV-55] Rethinking Image Processing for the Age of AI: A Problem-First Framework for Scientific Progress
链接: https://arxiv.org/abs/2608.26833
作者: Guoping Qiu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Modern AI has greatly expanded the capabilities of image processing. However, the ready availability of powerful models, public datasets, and benchmark leaderboards has also en- couraged a model-first research pattern: researchers increasingly begin with an available architecture and optimize it on a public benchmark, rather than beginning with the underlying real-world imaging problem. This can produce impressive benchmark results without necessarily improving our understanding or solution of the real problem. This paper argues for a problem-first approach that distinguishes the physical imaging problem, solution principle, statistical estimator, and computational implementation, while clarifying what modern AI can achieve and which fundamental problems remain unsolved. Through case studies of super- resolution and low-light enhancement, we show how benchmark datasets may define tasks that differ substantially from the real-world problems they are intended to represent, and why performance improvements must be interpreted within the conditions under which they are obtained. We propose a six-stage workflow that places problem formulation, image acquisition, information-loss analysis, assumptions, ambiguity, and evaluation before model and dataset selection. The paper also proposes clearer standards for evidence, reproducibility, uncertainty, and claims of state-of-the-art performance. More fundamentally, it calls for a change in research culture and education so that future researchers learn to understand imaging problems deeply and use modern AI to achieve genuine scientific and technical advancement.
[CV-56] SAGE: Variate-Wise Semantic Augmentation for Vision-Language Time Series Forecasting
链接: https://arxiv.org/abs/2608.26829
作者: Haizhao Fan,Xinyi Le
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 2 figures
Abstract:Time series forecasting models operate on raw numerical sequences, lacking the semantic knowledge that domain experts implicitly leverage, such as the physical meaning of each variable, its statistical behavior, and its temporal dynamics. Recent efforts to bridge this gap fall into two camps. Some rely on large language models at inference time, which is computationally expensive. Others apply uniform textual prompts at the dataset level, ignoring the heterogeneous semantics across individual variates. We propose SAGE (Seeing and Augmenting with Grounded Encoding), an end-to-end CLIP-based framework that jointly models temporal, cross-variable, textual, and visual information. The CLIP text encoder processes frequency-enhanced patches and variable tokens, while gated residual paths inject variable-specific descriptions and statistical descriptors. In parallel, the frozen CLIP vision encoder aligns rendered series with temporal representations through a training-only contrastive objective. This dual use of CLIP adds complementary semantic and visual supervision without placing an LLM in the forecasting loop. Across eight long-term benchmarks and M4, SAGE achieves state-of-the-art accuracy. Ablations confirm complementary gains from multimodal alignment and variable-level knowledge.
[CV-57] Evaluator-Dependent Patient-Adaptive ECG Lead-Channel Allocation
链接: https://arxiv.org/abs/2608.26827
作者: Xiaoyang Li,Zeyan Tao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Patient-conditioned acquisition policies for ECG lead-channel selection can outperform population-wide fixed protocols by tailoring the channel budget to each patient’s observed cardiac state. However, the value of acquiring any given channel is defined relative to a downstream diagnostic evaluator, so marginal utilities learned under one evaluator need not transfer when the evaluator is replaced. We study this evaluator dependence empirically on PTB-XL by freezing two policies (ECG-on-Demand and MGA) trained with a controlled arbitrary-mask logistic evaluator, then scoring their unchanged acquisition trajectories with a more predictive masked raw-waveform ResNet1D. Exhaustive search provides metric-matched population-wide fixed comparators separately for each evaluator, enabling a clean interaction contrast. At budget k=4 on a held-out evaluation fold, ECG-on-Demand shifts from D_\mathrmC=-0.011 (favoring adaptive under the controlled evaluator) to D_\mathrmS=+0.029 (favoring fixed under the strong evaluator), yielding an NLL interaction of +0.041 (95% CI [+0.030, +0.050] ). Across two policies, five budgets, and three probabilistic metrics, all 30 interaction estimates are positive with paired confidence intervals excluding zero. Three post-hoc sensitivity analyses – common-reference scoring, training the strong evaluator on a mixture of policy-generated and random masks, and evaluator-aligned Strong-MGA policy training – each preserve a positive interaction interval, making reference-choice and mask-distribution artifacts less plausible explanations. Evaluator-aligned training reduces but does not eliminate the gap. These results indicate that adaptive ECG channel allocation should be developed and validated jointly with its intended diagnostic backbone, and that jointly optimized sensing-diagnosis systems remain an open problem.
[CV-58] LLaVAFlow: Preserving Latent Alignment Flow for Parameter-Efficient Multimodal Fine-Tuning ACM-MM2026
链接: https://arxiv.org/abs/2608.26820
作者: Muyao Yuan,Muyan Jiao,Jiangyong Ying,Weizhan Zhang,Yuanhong Zhang,Lan Ma,Yuan Gao,Haipeng Du
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by ACM Multimedia 2026 (ACM MM 2026)
Abstract:While Multimodal Large Language Models (MLLMs) exhibit strong generalization, visual instruction tuning for downstream tasks inevitably causes catastrophic forgetting, impairing overall generalization. While existing methods regulate weight updates to reduce forgetting, they overlook the fundamental cross-modal alignment in MLLMs. Based on prior work and our observations, we argue that cross-modal alignment is implicitly captured in the information-compression trajectory. To preserve the alignment flow embedded in the trajectory, we propose LLaVAFlow, an information-theoretic distillation framework. First, we compress the mutual information between the extracted relations and MLLM embeddings, encouraging a learnable module to produce a refined alignment flow that benefits downstream tasks. Second, we maximize the mutual information between the extracted alignment flows of the pretrained and fine-tuned MLLMs, enabling the transfer of compact alignment information. Extensive experiments show that LLaVAFlow is an effective plug-and-play framework that preserves alignment flow and enhances both downstream performance and generalization.
[CV-59] Hyperspectral Diffusion Equivariant Imaging (HyDiff-EI): A Self-supervised Framework for Hyperspectral Image Inpainting
链接: https://arxiv.org/abs/2608.26812
作者: Shuo Li,Mike Davies,Mehrdad Yaghoobi
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 14 pages, 7 figures
Abstract:A novel Hyperspectral diffusion Equivariant Imaging (HyDiff-EI) framework for solving the hyperspectral image (HSI) inpainting problem has been presented here. Unlike conventional diffusion-based methods that rely on large-scale pretraining, HyDiff-EI is a test-time optimization framework that learns directly from a single corrupted HSI acquisition. This makes it flexible for different sensor configurations and particularly well-suited for practical remote sensing scenarios where large annotated hyperspectral datasets are limited. To address the ill-posed nature of unsupervised inpainting, we embed equivariant consistency constraints within the diffusion process. By leveraging the inherent geometric symmetries and intrinsic characteristics of HSIs, HyDiff-EI bridges the gap between generative diffusion modeling and self-consistent physical priors. We empirically show that coupling diffusion modeling with equivariant priors substantially enhances noise robustness and generalizability. Extensive experiments on real-world datasets including Chikusei, Botswana, and EMIT demonstrate that HyDiff-EI offers remarkable inpainting quality over existing self-supervised and diffusion-based algorithms in both noiseless and noisy cases.
[CV-60] hinking on Shots: Consistent Multi-Shot Video Editing with Agent ic Reasoning
链接: https://arxiv.org/abs/2608.26809
作者: Chenyang Wu,Fuchen Long,Binyuan Huang,Xinlong Sun,Xi Chen,Chun-Le Guo,Chongyi Li
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Project Page: this https URL and see source codes at this https URL
Abstract:While generative AI has significantly advanced video editing, existing methods primarily focus on single-shot or short video clips. Editing long videos with multiple instructions remains a formidable challenge. Naive chunking strategies, e.g., fixed-duration segmentation, often lead to entity fragmentation, severe editing hallucinations, and disrupted temporal continuity. To bridge this gap, we introduce the Multi-Instruction Multi-Shot Long-Video Editing (MMLVE) task, which is structured around three core objectives: Cross-Shot Editing Consistency (CSEC), Multi-Instruction Decoupling (MID), and Zero-Destruction on Spatiotemporal Structure (ZDSS). To tackle these three unique challenges, we introduce an agentic editing framework that leverages the synergy of Large Language Models (LLMs) and Vision-Language Models (VLMs) to achieve shot-level video decoupling and precise instruction parsing. Furthermore, to comprehensively evaluate this task, we construct MMLVE-Bench, which is an MMLVE-focused dataset characterized by complex real-world spatiotemporal dynamics, high-density heterogeneous instructions, and sparse, random entity distributions. Three MMLVE-focused evaluation metrics are further exploited to assess the quality of the editing results. Extensive experiments demonstrate that our MMLVE-Agent outperforms existing closed-source SOTA approaches (e.g., Seedance 2.0), successfully eliminating editing hallucinations, preserving cross-shot editing consistency, and attaining seamless spatiotemporal transitions.
[CV-61] Multi-Image Visual Token Pruning in Large Visual Language Models
链接: https://arxiv.org/abs/2608.26806
作者: Rongyang Zhang,Chengqiang Lu,Cong Li,Hongchao Gu,Tingjia Shen,Xuyang Zhi,Qimeng Wang,Yan Gao,Yi Wu,Yao Hu,Hao Wang,Enhong Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 3 figures
Abstract:With the growing demand for processing multiple image sequences in real-world applications, various visual token pruning methods have emerged to mitigate the computational and context length constraints faced by Large Vision Language Models (LVLMs). However, most existing pruning approaches rely on static strategies that struggle to adapt across different architectural LVLMs and multi-image scenarios, and are additionally constrained by their dependence on attention computations that are incompatible with efficient techniques like FlashAttention. To address these limitations, we propose a training-free, Adaptive Visual Token Pruning (AVTP) framework, applicable to diverse LVLM architectures. We strategically determine pruning layers based on empirical analysis of visual attention distributions across various LVLMs, and implement adaptive pruning ratios in multi-image contexts where images of higher importance retain proportionally more tokens. We conduct extensive experiments across different LVLMs to demonstrate the effectiveness and robustness of AVTP. Specifically, Qwen3VL-8B achieves 2 times inference speedup while maintaining 96.1% of its original accuracy on multiple multi-image benchmarks, InternVL3.5-8B retains 94.1% accuracy, and LLaVA-OV-7B even exceeds its original baseline performance. Our code is available at \hrefthis https URLthis link.
[CV-62] Ring Forcing: Towards Precise Long-Term Memory for Autoregressive Video Diffusion
链接: https://arxiv.org/abs/2608.26794
作者: Bowen Xue,Brandon Y. Feng,Chenguo Lin,Yuchen Lin,Yujia Zeng,Lvmin Zhang,Maneesh Agrawala,Honglei Yan,Panwang Pan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Scaling video generation to long durations reveals a critical bottleneck: current models lack robust long-term memory. This deficiency can be studied along two critical aspects: object permanence, the ability to precisely reproduce the appearance of objects upon re-entry; and memory capacity, the ability to process ultra-long context and use information from distant history. Robust long-term memory requires both: object permanence without sufficient context handling limits the temporal scope, while long context length without permanence fails to maintain identity. To address this, we present Ring Forcing, an autoregressive video diffusion framework designed to robustly construct and precisely utilize long-term memory. Our ring-structured training strategy enforces retrieval from distant history, effectively reconciling the trade-off between strict historical adherence and generative diversity. To expand memory capacity, we introduce a compression and timestep composition strategy. Under fixed sequence length constraints, this method extends the effective historical span to minutes-long durations and achieves a comprehensive receptive field over the entire history. Furthermore, we present a sparse RoPE mechanism to enable flexible, scalable memory adaptation while fully exploiting pre-trained priors. Extensive experiments demonstrate that Ring Forcing achieves superior minutes-long coherence and object permanence, significantly outperforming state-of-the-art methods.
[CV-63] Cross-Architecture Knowledge Distillation from a Vision Foundation Model to a Lightweight Visual State Space Model for Tea Leaf Disease Classification
链接: https://arxiv.org/abs/2608.26771
作者: Zibo Zhou,Zongsen Qiu,Rui Chen,Yujie Yao,Yue Zhou,Jianjun Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Automated tea leaf disease classification supports precision agriculture, yet deploying accurate models on edge devices remains challenging under tight compute budgets. Self-supervised vision foundation models such as DINOv2 provide strong features but are too large for field deployment, while lightweight models trained from scratch on small agricultural datasets often underfit. We study cross-architecture knowledge distillation (KD) from a fine-tuned DINOv2 teacher (Vision Transformer) to a compact bidirectional Visual State Space Model (LVSSM) student, an underexplored direction because the architectures use fundamentally different token-mixing mechanisms. We identify and fix two training-stability problems that prevent the from-scratch SSM student from learning on limited data: a single large patch-embedding convolution and a fusion layer that severs the residual path. With a progressive convolutional stem and gated bidirectional selective-scan block, the 4.45M-parameter student trains stably. Across three seeds, temperature-scaled logit distillation raises test accuracy from 92.32+/-2.14% to 95.41+/-1.17% (best single run: 96.20%; macro-F1: 94.45%), a +3.09 percentage-point mean gain. The student uses 5.0 times fewer parameters than the 22M-parameter teacher while retaining 98.3% of its accuracy. Ablations show that intermediate feature-alignment losses reduce accuracy, making simple logit-level KD the strongest configuration. A fair from-scratch comparison shows the gain is specific to students that start below the teacher. We report per-class metrics, confusion matrices, bootstrap confidence intervals, and FLOPs/latency measurements, and discuss limitations including the single-dataset scope and simplified non-official SSM implementation.
[CV-64] Glass Surface Detection Grounded in 3D Visual Geometry
链接: https://arxiv.org/abs/2608.26752
作者: Yiwei Lu,Ke Xu,Tao Yan,Xiaojun Chang,Radu Timofte,Rynson W. H. Lau
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 9 pages, 10 figures. Accepted by ACM Multimedia 2026
Abstract:Glass surface detection (GSD) is critical for scene understanding and reconstruction, and yet remains challenging due to the transparency and reflectivity of glass surfaces. Existing GSD methods typically rely on 2D appearance cues, which may fail in geometrically ambiguous scenes. In this paper, we propose a paradigm shift: grounding GSD in 3D visual geometry to explicitly model the physical existence of glass surfaces. Our method first distills rich 3D priors from the visual geometry grounded transformer (VGGT) and generates glass-aware 3D representations. It then exploits multi-tasking learning with a novel glass detection head, consisting of two core modules: a Frequency Self-Attention Module (FSAM) that identifies glass-specific spectral features for glass surface localization, and a Geometry Grounding Block (GeGB) that selectively grounds 2D features in 3D geometry for glass surface segmentation. Extensive experiments demonstrate that our method achieves state-of-the-art performance across seven standard GSD benchmarks, generalizes well to video/multi-modal data, and substantially improves reconstruction in glass scenes. Code is available in this https URL.
[CV-65] G2D: Generative-to-Discriminative Collaborative Inference for Zero-Shot Image Classification ACM-MM2026
链接: https://arxiv.org/abs/2608.26744
作者: Zehua Hao,Fang Liu,Qinliang Wang,Yaoyang Du,Xinyan Huang,Puhua Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at ACM MM 2026. 10 pages, 5 figures
Abstract:Zero-shot classification needs efficient label retrieval and fine-grained visual reasoning, yet discriminative and generative vision-language models fail in complementary this http URL CLIP’s top-1 prediction is wrong, the correct label often remains in its top- K shortlist, making disambiguation rather than recall the key this http URL generative models, however, are hindered by large label spaces and unconstrained this http URL complementarity motivates separating broad candidate retrieval from fine-grained, image-grounded this http URL propose G2D, a training-free framework that uses a generative VLM to verify CLIP-retrieved candidates against the this http URL names and CLIP probabilities provide a structured prior for resolving visually similar this http URL confidence routing, entropy-adaptive candidate sizing, and trie-constrained decoding focus generative reasoning on uncertain samples and ensure one valid output for each input at test this http URL eight benchmarks, G2D achieves 68.85% average accuracy, versus 59.35% for CLIP and 63.11% for the standalone this http URL seven generator configurations, candidate-set verification improves average accuracy by 1.08–27.42 percentage points.G2D also transfers to DCLIP, WaffleCLIP, and CuPL, supporting a practical interface between discriminative proposal and generative visual reasoning. Code: this https URL
[CV-66] Generative Semantic Scene Completion
链接: https://arxiv.org/abs/2608.26737
作者: Shi Chen,Weifeng Ge
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注: 18 pages, 12 figures, 4 tables. Supplementary material (29 pages) is included as an ancillary file. Project page: this https URL - Code, models and the PS3 dataset: this https URL
Abstract:Outdoor LiDAR semantic scene completion (SSC) recovers a dense semantic voxel grid from a scan observing 1% of the target volume, under class imbalance beyond 7,000x. We recast SSC as generative semantic scene completion (GSSC): a single discrete-diffusion formulation in three roles. First, paired sparse-dense scene synthesis (PS ^3 ) generates matched sparse LiDAR observations with their dense semantic completions, addressing the long tail at its source and yielding the PS ^3 -SemanticKITTI corpus we train on alongside SemanticKITTI. Second, semantic-guided generative scene completion (SGSC) generates the scene from noise with multinomial discrete diffusion, conditioned on the sparse scan through a bird’s-eye-view semantic map and a sparse 3D feature stream. Third, the same framework instead refines an existing completion in one flow-matching step: structured source discrete diffusion (S ^2 D ^2 ). S ^2 D ^2 improves the mIoU of SGSC’s own output and every external SSC base tested, without base retraining or test-time adaptation. On the strongest base, one step without test-time augmentation reaches 38.8% mIoU on the SemanticKITTI hidden test. To our knowledge that is the best causal, single-sweep, single-sample result on that leaderboard, +2.1 pp over the previous best published score under the same restriction. Four correction steps with eight-view test-time augmentation reach 39.2%, outside that restriction.
[CV-67] GeoMAD: Geometry-Aware Multi-View Anomaly Detection via Deformable Fusion and Distributional Alignment
链接: https://arxiv.org/abs/2608.26724
作者: Shang-Fu Chen,Jhih-Ciang Wu,Kuan-Chuan Peng,Wen-Huang Cheng,Kai-Lung Hua
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multi-view anomaly detection (MvAD) detects defects by exploiting complementary observations from multiple camera viewpoints. The central challenge is to fuse views with sufficient geometric awareness while remaining scalable to multi-class industrial settings. Existing methods typically fall into two extremes: voxel-based fusion provides explicit geometric alignment but requires costly 3D construction and class-specific assumptions, whereas lightweight patch-based fusion is efficient but relies on discrete candidate matching and lacks continuous cross-view correspondence. In this paper, we propose GeoMAD, a unified multi-view, multi-class AD framework that addresses both geometric correspondence deficiency and distributional inconsistency. Our \textitCross-view Deformable Fusion Module (CDFM) learns content-adaptive, view-pair-specific sampling offsets directly on 2D feature maps and arranges them across a multi-scale window pyramid with image-global reference sampling, enabling hierarchical cross-view correspondence without camera calibration, voxel construction, or class-specific 3D supervision. We further introduce \textitDistributional View Alignment (DVA), a self-supervised cross-view regularization loss that aligns each view’s bottleneck distribution against a per-instance view-centric target, enforcing global consistency without pixel-level correspondence. Together, CDFM and DVA bridge local geometric correspondence and global distributional consistency, providing geometry-aware and distribution-consistent fusion while preserving the efficiency of 2D feature-space learning. Extensive experiments on Real-IAD and MANTA-Tiny show that GeoMAD achieves strong detection and localization performance in unified MvAD.
[CV-68] UniGeo: A Multi-modal Large Language Model for Text-Guided Cross-View Geo-Localization
链接: https://arxiv.org/abs/2608.26722
作者: Jiahao Wen,Hang Yu,Zhedong Zheng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Text-guided drone geo-localization aims to identify a target region in a large-scale image gallery from a natural-language description. Existing methods mainly formulate this task as direct matching between an open-ended text query and candidate images. However, incomplete queries and highly similar candidates often make global cross-modal matching insufficient for reliable fine-grained localization. We propose UniGeo, a unified multimodal large language model (MLLM) for text-guided drone geo-localization. Built on a shared vision-language framework, UniGeo jointly supports geo-semantic understanding, cross-view semantic generation, and candidate-level verification. Specifically, it establishes stable correspondences among local scene elements, spatial relations, and language descriptions through geo-semantic learning, and further models semantic mappings between drone and satellite views through cross-view generation. Based on these capabilities, a plug-and-play verification module performs fine-grained discrimination among highly confusable candidates. We further introduce a multi-stage training strategy that progressively learns geo-semantic understanding, cross-view generation, and candidate verification, improving adaptation to text-guided geo-localization. Experiments demonstrate consistent improvements across multiple retrieval backbones. On GeoText-1652, UniGeo improves R@10 and mAP by 13.59 and 2.83 percentage points, respectively, validating its effectiveness for fine-grained text-guided drone geo-localization.
[CV-69] Parameter Efficient Continual Learning for Sparse Event-Based Transformers
链接: https://arxiv.org/abs/2608.26720
作者: Vaishnavi Nagabhushana,Kartikay Agrawal,Ayon Borthakur
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Robotic and edge intelligence systems operate in dynamic environments where data arrives continuously, requiring models to adapt while preserving previously learned knowledge under strict memory and energy constraints. While parameter-efficient fine-tuning has shown promise for continual learning with vision transformers, conventional architectures rely on dense computation and remain costly for real-world deployment. Sparse event-based vision transformers provide energy-efficient event-driven computation, yet their continual learning capabilities remain largely unexplored. We here introduce sLoTh, a parameter-efficient continual learning framework for pretrained sparse event-based (spiking) vision transformers. sLoTh freezes the backbone and restricts plasticity to scalable-efficient low-rank attention updates (seLoRA) and shared neuronal threshold modulation, enabling adaptation without replay buffers by updating less than 1% of model parameters. Experiments across CIFAR-100, Tiny-ImageNet, ImageNet-100, and ImageNet-R with up to 100 tasks demonstrate competitive rehearsal-free performance in class-incremental learning and online continual learning, while enabling approximately 6.5x lower energy consumption than conventional dense vision transformers.
[CV-70] Beyond Atomic Layouts: Compositional Design Understanding with Vision-Language Models ECCV2026
链接: https://arxiv.org/abs/2608.26716
作者: Yiyang Huang,Zhaowen Wang,Simon Jenni,Jing Shi,Yitian Zhang,Yizhou Wang,Yun Fu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ECCV 2026
Abstract:Layout understanding, or the interpretation of element organization, is essential for document analysis, user interface (UI) creation, and graphic design. While recent vision-language models (VLMs) excel at interpreting atomic layouts composed of independent elements, they struggle with compositional layouts that require reasoning over visually entangled elements within hierarchical multi-layer structures. In this paper, we introduce a new task, compositional layout understanding, and present CoDeLayout, a VQA dataset of ~20K real-world multi-layer layouts annotated with compositional element pairs and design intent. Through empirical analysis on CoDeLayout, we identify two key challenges for existing VLMs: semantic drift between textual metadata and visual content, and structural ambiguity in hierarchical inter-element relationships. To address these challenges, we propose MASON, a post-training paradigm that integrates multimodal alignment (MA) and structural perception (SP). MA enhances element interpretation by grounding metadata-defined elements to their visual counterparts, mitigating semantic drift, while SP models layer-aware inter-element spatial relationships to improve hierarchical understanding and reduce structural ambiguity. Experiments reveal substantial gaps in existing VLMs: even the strongest baseline, GPT-o3, achieves only 79.68% accuracy, whereas Qwen2.5-VL 7B with MASON reaches 91.66%. Notably, MASON surpasses full-data Direct Finetune using only 30% of the training data and scales better with additional data.
[CV-71] LiveVVT: High-Fidelity Video Virtual Try-On in Real Time
链接: https://arxiv.org/abs/2608.26714
作者: Yushe Cao,Shikun Feng,Ruxiang Duan,Liyong Wang,Dianxi Shi,Chun Yu,Junliang Xing
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 16 pages, 13 figures,
Abstract:Diffusion-based Video Virtual Try-On (VVT) achieves high visual fidelity through bidirectional spatio-temporal modeling, but complete-clip dependence incurs prohibitive latency and computational overhead in practical continuous deployment. Naively enforcing causality disrupts pretrained bidirectional priors and substantially degrades synthesis quality. We introduce LiveVVT, a rolling streaming diffusion framework that preserves bounded bidirectional modeling within causal recurrent generation. Within a fixed-size window, LiveVVT jointly denoises multiple video chunks under bounded look-ahead, preserving local bidirectional interactions while emitting one clean chunk per iteration. Beyond the window, two complementary memories sustain long-term consistency: a bounded temporal memory propagates recent dynamics and occlusion context, whereas a persistent global appearance memory, constructed once from the target garment and a frontal try-on keyframe, anchors garment details and dressed appearance throughout the stream. We further introduce a progressive distillation framework integrating bidirectional VVT learning, teacher-trajectory regression for causal few-step adaptation, and Collaborative Matching Distillation, which couples teacher-distribution matching with rolling flow matching on real videos to align optimization with recurrent inference. Experiments on paired and unpaired long-sequence benchmarks demonstrate superior generation quality over similarly sized models, with 26\times lower latency and 11\times higher throughput, enabling high-fidelity real-time streaming VVT.
[CV-72] AesCanvas: A Large-Scale Dataset and Benchmark for Aesthetic Critique and Contextual Suitability
链接: https://arxiv.org/abs/2608.26713
作者: Xuanwei Hu,Haoyu Dong,Kejun Wu,Tianyi Liu,Jianjun Gao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 10 pages, 4 figures, 6 tables. Supplementary material included
Abstract:Recent advances in Multimodal Large Language Models (MLLMs) have extended Image Aesthetic Assessment (IAA) beyond scalar scores toward interpretable critique and guidance. Yet existing benchmarks mainly assess intrinsic visual quality or fixed domain criteria, leaving open whether an appealing image is appropriate for a specific purpose, audience, cultural setting, or domain convention. We introduce AesCanvas, a unified suite with two complementary components: CritiqueCanvas with 519,136 instruction-response pairs from 54,300 images supports long-form, multi-dimensional critique across photography, painting, and virtual imagery, whereas ContextCanvas with 301 expert-reviewed use scenarios evaluates contextual aesthetic suitability in realistic use scenarios. Under a unified protocol, we evaluate closed-source frontier, open-weight general, and aesthetic-specific MLLMs. Results reveal a clear separation between critique generation and context-sensitive judgment: reference-based lexical and semantic metrics only partially capture critique quality, while aesthetic specialists remain competitive on selected critique metrics yet substantially lag strong general-purpose MLLMs on ContextCanvas. Further analyses show that aesthetic specialization does not reliably transfer to contextual suitability and that model decisions may fail to track or ground themselves in decisive contextual visual cues. These findings establish culturally situated, evidence-grounded suitability as a distinct objective for aesthetic modeling.
[CV-73] Domain-Specific Self-Supervised Representation Learning for Retinal Fundus Classification
链接: https://arxiv.org/abs/2608.26686
作者: Bekzat Nurlanbekova,Fung Fung Ting
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Despite the growing number of public datasets, annotated medical images remain scarce. Supervised learning methods achieve strong performance on many benchmarks, however require large amounts of labeled data, which are costly and time-consuming to obtain in the medical domain. To address this limitation, contrastive self-supervised learning (SSL) has emerged as a promising alternative for learning useful representations from unlabeled data. In this work, we investigate two SSL frameworks, SimSiam and SimCLR, for retinal disease classification from fundus images. We focus on understanding how augmentation strategies and training parameters influence representation learning under resource-constrained settings. Given limited data and computational capacity, we explore the feasibility of training SSL models with small batch sizes incorporated with retinal-specific augmentation techniques. Through a series of experiments, we assess the quality of learned representations via linear evaluation and fine-tuning across downstream tasks, including multi-disease classification and diabetic retinopathy grading. Our results show that tailoring augmentation strategies to the characteristics of retinal images plays a critical role in improving performance. Even under constrained settings, lightweight SSL frameworks can learn transferable representations that reduce dependence on large annotated datasets and achieve competitive results.
[CV-74] Reason in the Words You Speak: Idiolectal Paraphrasing Off-Policy Traces for Reasoning Distillation in VideoLLM s
链接: https://arxiv.org/abs/2608.26684
作者: Ji Soo Lee,Jinyoung Park,Seohyun Lee,Jongha Kim,Joonmyung Choi,Jinsung Yoon,Hyunwoo J. Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Work in progress
Abstract:Recent large language models achieve strong performance on complex reasoning tasks, where reinforcement learning with Group Relative Policy Optimization (GRPO) has emerged as a leading paradigm for optimizing models on self-generated trajectories. However, the on-policy nature of GRPO bounds the model to the reasoning skills it can already produce, restricting to learn more advanced capabilities. Prior works inject privileged reasoning traces from a stronger teacher policy to guide training, yet these traces are inherently out of distribution with respect to the student policy. We observe that this mismatch between on-policy and off-policy causes gradient clipping on semantically critical reasoning tokens, ultimately rewarding correct answers while leaving the reasoning that justifies them unlearned. Hence, we propose \textbfEcho-GRPO, a framework that lets the model reason in the words it speaks. Rather than imitating low-probability privileged traces from the teacher model, Echo-GRPO rewrites them into the student policy’s own \textitidiolect, that is, its own characteristic vocabulary and expression patterns, while preserving their semantics via Dual-Reference Decoding. We instantiate this framework as \textbfVideoEcho-R1 for video reasoning distillation, achieving consistent improvements across three multimodal LLM backbones and five benchmarks. Finally, we show that our idiolectal paraphrasing is a plug-in module that consistently improves both RL and supervised fine-tuning frameworks for reasoning distillation, demonstrating that policy-aligned supervision extends beyond GRPO.
[CV-75] RECAP-Forcing: Retaining Content Appearances for Long Video Generation
链接: https://arxiv.org/abs/2608.26671
作者: Haiyang Xu,Zheng Ding,Zhuowen Tu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Long autoregressive video generation faces a fundamental memory challenge: with a finite attention window, a model must decide which information from an ever-expanding history to retain. Existing methods organize memory temporally, preserving recent frames while compressing or discarding older ones. We instead propose RECAP-Forcing, organizing memory by appearance novelty. A long video is not merely a sequence of frames, but an evolving cast of subjects, objects, and scenes whose identities must remain consistent over time. We organize memory by retaining the KV cache associated with newly appearing content–such as entering subjects, disoccluded regions, and newly introduced scenes–at the moment it first becomes visible, prioritizing novelty over recency. Memory should scale with the amount of newly introduced content, rather than with video length. This appearance-indexed memory makes long-range consistency an explicit property of the memory structure. Our framework unifies two mechanisms under this single principle. At the beginning of a video, when all visible content is novel, an attention sink preserves the initial scene. As the video evolves, an optical-flow-based novelty bank extends the same principle by selectively retaining newly revealed content. As a training-free inference method with no additional learnable parameters, RECAP-Forcing consistently improves visual quality and semantic fidelity across multiple strong baselines and outperforms existing memory methods.
[CV-76] Beyond the Proving Ground: Independent Public-Road Testing of Assisted Lane Change Systems using LiDAR
链接: https://arxiv.org/abs/2608.26669
作者: Marcello Cellina,Akos Kriston,Antonio Migneco,Davide Maggi,Stefano Favelli,Fabrizio Re,Fabrizio Minarini,Andrea Nuovo,Riccardo Dona,Biagio Ciuffo
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Testing of commercial Advanced Driver Assistance Systems is essential to ensure safety and compliance during type approval and in service operation. However, proving ground scenarios may not reflect real world driving complexity, while geo fencing can require manufacturer collaboration and limit assessment independence. This work presents a methodology for independently testing Assisted Lane Change systems on public roads. A campaign on the A31 French motorway used a test vehicle equipped with a LiDAR based vehicle detection and tracking system. Tests covered combinations of inter vehicle distance and speed between the test vehicle and the take over vehicle. Real time kinematic global navigation satellite system receivers assessed detection and tracking performance. Recorded lane change trajectories were compared with the lane change suppression requirements of UNECE Regulation Number 79. Of 27 predefined lane change manoeuvres, 18 were completed and 9 suppressed. In 6 cases, the system allowed manoeuvres that did not meet regulatory minimum distance requirements. In 3 cases, the deviation remained statistically significant after accounting for measurement uncertainty. To the authors knowledge, this is the first public road campaign designed to assess Assisted Lane Change compliance with Regulation Number 79 safety distance requirements. The results demonstrate the suitability of LiDAR based sensing for this purpose. The methodology can support market surveillance and future regulatory revisions by revealing real world behaviours not covered by approval procedures.
[CV-77] Hull First Wake Second: Wake-Reliance Suppression for Robust Maritime Vessel Detection
链接: https://arxiv.org/abs/2608.26665
作者: Yefan Wang,Xingyu Wang,Ruibiao Zhu,Yusen Wu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by ICIG 2026 (The 14th International Conference on Image and Graphics)
Abstract:Maritime vessel detectors often face scenes where hulls are small, low-contrast, or blurred, while wakes are longer and easier to detect. This creates a wake-reliance problem: detectors may miss slow or stationary vessels with weak wakes, or produce false positives on wake-like water clutter. We propose HullWake, a hull-first wake-second framework for robust maritime vessel detection. HullWake separates proposal-centered hull evidence from directional wake context, extracts wake cues with bidirectional proposal-anchored corridors, and suppresses wake-dominant predictions through wake response supervision, wake-attenuated consistency, wake-only confidence suppression, and hull–wake decorrelation. We also introduce a wake-oriented evaluation protocol covering weak/no-wake vessels, wake-like hard negatives, worst-group AP, and confidence drop after wake attenuation. Experiments are conducted on Curated-Wake, a wake-oriented maritime dataset of about 10,000 images curated from Ships/Vessels in Aerial Images, the SMD benchmark, and SeaDronesSee, with newly added detection- and segmentation-level wake annotations. Compared with box-only detectors and mask-supervised segmentation baselines, HullWake improves overall AP, weak/no-wake robustness, wake-like false positives, worst-group AP, and confidence stability after wake attenuation.
[CV-78] CoGeo-GS: Concept-Driven and Geometry-Aware Multi-Object Removal in 3D Scenes ICME2026
链接: https://arxiv.org/abs/2608.26656
作者: Yuanxiang Ni,Xianliang Huang,Chenhang Ma,Chen Xiao,Yuewen Ma,Ruxin Wang,Hao Zhang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 6 pages, 4 figures, accepted at ICME 2026
Abstract:Multi-object removal in 3D scenes is challenging due to severe occlusions, semantic entanglement, and the difficulty of maintaining geometric and multi-view consistency. Existing 3D Gaussian Splatting (3DGS) methods perform well for single-object editing but scale poorly to multi-object scenarios, often requiring repetitive optimization and yielding unstable geometry in removed regions. We propose CoGeo-GS, a concept-driven framework for controllable multi-object removal in 3D scenes. CoGeo-GS assigns concept-aware semantic tags to Gaussians, enabling flexible object selection and reducing interference between foreground objects and background structures within a single optimization stage. To recover plausible geometry, we introduce a geometry-aware completion pipeline that combines monocular depth priors with diffusion-based refinement and boundary-aligned blending. A geometry-regularized refinement strategy further stabilizes reconstruction and preserves multi-view consistency. Experiments demonstrate that CoGeo-GS outperforms existing methods in visual quality and reconstruction fidelity.
[CV-79] Hierarchical Channel Stacking: A Structured Decision Framework for AI-Generated Image Detection ICANN2026
链接: https://arxiv.org/abs/2608.26648
作者: Saifullah Shoaib,Akash Borigi,Rupendra Lekkala,Amaury Lendasse,Edward Ratner,Sai Sowjanya Bhamidipati,Alexander Schlager,Peggy Lindner
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 12 pages, 4 figures. Accepted for publication at ICANN 2026
Abstract:Many synthetic-image detectors produce accurate predictions but offer limited insight into how those decisions are formed. This paper introduces Hierarchical Channel Stacking (HCS), a compact framework for AI-generated image detection that converts intermediate CNN activations into a structured 60-dimensional representation organized across three progressively deeper backbone stages. HCS uses per-channel Level-1 classifiers and a Level-2 aggregator to produce image-level predictions while preserving explicit hierarchical structure for analysis. On a benchmark spanning GAN and diffusion generators, HCS achieves 86.7% accuracy and 86.7% macro-F1 on the held-out test set. Stage ablation shows that the full three-stage system outperforms reduced single-stage and two-stage variants, indicating that the hierarchy carries complementary predictive information. Stage-level contribution analysis further shows that, in the analyzed detector setting, fake GAN and fake diffusion images exhibit distinct stage-level contribution profiles. These results position HCS not simply as a compact detector, but as a structured framework for studying how synthetic-image detectors assemble evidence across representation levels.
[CV-80] ssue-Mixture Entropy-Weighted Reconstruction for Partial-Volume-Aware Brain MRI Super-Resolution
链接: https://arxiv.org/abs/2608.26647
作者: Xiao Tong,Wenyun Yang,Ziheng Zhang,Jingzhi Han,Zhaochu Luo,Jinbo Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 6 figures, 8 tables
Abstract:Full-image objectives in brain magnetic resonance imaging (MRI) super-resolution (SR) can underweight tissue-transition regions affected by the partial-volume effect (PVE), as these regions occupy only a small fraction of the image. Binary boundaries also do not capture the continuous mixture of cerebrospinal fluid, gray matter, and white matter within a voxel. We propose Anatomy-Guided Gaussian-Parameter Warping with PVE-Balanced Reconstruction (AGW-PBR), which combines a low-resolution (LR)-only reconstruction backbone with a training-time objective that emphasizes tissue transitions. The backbone integrates LR-derived Sobel guidance, soft latent-basis assignment, and bounded grid-anchored residual warping. Fixed, quality-controlled tissue fractions derived from registered T1/T2/PD IXI images are converted into tissue-mixture entropy, which defines mean-normalized reconstruction weights within validated PVE support. These sidecars are used only during training, and inference requires only the LR image. AGW-PBR is evaluated on T2-weighted IXI images at 2x, 4x, and 6x using three seeds and subject-level paired analyses. At 4x, test-only SynthSeg masks independently assess reconstruction in tissue-interface and non-interface regions. Targeted ablations examine valid-support supervision, spatially aligned entropy weighting, and soft latent assignment. The AGW-backbone is also trained from scratch on fastMRI at 4x without PVE supervision. AGW-PBR improves full-image reconstruction across the tested IXI scales and regional fidelity at 4x, while the PVE-free backbone retains strong performance on fastMRI. These findings support tissue-mixture entropy weighting for partial-volume-aware brain MRI SR.
[CV-81] Real-time Unsupervised Object Discovery from Asynchronous Event Streams
链接: https://arxiv.org/abs/2608.26644
作者: Pratham G. Shenwai,Hemant Kumar Singh,Sridhar Ravi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Event cameras capture pixel-level intensity changes with microsecond resolution to produce highly sparse asynchronous data streams. For visual perception in latency-critical environments, we propose a lightweight, training-free framework for discovery of moving objects based on spatio-temporal clustering. This framework is driven by two core contributions. First, a linear-time Spatio-temporal Probabilistic Event Filter (SPEF) that introduces an adaptive event acceptance threshold to distinguish salient motion structures from background noise. Second, an Event Morton Code Clustering (EMCC) module that bypasses expensive distance matrix computation to efficiently group events for unsupervised discovery of moving objects. On the E-MLB dataset benchmark, SPEF achieves the best denoising performance among classical filtering methods and remains competitive with learning-based approaches without requiring any offline training. On object discovery, EMCC achieves the highest overall accuracy and lowest execution time across the FRED and eTraM datasets, outperforming established density-based clustering baselines by a substantial margin. Overall, this work establishes a new performance benchmark for classical object discovery in event data, providing a highly scalable, training-free solution for resource-constrained visual perception. The code is available at this https URL
[CV-82] Who Remains What Changes: Identity Anchored Composed Gait Retrieval
链接: https://arxiv.org/abs/2608.26632
作者: Jingchen Fei,Zengbin Wang,Yukun Liu,Muyi Sun,Shibiao Xu,Man Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Gait recognition has achieved remarkable progress, yet existing methods remain confined to rigid visual matching and often overlook the potential of natural language instructions for interactive retrieval. In this paper, we introduce Composed Gait Retrieval (CoGR), a novel task that retrieves a target gait sequence based on a reference sequence and a natural language modification query. To address the absence of existing datasets for this task, we design an automated annotation pipeline powered by large vision-language models (VLMs) to construct the first gait-language datasets: Language-Augmented CCPG and Language-Augmented CASIA-B. Building on this, we propose ComposeGait, an identity-anchored composition framework designed to prevent the identity drift that arises when generic composed retrieval follows the instruction but returns the wrong person. Its Part-aware Identity Adapter (PIA) aggregates multi-frame, part-aware identity evidence into a sample-specific ID token. We inject the ID tokens into both branches of a shared Q-Former to preserve identity, while excluding the ID-token outputs from the final retrieval embeddings. Joint identity and task-adapted composed-retrieval objectives optimize this space end to end. We evaluate ComposeGait on both benchmarks and show that it achieves the best R@1 among the compared methods, reaching 72.38% on Language-Augmented CCPG and 83.61% on Language-Augmented CASIA-B. These results establish ComposeGait as a strong baseline for CoGR. The datasets and code will be made publicly available.
[CV-83] xt-to-seed generation: Training-free open-vocabulary seeded semantic segmentation via re-purposing diffusion as text-guided seed generator
链接: https://arxiv.org/abs/2608.26624
作者: Kumju Jo,Heesun Jung,Sungyong Baik
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Preprint version of the work accepted for publication in Knowledge-Based Systems
Abstract:Open-vocabulary semantic segmentation (OVSS) aims to segment image regions corresponding to arbitrary text queries. Although the Segment Anything Model (SAM) is a powerful foundation model for segmentation, its standalone performance on OVSS remains limited. Existing methods therefore often use SAM to refine coarse masks predicted by other models, but this strategy is unreliable when the initial masks are inaccurate. In this work, we argue that more reliable segmentation can be achieved by exploiting SAM as a region expansion module guided by accurate object points (i.e., seeds) rather than inaccurate coarse masks. Inspired by classical seeded segmentation, we reformulate OVSS as text-guided seed localization followed by seed-based region expansion. To realize this idea, we propose Text-to-Seed (T2S), a training-free framework that leverages the text-to-region correspondence of Stable Diffusion to generate attention-based seed points for target categories described by text. These sparse seeds are then used as point prompts for SAM to produce full object masks. Without task-specific training or additional annotations, T2S achieves strong performance on standard OVSS benchmarks, demonstrating the effectiveness of combining semantic grounding with seed-driven spatial segmentation.
[CV-84] FU-Mamba: A Frequency-Enhanced Dynamic Scanning Framework for Oralscan Image Segmentation
链接: https://arxiv.org/abs/2608.26607
作者: Xinxin Zhao,Jinpeng Ye,Bo Wei,Liqin Wu,Mahmoud Hassaballah,Karen Egiazarian,Aura Conci,Victor Hugo C. de Albuquerque,Abdulkadir Sengur,Leszek Rutkowski,Yan Tian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by Neurocomputing
Abstract:Oralscan image segmentation is essential for computer-aided diagnosis and treatment planning in digital dentistry. However, existing visual state space models (SSMs) often rely on manually designed scanning orders to flatten image patches into sequences, which disrupts the semantic spatial continuity and hinders coherent feature extraction from key foreground regions. Moreover, elements such as inconsistent lighting, reflective surfaces, and noise during data acquisition disrupt the frequency distribution by diminishing high-frequency details while enhancing low-frequency components, consequently hindering the accurate localization of boundaries. In response to these challenges, we introduce FU-Mamba, an innovative framework that incorporates dynamic scanning and frequency domain enhancement within the SSM architecture. Specifically, the Dynamic Mamba Block (DMB) adaptively learns sampling offsets via a trainable offset prediction network and performs flexible bilinear interpolation, enabling content-aware scanning that preserves spatial coherence. Furthermore, a frequency domain enhancement block balances spectral components through wavelet-guided decomposition and spectrum pooling, improving robustness under adverse imaging conditions. Experimental findings indicate that FU-Mamba attains a notable enhancement in segmentation accuracy, evidenced by a 1.1% increase in the mean intersection over union (mIoU) metric when evaluated on the dental segmentation dataset. Project page: this https URL
[CV-85] DPA-I2P: Depth-Guided Projective Alignment for Image-to-Point-Cloud Registration in Autonomous Driving
链接: https://arxiv.org/abs/2608.26589
作者: Wenxin Zhang,Hang Li,Zhiwei Xu,Qiankun Dong,Gang Wang,Tao Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Image-to-Point Cloud Registration aims to estimate the camera pose of a given image within a 3D scene point cloud, which is a fundamental task in autonomous driving and large-scale outdoor localization. Recent implicit correspondence learning methods have improved registration performance by learning cross-modal alignment in an end-to-end framework, leading to more accurate camera pose estimation. However, due to the inherent modality discrepancy between images and sparse LiDAR point clouds, reliable cross-modal correspondence learning remains challenging. To address this issue, we propose Depth-Guided Projective Alignment for Image-to-Point-Cloud Registration (DPA-I2P). Unlike naive depth or feature concatenation, Ray-Conditioned Metric Depth Encoding (RMDE) and Projection-Consistent Vision Lifting (PVL) exploit depth and visual cues in a structured, geometry-aware manner. In addition, Cross-Modal Query Pruning (CQP) suppresses unreliable queries during early refinement to improve matching stability. Experiments on KITTI and nuScenes demonstrate the effectiveness of the proposed method. On KITTI, DPA-I2P reduces RTE and RRE by 45.0% and 55.6% over the strongest implicit baseline, respectively. On nuScenes, DPA-I2P also improves registration accuracy over the evaluated baselines, suggesting better transferability to different driving scenes.
[CV-86] rapVLA: Trapping Vision-Language-Action Models in Configured Failure Modes
链接: https://arxiv.org/abs/2608.26578
作者: Jun-Hui Liu,Kun-Yu Lin,Yi-Lin Wei,Xu-Han Chen,Yinghao Li,Zhuohao Li,Yuan-Ming Li,Qing Zhang,Xiaoyi Fan,Dongmei Jiang,Yan Li,Wei-Shi Zheng
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:This work introduces Configured Failure Trapping, a novel backdoor attack task against Vision-Language-Action (VLA) models, which aims to activate attacks through stealthy textual triggers and induce configured failure modes. Unlike prior backdoor attacks that treat any task failure as a successful attack, Configured Failure Trapping requires the attacker to control how the robot fails (e.g., causing the robot to grasp with a specified positional offset), making it substantially more challenging and hard to detect. To support the new task, we propose an effective data engine for synthesizing high-quality target trajectories and an automated suite for measuring configured-failure fidelity. Then, based on this foundation, we construct two new benchmarks, namely Trap-LIBERO and Trap-RoboTwin, that instantiate Configured Failure Trapping across four representative failure modes. To address this task, we identify sparse action deviation as a critical challenge and accordingly propose a novel method named TrapVLA, which explicitly learns trigger-induced action residuals to steer the policy toward the configured failure behavior. Extensive experiments across simulation benchmarks and real-world robotic settings show that TrapVLA effectively injects configured failure modes into VLA models while largely preserving performance on clean data. Project page: this https URL
[CV-87] Camera Calibration Using Inaccurate and Asynchronous Discrete GPS Trajectory from Drones
链接: https://arxiv.org/abs/2608.26548
作者: R. Yang,Y. Bar-Shalom,H.A.J. Huang
类目: ystems and Control (eess.SY); Computer Vision and Pattern Recognition (cs.CV)
备注: 11 pages, 12 figs, published on JAIF
Abstract:This paper considers a stationary camera calibration problem, which estimates the camera orientation angles yaw, pitch and roll, using a drone trajectory recorded by a GPS. There are three challenges in using a GPS trajectory as ground truth for camera calibration. One, the altitude of GPS data is inaccurate with an unknown bias. Two, the GPS receiver and camera are not time synchronized, and there is an unknown time offset between the two systems. Three, the GPS trajectory is time-discrete and accurate interpolation is needed. This is actually an estimation problem since velocity is also needed. To address the first two challenges, we formulate the problem as a parameter estimation problem to estimate a vector consisting of the GPS altitude bias and time offset in addition to the camera yaw, pitch and roll biases. We then develop a special maximum likelihood estimator using the Iterated Least Squares algorithm which can work with a non-synchronized time-discrete GPS trajectory for the third challenge. Since the camera measurement errors are usually small, this requires a high calibration accuracy so that the residual bias error following the calibration should not be significant compared to the measurement error standard deviation. The calibration accuracy depends highly on the drone trajectory. This paper also recommends an appropriate drone trajectory which can yield a good calibration accuracy, namely, 14% of the measurement error standard deviation. Simulation tests are conducted to demonstrate the algorithm performance. The estimation results meet the Cramer-Rao Lower Bound (CRLB) since the Normalized Estimation Error Squared w.r.t.\ the CRLB is statistically acceptable.
[CV-88] FAN-LoRA: A Fourier-Adaptive Nonlinear Low-Rank Adaptor for Medical Foundation Model Domain Adaptation
链接: https://arxiv.org/abs/2608.26531
作者: Ziquan Liu,Zhewei Zhu,Xuyang Shi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 9 pages, 4 figures
Abstract:The advent of vision foundation models, notably the Segment Anything Model (SAM), has catalyzed significant advancements in natural image segmentation. However, their direct transfer to medical imaging remains severely bottlenecked by profound domain gaps, such as cross-modality and cross-center shifts. Existing Parameter-Efficient Fine-Tuning (PEFT) methods facilitate the adaptation of SAM to medical domains; nevertheless, they frequently suffer from performance degradation under severe distribution shifts. This vulnerability primarily stems from the implicit entanglement of heterogeneous frequency components within a shared low-rank subspace, which directly exacerbates sub-optimal structural alignment and localized boundary blurring. To overcome this representational bottleneck, we propose the Fourier-Adaptive Nonlinear Low-Rank Adaptor (FAN-LoRA), a novel frequency-decoupled fine-tuning architecture. FAN-LoRA explicitly separates the optimization space by employing a B-spline-driven low-pass branch for global structural alignment, synergistically coupled with a discrete Fourier high-pass branch for local textural compensation. Extensive experiments across three challenging cross-modality and cross-center benchmarks demonstrate that FAN-LoRA consistently outperforms state-of-the-art PEFT baselines. Compared to the strongest competitors, our method achieves consistent improvements in average Dice scores and notable reductions in boundary errors, while maintaining a compact module size without compromising computational efficiency.
[CV-89] HUG-VIS: A Multimodal Benchmark for Human-centered Understanding and Generation in Visual Intelligence
链接: https://arxiv.org/abs/2608.26517
作者: Fei Ma,Zebang Cheng,Minghui Li,Hongbo Xu,Yuyong Tan,Yihua Shao,Hanling Wang,Zhou Liu,Yuqing Gao,Dong Wang,Long Ma,Laizhong Cui,Nicu Sebe,Qi Tian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Visual intelligence seeks to perceive, interpret, and synthesize the visual world and is central to modern computer vision. Human-centered visual intelligence is especially demanding because it studies people as expressive, socially situated subjects whose meaning is rarely conveyed by appearance alone. It couples vision with audio and language across four representative tasks: human emotion recognition, human video generation, human voice cloning, and human video matting. Yet existing resources remain task-specific, providing modalities and annotations for individual problems rather than a shared foundation coordinating understanding and generation. This limits multimodal signal use and broader research. We address this gap with HUG-VIS, a unified benchmark for Human-centered Understanding and Generation in Visual Intelligence. It contains 8,400 seated half-body videos of 30 professional actors, each performing the same 280 emotion-action-prompt assignments under a controlled Mandarin studio protocol, with synchronized video, audio, text, and alpha mattes. We evaluate diverse open- and closed-source models across the four tasks under a unified zero-shot protocol using automatic metrics, criterion-specific mean opinion scores, and multiple cross-task analyses. Results show that (i) linguistic content dominates current emotion recognition, while purely visual affect recognition is weakest; (ii) in video generation and voice cloning, automatic metrics and human judgment agree overall but differ in their top rankings, requiring joint reporting; (iii) boundary fidelity under motion is the main remaining obstacle for human matting; and (iv) task difficulty varies across emotions, models, and metrics, with notable cross-task correlations. The dataset and results are available at this https URL.
[CV-90] NeuDonatello: Uncertainty-Aware Framework for Accurate Neural SDF Learning BMVC2026
链接: https://arxiv.org/abs/2608.26504
作者: Alvin Jinsung Choi,Wanhee Kim,Taeyun Kim,Dasol Hong,Wooju Lee,Hyun Myung
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to BMVC 2026
Abstract:Neural surface reconstruction has emerged as a powerful paradigm for recovering high-quality 3D surfaces from multi-view images. However, recovering accurate geometry solely from RGB images remains challenging due to uncertainties arising from textureless regions, occlusions, and inherent scene ambiguities. Existing methods often overlook such uncertainties, leading to inaccurate estimates of the signed distance function (SDF). We introduce NeuDonatello, a novel framework that models and leverages SDF uncertainty to improve surface reconstruction. Central to our approach is to model spatially varying uncertainty using a Monte Carlo sampling strategy. Using this uncertainty, we develop an adaptive regularization that selectively strengthens geometric constraints where RGB supervision is unreliable, avoiding incorrect surface reconstruction. We further introduce an uncertainty-aware scale parameter for the SDF-to-density conversion. Conditioned on uncertainty, this design enables more accurate modeling of spatially varying densities. Extensive experiments demonstrate that NeuDonatello achieves state-of-the-art reconstruction accuracy, with robust performance across diverse scenes using only posed RGB images.
[CV-91] Systematic Literature Review of Machine Learning Models and Applications for Text Recognition
链接: https://arxiv.org/abs/2608.26500
作者: Nuzhat Khan,Ab Al-Hadi Ab Rahman,Shahriyar Masud Rizvi,Ibrahim Yousef Alshareef,Muhammad Nadzir Marsono,Muhammad Paend Bakht,Mohd Shahrizal Rusli,Shahidatul Sadiah
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Published in IEEE Access, 2025. 24 pages, 16 figures, 5 tables
Abstract:Optical Character Recognition (OCR) for text recognition using machine vision has significantly improved, particularly when handling heterogeneous textual data. Traditional OCR models struggle with script variations, writing styles, and degraded documents. Advancements in technology are leading to new AI models with improved architecture for handling multiple languages and complex data formats. Despite this progress, a comprehensive evaluation of OCR advancements remains limited. Based on the established preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines, this literature review presents an extensive assessment of OCR research to trace the evolution of AI models over the past decade. It explores the transition in AI models, application domains, data types, linguistic coverage, and challenges. Through a detailed analysis of 97 selected studies published during January 2015 - January 2025, key OCR models are identified, and their performance, strengths, and limitations are analyzed. The findings highlight how OCR technologies have evolved to address structured and unstructured text, scene text recognition, and multilingual processing. Unresolved challenges include limited resources for underrepresented languages, high variability in handwritten text, visual similarity among characters, and constraints in real-time OCR applications. To address these issues, several promising approaches are proposed. Key suggestions include self-supervised learning, multimodal AI, automated machine learning (AutoML), AI-assisted postprocessing, tiny machine learning (TinyML), and the creation of joint corpora for script matching. The future recommendations aim to enhance OCR accuracy and tackle the challenges identified for real-time industrial applications. This study will guide future research and establish a foundation for OCR field.
[CV-92] RTNav: Towards Real-Time Zero-Shot Object Navigation
链接: https://arxiv.org/abs/2608.26496
作者: Easop Lee,Lingyu Zhang,Boyuan Chen
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Navigation in unknown environments to find unforeseen objects has become increasingly feasible with capable vision and language foundation models. However, these models also introduce non-negligible inference latency, which becomes an important concern when agents must operate continuously in the real world. Most state-of-the-art methods are still developed in synchronous simulators, where the environment waits for the agent to act and inference time is effectively free. As a result, agents are often designed around the sequential execution of perception, reasoning, and action, with little regard for time constraints. Under real-time execution, where wall-clock time counts towards the task budget, the inefficiencies of these architectures become clear. We show that recent zero-shot object navigation methods suffer consistent performance degradation under such realistic timing conditions. Motivated by this observation, we propose RTNav, a simple but effective architecture that treats inference latency, asynchronous environment stepping, and bounded compute as explicit design considerations. Evaluated on real-time variants of HM3D-v1, HM3D-v2, and HM3D-OVON, RTNav improves the success rate by up to 11% and the Success weighted by Completion Time by up to 5.1 points over prior work.
[CV-93] Video-FLAIR: Not Whether to Reason But How
链接: https://arxiv.org/abs/2608.26495
作者: Yogesh Kulkarni,Pooyan Fazli
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal queries can require different types of reasoning. Some can be answered via perceptual reasoning, extracting information directly from the visual signal, while others require compositional reasoning that combines observations or deliberative reasoning that evaluates competing hypotheses. However, many existing methods apply a uniform reasoning strategy across queries, leading to unnecessary computation on simple tasks and insufficient reasoning on complex ones. We introduce Video-FLAIR, a training framework that learns to select the appropriate reasoning mode for each query using reinforcement learning. During training, the model generates responses under all three modes for the same prompt, enabling direct comparison. A composite reward compares these responses to favor the most effective one based on correctness, grounding, and cost, while discouraging unsupported or misaligned deliberation. This yields a supervision signal for learning adaptive reasoning without per-query annotations. Video-FLAIR improves accuracy over the Qwen2.5-VL base model by +5.4 on MathVista, +4.8 on Video-Holmes, and +4.8 on Video-MMMU, while reducing average token usage to 95 compared to 417 for always-thinking baselines.
[CV-94] Learning Woody Clearing With Loss Alignment for Zero-Shot Regrowth and Woody Segmentation
链接: https://arxiv.org/abs/2608.26489
作者: Kal Backman,Jared Wood,Adam Roff
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Published in IEEE Transactions on Geoscience and Remote Sensing
Abstract:Detecting woody clearing is vital for managing biodiversity. Deep learning models can detect change in woody vegetation from bitemporal remote sensing imagery, however generated products may not meet end-user specifications due to unaligned loss definitions. Further limitations of deep learning models are the reliance on large datasets which can be difficult to attain for spatially rare and ambiguous events such as regrowth detection. In this work we train a model to detect woody change using bitemporal Sentinel-2 imagery consisting of 7 years’ worth of annual imagery across the state of New South Wales, Australia. To align the objective of the model with end-user metrics, we introduce the loss scaling coefficient \alpha which transforms the objective to optimize for specific F_\beta scores. Introducing \alpha was found to increase precision by 1.85x or recall by 1.12x. We propose input imagery augmentation and generation techniques that allow the woody change detection model to zero-shot transfer to regrowth and woody segmentation tasks. For woody segmentation, image generation techniques using activation maximization with low \alpha values for stability and image generation techniques derived from handcrafted features utilizing a mosaic of clearing patches and artificial trees for contextual grounding were found to outperform prior woody segmentation works of the study area, reducing the overall error by up to 18.2%. For zero-shot woody regrowth, creating pseudo-post and prior images resulted in the model achieving an F1 score of 0.845, creating a foundation for future regrowth detection work.
[CV-95] Zero-Shot Video Restoration and Enhancement with Text-to-Image Latent Diffusion Models and Multi-Modal References
链接: https://arxiv.org/abs/2608.26476
作者: Cong Cao,Huanjing Yue,Xin Liu,Jingyu Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Zero-shot image restoration methods with text-to-image latent diffusion models have achieved great success in universal image restoration tasks without training. However, applying them to video restoration will result in severe temporal flickering. In this paper, we propose a novel framework for zero-shot video restoration and enhancement which uses a text-to-image latent diffusion model and multi-modal references. Through the proposed dual prompt tuning inversion and sampling, the inference time can be reduced to nearly 1/3 of the original. The performance and temporal consistency can be also significantly stregthened. By using the proposed texture-aware video token merging, the temporal correlation between frames can be further utilized to improve the temporal consistency. We futher propose the referenced self-attention and referenced token merging to support image reference. Experimental results demonstrate the superiority of the proposed method in restoring and enhancing temporally consistent videos.
[CV-96] Mapping Woody Vegetation from Multi-Source Imagery and Prediction Fusion for Enhanced Data Efficiency and Accuracy
链接: https://arxiv.org/abs/2608.26471
作者: Kal Backman,Jared Wood,Adam Roff
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Published in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Abstract:Tree cover maps are a fundamental remote sensing product, used to derive ecological insights about the landscape and are essential to change detection, vegetation mapping and fire monitoring programs. However, comprehensive tree cover mapping requires reliable and high-quality imagery, free of cloud and weather defects to ensure accurate model outputs. Deep learning approaches can generate high quality maps with minimal human intervention but require large amounts of human annotated data to be successful. In this work we propose a framework consisting of methods that aim to improve the data efficiency and robustness of deep learning models using data fusion techniques to segment woody vegetation defined as vegetation over the height of 2m across the state of New South Wales, Australia. To improve robustness against varying image quality, we propose an image composition method that normalizes the imagery and removes defects, whilst also minimizing the reliance on individual image quality by proposing a prediction fusion method. The two methods resulted in an error reduction of 38.2% and 53.6% respectively compared to single-source imagery. To address deep learning approaches’ limitation of requiring large amounts of data, we apply label transfer to multiple sources of imagery as a form of data augmentation to improve data efficiency. Learning from multiple image sources was shown to be the biggest improvement in performance, resulting in an error reduction between 28.1% to 76.2% across the different validation experiments, whilst reducing the standard deviation of performance across image dates by a factor of 13.
[CV-97] VIPER: An Expert-Curated Benchmark for Vision-Language Models in Veterinary Pathology
链接: https://arxiv.org/abs/2608.26382
作者: Luca L. Weishaupt,Simone de Brot,Javier Asin,Llorenç Grau-Roma,Nic G. Reitsam,Andrew H. Song,Dongmin Bang,Stefan T. Kaluziak,Long Phi Le,Jakob Nikolas Kather,Faisal Mahmood,Guillaume Jaume
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Pathology vision-language models are advancing rapidly, yet existing benchmarks remain focused on human tissue, particularly oncology, leaving non-human pathology largely unaddressed. This gap is especially important in toxicologic pathology, where microscopic tissue examination of laboratory animals is a core component of preclinical drug safety assessment. To address it, we introduce VIPER, the first expert-curated benchmark for vision-language model evaluation in toxicologic pathology. VIPER contains 1,251 questions associated with 419 HE-stained rat histology images across seven organ systems, covering multiple-choice, KPrim, and free-text formats. All questions were curated and validated by board-certified veterinary pathologists. In total, we benchmarked 16 models, including two newly introduced veterinary-pathology models, seven human pathology-specialized models, and seven general-purpose frontier models. The results identify a substantial domain gap between veterinary and human pathology, expose the risk of over-diagnosis of normal tissue in frontier models, and show that domain-specific training remains critical for visually grounded predictions. VIPER data and evaluation code are available at this https URL.
[CV-98] A Unified Framework for the Mechanics of Information in Convolutional Neural Network Image Space
链接: https://arxiv.org/abs/2608.26363
作者: Aryan Shukla,Matthew Toews
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:This paper introduces a unified mathematical framework for modeling information propagation through convolutional neural networks (CNNs), with the aim of connecting descriptions of physical space and information space. A correspondence is presented linking discrete filter symmetry and the relativistic energy–momentum relation under the widely used nonlinear rectified convolution operation. Specifically, symmetric filter components (e.g. the sum \Sigma = [1,1] ) operate analogously to rest energy mc^2 in preserving the image centre of mass (e.g. isotropic diffusion), whereas antisymmetric components (e.g. the gradient \nabla = [-1,1] ) operate analogously to the momentum term pc in generally inducing a displacement (e.g. vibration or translation). For typical small discrete filters, this displacement is determined by the ratio of antisymmetric to total filter energy, analogously to how the displacement of a relativistic particle relates to a Lorentz transform with beta parameter \beta = \fracvc=\fracpcE equal to the ratio of momentum pc to total energy E . Repeated filtering leads to the Gaussian scale-space and emergent scale-invariant features. These constructions share a Laplacian-driven structure with the classical heat (diffusion) equation and, via standard mathematical correspondences, with the Schrödinger equation and aspects of the Friedmann equations, together with emergent Morse topological structure. Demonstrations in 3D images reveal blob-like, scale-invariant Morse critical points in images spanning a wide range of physical scales, including organic sugar molecules and inorganic silicon crystals, human and primate brains in magnetic resonance images (MRI), galaxies and the cosmic microwave background (CMB). Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2608.26363 [cs.CV] (or arXiv:2608.26363v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2608.26363 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-99] Finding the Right Evidence: Factor-Guided Coarse-to-Fine Reasoning for Long Videos
链接: https://arxiv.org/abs/2608.26355
作者: Baixuan Xu,Yinyui Xu,Tianshi Zheng,Zhaowei Wang,Weiqi Wang,Haochen Shi,Jiayu Liu,Qing Zong,Xiyu Ren,Xinyu Geng,Zhitao He,Yangqiu Song
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:While LVLMs rapidly improve, long-video question answering still remains challenging: relevant evidence is sparse, and question-relevant context often fails to provide cues that discriminate the correct answer from plausible alternatives. Diagnostic analysis on a manually annotated subset of MMR-V shows that prior agentic systems substantially improve cue retrieval over direct VLM inference yet fail to achieve a corresponding gain in answer accuracy, indicating that the bottleneck lies in option-discriminative evidence rather than topical relevance alone. We propose PACE (Progressive Acquisition of Critical Evidence), a factor-guided framework for long-video evidence acquisition. PACE proceeds in two stages: it first indexes clip-level descriptions guided by question-derived factors without observing the candidate answers; it then uses the candidate answers to derive contrastive cues and queries the index for verification. On MMR-V with the open-source Qwen3-VL backbone, PACE achieves 42.6% accuracy, outperforming direct inference and prior agentic baselines including Deep Video Discovery (DVD). On the same diagnostic subset, PACE recovers 66.9% of the annotated cues, providing empirical evidence that its gains are associated with improved evidence recovery rather than stronger answer-side priors alone. Consistent gains over DVD on LVBench, Video-MME, EgoSchema, and LongVideoBench suggest that option-aware evidence acquisition transfers beyond MMR-V. Code is available at this https URL.
[CV-100] StreamAV-Bench: A Comprehensive Benchmark for Streaming Audio-Video Generation
链接: https://arxiv.org/abs/2608.26336
作者: Kaiqi Liu,Haoxuan Zeng,Jingqi Liu,Jiacong Fang,Ziqi Cai,Yunyao Mao,Henglin Liu,Yu Sheng,Shuchen Weng,Boxin Shi
类目: ound (cs.SD); Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注:
Abstract:Recent advancements in generative models are pushing video generation toward unbounded streaming audio-video generation for real-time interactive worlds. However, existing benchmarks primarily evaluate completed sequences and struggle to capture streaming properties. To bridge this gap, we introduce StreamAV-Bench, the first comprehensive benchmark tailored for streaming audio-video generation. StreamAV-Bench establishes a unified evaluation framework, including the progressive track for instruction adherence and long-horizon stability, and the interactive track for interactive response and state retention and reuse. With expert-verified evaluation cases across 32 fine-grained dimensions, we conduct an extensive evaluation of 13 representative systems. Our analysis reveals that current models suffer from temporal drift in progressive generation and responsiveness bottlenecks during interactive control. Based on a comprehensive failure analysis, we share insights to advance the development of native joint audio-video streaming models.
[CV-101] Modality Maturity Index: A benchmark for assessing multimodal capabilities of omni models
链接: https://arxiv.org/abs/2608.26317
作者: Rohit Patel,Dieuwke Hupkes,Sloan Strader
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Multimedia (cs.MM)
备注: 26 pages, 6 figures. Code and dataset available
Abstract:Frontier language models are increasingly marketed as omni systems that can perceive and respond across modalities. Existing evaluation frameworks, however, focus almost exclusively on bimodal understanding, typically text plus one other modality. We propose the Modality Maturity Index (MMI), a benchmark designed to evaluate the multimodal capabilities of large language models across five modalities (text, image, audio, video and document) and combinations of up to three modalities in both inputs and outputs. MMI consists of 893 questions, each carefully crafted to require the model to demonstrate its understanding of multiple input modalities and to generate responses that incorporate various output formats. The questions are designed to be self-contained, with clear expectations for the correct modality or mix of modalities required for an accurate response. Every MMI prompt carries human-authored rubric criteria for each output modality expected in the response; a model’s MMI Value expresses the average of the per-modality scores for each prompt. Because low scores can reflect either failure to generate a modality (lack of presence) or failure to generate correct content, we introduce also a supplementary Modality Presence Score (MPS), a per-prompt F1 over the expected output modalities. Applying MMI to five frontier multimodal models, we find that the MPS ranges from only 15.6 (Claude Opus 4.6) to 34.9 (GPT-5.4). Given the low availability of returned modalities to even grade, we report MPS as our main result pending model improvements. To assess the viability of judging output correctness with LLM judges and rubrics, we run a separate experiment with custom generation tools. On the assets that generates, we find that an LLM judge applying the rubrics agrees with rubric-blind human annotators (who score the outputs directly and never see the criteria) on 70.8% of judgments.
[CV-102] Procedura: Agent ic 3D Modeling with Procedural Control
链接: https://arxiv.org/abs/2608.26238
作者: Youtian Lin,Yikang Yang,Zhanpeng Hu,Mengqi Zhou,Feihu Zhang,Xun Cao,Jiaheng Liu,Yao Yao
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: Project page: this https URL
Abstract:Native 3D generators now recover impressive mesh geometry from a single image. However, a dense mesh stays soft where a machined object should be sharp, it carries no part decomposition, and it exposes no parameter a user could edit. To address this, we explore the paradigm of 3D shape as code, leveraging and scaling the coding ability of an LLM for 3D modeling. We introduce Procedura, a novel 3D modeling agent framework that writes an object as a procedural assembly, a parametric program whose named parts are joined by typed, machine-checkable mates. From a text prompt, the agent plans the object as an assembly graph and writes the program part by part, solving each placement from the mated frames rather than guessing it, and admitting a part only once compile, mate, and connectivity checks pass. A decoupled vision critic then refines the assembly one diagnosed fix at a time. Moreover, the same graph carries per-part materials and a simulator-validated articulation. We evaluate on P3D-Bench under its assembly judge, and with the same judge on MechBench-36, our hard-surface benchmark. On both, Procedura outperforms state-of-the-art native 3D generators and every prior 3D-code agent on judged quality, produces the sharpest edges of any method we evaluate, and is the only one whose output is an editable, part-structured program.
[CV-103] Surgical Video Generation From Diffusion to World Models: A Survey
链接: https://arxiv.org/abs/2608.26214
作者: Fuxiang Huang,Chenxu Zhang,Liang Han,Lei Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 4 pages, 1 figures, 3 tables. Accepted for oral presentation at the 2026 3rd International Conference on Intelligent Perception and Pattern Recognition (IPPR 2026)
Abstract:Surgical video data provides the primary training resource for models of intraoperative perception, surgical workflow understanding, and robotic decision-making. However, clinical data acquisition remains constrained by privacy, cost, and class imbalance. Surgical video generation has emerged as a transformative approach to addressing data scarcity and as a foundation for surgical simulation, training, and robotic policy learning. The field has developed rapidly without a clear conceptual framework. This survey organizes the 2024-2026 literature into three categories: unconditional generation, conditional generation, and world modeling generation, revealing a fundamental shift in how the task is defined from synthesizing visually plausible frames to modeling the causal dynamics of surgical scenes. We examine the persistent gap between pixel-level fidelity and clinical plausibility, and identify generalization, physical realism, controllability, and interpretability as bottlenecks. We further summarize experimental results of representative methods on public datasets to provide a quantitative reference for the field. This survey provides a structured overview of the current state and open challenges, offering a reference for researchers working at the intersection of intelligent perception, multi-modal fusion, generative AI, and surgical data science.
[CV-104] Attention-Guided Reliability Scaling for Contrastive Decoding in Robust Audio-Visual Speech Recognition INTERSPEECH2026
链接: https://arxiv.org/abs/2608.26213
作者: YoungChae Kim,Da-Hee Yang,Joon-Hyuk Chang
类目: ound (cs.SD); Computer Vision and Pattern Recognition (cs.CV); Audio and Speech Processing (eess.AS)
备注: Accepted to Interspeech 2026
Abstract:Large language model (LLM)-based audio-visual speech recognition (AVSR) systems are robust under noise. Contrastive decoding (CD), originally introduced to stabilize LLM generation by contrasting a weaker model against a stronger one at inference time, adjusts predictions without additional training. In this work, we apply CD to AVSR by contrasting audio-only conditioning with full audio-visual conditioning within the same underlying model. However, using a fixed contrastive strength introduces a trade-off across noise levels: stronger intervention helps under severe noise but may over-correct reliable predictions in clean conditions. We propose reliability-aware scaling of CD for AVSR. Instead of using a fixed strength, we adaptively modulate the contrastive influence at each token based on reliability signals derived from attention dynamics and inter-model predictive divergence. Experiments on LRS3 show consistent improvements across clean and low-SNR conditions.
[CV-105] GameWAM: A World Action Model for Video Games
链接: https://arxiv.org/abs/2608.26200
作者: Yuncheng Guo,Zhanqiu Zhang,Yiwen Guo,Weijia Li
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 44 pages, 23 figures, 7 tables
Abstract:Modern video games combine first-person perception, rapid visual changes, persistent world state, and heterogeneous native controls. Existing game agents map visual and task context directly to actions but lack explicit world dynamics modeling, whereas interactive game world models predict visual futures from supplied actions but do not serve as task policies. World-Action Models (WAMs) unify these objectives, but remain largely unexplored under the dynamics and open-ended interaction of video games. We introduce GameWAM, to our knowledge the first WAM for native closed-loop gameplay and GUI control. GameWAM jointly generates future visual observations and executable keyboard-mouse trajectories through parallel visual and action generative processes with block-causal conditioning and flow matching. To support joint world-action learning, we construct synchronized gameplay and GUI trajectories. To handle heterogeneous native control, GameWAM predicts a gameplay/GUI mode at each action step and generates actions with mode-specific prediction distributions and continuous-action normalization. For long-horizon interaction, block-cycle control predicts beyond the committed horizon, executes only a short action prefix, and replans from new observations, while fine-grained within-cycle context and hierarchical cross-cycle history preserve temporal continuity. Experiments demonstrate competitive task success with fewer executed native actions than the compared agents. We further uncover Low-Frequency Action Source Imprinting (LASI), in which low-frequency components of the sampled action source systematically steer coarse generated camera motion under fixed conditioning, revealing a source-sensitivity failure mode in generative control. Project page is available at this https URL.
[CV-106] ClassVision: AI-Powered Classroom Attendance System
链接: https://arxiv.org/abs/2608.26173
作者: Ankit Kumar Aggarwal,Veerabhadra Rao Marellapudi,Ovadia Sutton,Youshan Zhang
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Students and working professionals have to go through the attendance process every day. Traditional methods of marking attendance using pen and paper or online platforms are human-intensive and time-consuming. To address the challenges in manual attendance processes, this research explores the use of face detection (FD) and face recognition (FR) technology to automate the attendance process, particularly in educational settings, and build a ClassVision course attendance system. We also propose an automated attendance system featuring a human-computer interaction (HCI) and user-friendly web interface that utilizes real-time image processing to identify and recognize students in classrooms and automatically record their attendance. We identified RetinaFace as the best face detection model, and when combined with Face Recognition for verification, it provided the most promising results with a cropped embedding of 50x50 pixels.
[CV-107] Data-efficient crack quantification in lithium-ion cathodes using foundation model transfer
链接: https://arxiv.org/abs/2608.27162
作者: Thorsten Tegetmeyer-Kleine,Thomas Schmitt,Phillip Aquino,Christiane Rahe,Dirk Uwe Sauer,Weihan Li
类目: Materials Science (cond-mat.mtrl-sci); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 38 pages, 15 figures, 2 tables. Supplementary Information included as Appendix B. Under review at Energy Storage Materials
Abstract:Battery lifetime is central to sustainable electrification, yet the particle cracking that drives lithium-ion cathode aging is hard to measure: quantitative microscopy of this degradation is bottlenecked by annotation, because each destructive electron-microscopy cross-section spans hundreds of megapixels and pixel-level expert labelling requires hours per image. We show that a frozen self-supervised vision-transformer encoder, combined with a lightweight trainable decoder and iterative model-assisted annotation, turns this sparse labelling budget into population-scale degradation measurements. Applied to three 120-megapixel NMC cathode cross-sections representing initial, cycled-aged and calendar-aged states, the framework distinguishes intragranular cracks from early- and late-stage intergranular cracks and yields per-particle distributions of crack width, tortuosity and area fraction. Late intergranular crack coverage reaches 4.6% in the cycled sample versus 0.5% in the initial and calendar-aged samples, forming more tortuous, higher-coverage networks, consistent with degradation from repeated electrochemical cycling rather than elevated-temperature storage alone. A single destructive image yields the population-level statistics needed for lifetime-extending design, aging assessment and second-life decisions.
[CV-108] Dose-PlanNet: Physics Based Radiotherapy Dose Prediction with Deep Learning
链接: https://arxiv.org/abs/2608.26901
作者: Ankit Bhattacharjee,Sougata Maity,Santam Chakraborty,Indranil Mallick
类目: Medical Physics (physics.med-ph); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: The paper consists of 22 pages, 4 figures, 6 tables. The end-to-end pipeline of Dose-PlanNet will soon be made available on the GitHub repository of CHAVI-India ( this https URL )
Abstract:Automating prostate radiotherapy treatment planning is dosimetrically complex, particularly for extreme hypofractionated regimens. In this study, we introduce Dose-PlanNet, a physics-guided 3D deep learning architecture designed to predict dose distributions. This model’s performance was evaluated on a cohort of patients treated in a prospective trial where two different dose fractionation regimens were employed. Dose-PlanNet achieved comparable target coverage ( D_95 ), though statistical analysis revealed a marginal reduction in target homogeneity ( p0.001 ) offset. However the model achieved statistically significant improvements in high-dose organ-at-risk sparing ( p0.001 ). When evaluated against strict Prospective Randomized protocol volumetric constraints, automated plans met prespecified clinical acceptance criteria in 11 out of 14 Moderate Hypofraction Arm plans and 9 out of 12 Stereotactic Body Radiation Therapy Arm plans. This pipeline demonstrates that physics-informed deep learning can accelerate radiotherapy workflows while safely maintaining the stringent dosimetric quality required for high-precision clinical deployment.
人工智能
[AI-0] RedEvoAgent : Automatic Red-Teaming Agent with Experience-Driven Skill Evolution
链接: https://arxiv.org/abs/2608.27439
作者: Junjie Zhang,Hui Liu,Kecheng Chen,Xianbo Mo,Changsheng Chen,Haoliang Li
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:LLM-based agents are increasingly deployed in product-level execution harnesses, where jailbreaks can trigger harmful tool use and persistent state changes, creating greater risks than unsafe text generation alone. Existing automatic red-teaming methods often rely on fixed attacks, while recent agentic attackers coordinate multiple jailbreak tools and show stronger potential through trajectory-based retrieval. However, such retrieval can reuse misleading experiences due to retrieval bias and unclear tool credit, and full trajectories add context overhead while reducing interpretability. We propose RedEvoAgent, a black-box red-teaming agent that distills cross-case attack trajectories into a concise, human-readable attack skill. The attack skill adaptively evolves through tool-effectiveness profiling and Deciding-Tool Attribution for skill updates, and a validation ratchet that retains only updates improving validation performance. Experiments on multiple benchmarks, target models, and target execution harnesses show that RedEvoAgent outperforms fixed and agentic baselines, improves tool efficiency, and transfers across attacker models and target execution harnesses.
[AI-1] Mechanistic Reaction Prediction via Discrete Flow Matching on Graph-Structured Electron Occupation
链接: https://arxiv.org/abs/2608.27429
作者: Nguyen Xuan-Vu,Octavian Susanu,Daniel Armstrong,Philippe Schwaller
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Chemical reactions are fundamentally transformations in electron space, yet most machine learning approaches model them either through \textitde novo generation of product molecules or through heuristic graph edits that operate directly on molecular topology. We introduce MAELLE (\textbfMech\textbfAnistic \textbfEdit f\textbfLow-matching on e\textbfLectron r\textbfEarrangements), which instead models reactions as discrete flow matching over electron occupation vectors. Concretely, we formulate the reactant-to-product mapping as a Continuous-time Markov Chain (CTMC) over the graph-structured integer-valued electron occupation space defined on all bonding, non-bonding, and hydrogen sites. To construct the intermediate edit trajectories, we generalize the discrete flow matching mixture path to discrete electron rearrangements using Optimal Transport, yielding a sequence of mechanistically interpretable edit moves without requiring elementary step annotations. MAELLE achieves competitive performance on the USPTO-480K benchmark compared with leading reaction prediction models. Beyond in-distribution accuracy, we evaluate robustness across two out-of-distribution settings - structural complexity and reaction type - and find that MAELLE maintains strong performance where existing methods degrade. Finally, because the learned flow operates over the full electron redistribution, MAELLE naturally recovers mechanistic trajectories that align with known chemistry and can predict side products of a reaction. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2608.27429 [cs.AI] (or arXiv:2608.27429v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.27429 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Xuan Vu Nguyen [view email] [v1] Thu, 27 Aug 2026 17:50:44 UTC (1,078 KB) Full-text links: Access Paper: View a PDF of the paper titled Mechanistic Reaction Prediction via Discrete Flow Matching on Graph-Structured Electron Occupation, by Nguyen Xuan-Vu and 3 other authorsView PDFTeX Source view license Current browse context: cs.AI prev | next new | recent | 2026-08 Change to browse by: cs References Citations NASA ADSGoogle Scholar Semantic Scholar export BibTeX citation Loading… BibTeX formatted citation loading… Data provided by: Bookmark checked="checked"class=“labs-tab-input”> Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv’s community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) mathjaxToggle(); We gratefully acknowledge support from our major funders, member institutions, , and all contributors. About Help Contact Subscribe Copyright Privacy Accessibility Operational Status (opens in new tab) Major funding support from
[AI-2] Persona-Execution Separation: An Architecture Pattern for Evolving LLM Agents under Execution Audit
链接: https://arxiv.org/abs/2608.27427
作者: Yisen Xi
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 36 pages
Abstract:Large language model (LLM) agents in governed organizations must let the persona (instructions, tone, self-presentation) evolve freely, while keeping execution (stateful, audited work) traceable. A single trust domain does not satisfy both cheaply. We present Persona-Execution Separation (PES): persona and execution reside in different trust domains, connected by a governed contract bridge. The persona is singly-homed and may drift; execution is faceless and audited. Status summaries may return; data bodies remain in the restrictive domain except a graded data-loss-prevention (DLP) exception; identity stays continuous. An approval matrix, DLP, and audit enforce the crossing. PES follows from three goals—free drift, execution traceability, and decoupling. Under LLM representational indistinguishability, any single-domain mechanism that meets all three must re-introduce typed change objects, an external gate, and a stable audit anchor: PES rebuilt at higher coupling cost. A development/pilot case in a regulated digital-employee platform records five decisions over one month, each with a rejected alternative. A mechanism check on the shipped implementation found no execution-side re-validation under persona perturbation (five model configurations) and no persona fingerprint on hard-asserted fields. A probe of a recovered pre-separation build found the governed execution path decoupled from the persona by omission, not by construction; a later wiring change could reverse that isolation, which PES makes an audited architectural rule. The pattern applies when multi-user deployment, execution audit, and expected persona churn hold jointly.
[AI-3] Beyond F1: Evaluating Coverag e and Failure Recovery in AI Model Security Scanners
链接: https://arxiv.org/abs/2608.27424
作者: Qianlong Lan,Vinothini Pandurangan,Anuj Kaul,Indranil Sanyal
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Static scanners are increasingly used to identify executable or otherwise unsafe content in machine- learning artifacts, yet conventional evaluation metrics characterize only cases where a scanner yields a usable security judgment. We evaluate ModelScan, ModelAudit, and Fickling using a controlled, artifact-backed benchmark on a synthetic corpus of 170 Pickle and PyTorch focused artifacts across 145 specimen families, 135 of which have binary security ground truth and 10 of which are intentionally malformed without labels. We explicitly distinguish non-N/A coverage, analysis completion, definitive security decisions, non-security findings, and unsupported outcomes. On labeled families, ModelAudit produced definitive security decisions for all 135 families (100%), Fickling for 110 (81.5%), and ModelScan for 67 (49.6%). Conditional on making a definitive judgment, ModelScan achieved 100% precision, recall, and F1. Fickling identified no unique true- positive families beyond those found by the combination of ModelAudit and ModelScan. Furthermore, for the 48 malicious families where ModelScan failed to complete its analysis, both ModelAudit and Fickling generated detections consistent with ground truth. These findings underscore the need to separate judgment accuracy from judgment availability, as well as incremental detection coverage from tool-level redundancy.
[AI-4] Learning a Continuous Sepsis Severity Score Without Hour-by-Hour Supervision: A Two-Site Retrospective Study
链接: https://arxiv.org/abs/2608.27421
作者: Kevin Zhu,Ryan Zhang,Baraa Abed,Tilendra Choudhary,Malvern Madondo,Mehak Arora,Yixuan Yang,Alasdair Gent,Aditya Nagori,Omer T. Inan,Krista L. Haines,Patrick Georgoff,Suresh M. Agarwal,Vijay Krishnamoorthy,Tetsu Ohnuma,Mihai V. Podgoreanu,Michael R. Pinsky,Gilles Clermont,Craig M. Coopersmith,Craig S. Jabaley,Rishikesan Kamaleswaran
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Currently used sepsis severity indices rely on fixed variables and weights established decades ago, which are coarsely discretized and calibrated to a cohort that no longer reflects contemporary critical care. No alternative learned directly from patient trajectories is in routine use. We conducted a retrospective two-cohort study on a total of 29,116 and 7,691 adult patients meeting Sepsis-3 criteria from two hospital systems in Massachusetts and Georgie, respectively. We developed a sepsis index using 43 routinely charted variables over a 72-hour treatment window. Unlike previous studies, we use mortality as a treatment-level ranking signal rather than a per-state target, allowing credit to be redistributed non-uniformly across timesteps. Evaluation was done on a permanent 20% test holdout, using clinical vignettes and Spearman correlation. Uncertainty intervals were obtained by bootstrap resampling of whole patients. Under this ranking scheme, non-survivors scored 1.19-1.64 points higher than survivors on a 0-10 scale within all strata of baseline SOFA-2, with similar results stratifying within lactate, mean arterial pressure (MAP), and creatinine. Within-patient change in the index correlated with change in lactate (Spearman rho = 0.39; n = 1,854). Similar, weaker correlations were found for MAP and creatinine. On a cohort level, cross-institutional agreement measured by Spearman correlation between models trained on different sites, were 70-77% of same-site correlation. External within-patient correlations were 0.54 and 0.59 against ceilings of 0.92 and 0.90. Our index also correlated with established indices, while null controls stayed near zero. Our index demonstrated hourly prognostic information that meaningfully separates patient outcomes and is consistent with clinical expectation, indicating potential as a decision support tool complementing clinical judgement. Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2608.27421 [cs.AI] (or arXiv:2608.27421v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.27421 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-5] Sophistication in GenAI Use: Field Evidence from a Large Firm
链接: https://arxiv.org/abs/2608.27364
作者: Nicholas J. Hallman,Zachary T. Kowaleski,Anu Puvvada,Jaime J. Schmidt
类目: Artificial Intelligence (cs.AI); General Economics (econ.GN)
备注: 59 pages, 4 figures, 8 tables. Includes Appendix A (variable definitions), Appendix B (supplementary tables), and Online Appendix C (meta prompts)
Abstract:We study how sophistication in generative AI (genAI) use varies among the back-office workforce of a large firm. Using proprietary data, we observe 713,564 employee prompts and their corresponding large language model responses from nearly 4,000 back-office employees across 15 functional areas over eight months in 2025. We document three main findings. First, senior employees exhibit more sophisticated genAI use, consistent with domain expertise complementing genAI capabilities. Second, sophistication varies considerably across functions and is highest in Strategy, Digital Innovation, and Project Management, three groups that share a focus on firmwide strategic initiatives and organizational change. Third, we observe neither improvements in sophistication over time nor lasting improvements following formal AI training, suggesting that sophisticated use can be difficult to change. Together, our study provides measures of and insights into sophisticated genAI use that managers can use to improve outcomes and that researchers can use in future research.
[AI-6] Not All Eval-Awareness Is Equal: Capabilities Framing Predicts Compliance ICML
链接: https://arxiv.org/abs/2608.27340
作者: Allison Zhuang,Santiago Aranguri
类目: Artificial Intelligence (cs.AI)
备注: 5 pages (14 appendices), 5 figures, presented at 2026 ICML Mechanistic Interpretability Workshop
Abstract:Steering interventions targeting eval-awareness, a model’s recognition that it is being tested, are increasingly used in safety evaluation pipelines, where evaluation-awareness is treated as a single quantity to be suppressed. We show that verbalized eval-awareness in chain-of-thought can be identified as capabilities-flavored (“the user is testing my ability to follow instructions”), safety-flavored (“the user is testing my boundaries”), both, or neither: framings that predict compliance very differently. On Qwen3-32B over the FORTRESS dataset, capabilities-framing predicts compliance with a +24 to +46 percentage-point gap over safety-framing across all tested steering conditions. A CoT-prefill intervention on eval-awareness-negative rollouts suggests the link is causal, with 10 of 11 prefills shifting compliance in the predicted direction. Then, eval-awareness is not behaviorally uniform: aggregate suppression rates can move while the safety-relevant component does not, and the same “X% suppression of eval-awareness” can correspond to qualitatively different behavioral outcomes.
[AI-7] Verify Smarter Evolve Further: Efficient Harness Evolution through Behavior-Aware Verification
链接: https://arxiv.org/abs/2608.27311
作者: Jinghan Xu,Yikai Zhang,Aili Chen,Weiyuan Li,Jiaqing Liang,Deqing Yang
类目: Artificial Intelligence (cs.AI)
备注: 17 pages, 6 figures
Abstract:Agent harnesses shape how language-model agents use instructions, tools, and runtime components, but adapting these harnesses requires costly verification. Existing propose-and-verify methods typically score every candidate on a fixed task set, wasting rollouts on unrelated behaviors and allowing aggregate scores to obscure specific regressions. We introduce HarnessLens, a budget-aware framework for automated harness evolution. HarnessLens jointly explores the task space and user-configurable components, derives candidate modifications from execution trajectories, and selectively verifies each candidate on behavior-relevant tasks using an attributable-evidence gate. Across three agent harnesses and four benchmarks, HarnessLens improves average held-out performance by 7.6-13.6% while consuming substantially less evaluation budget than competing baselines. These results demonstrate that behavior-aware verification with explicit attribution enables more reliable and sample-efficient harness evolution under constrained interaction budgets. Our code is available at this https URL.
[AI-8] LLM s Can Design Near-Optimal OR Algorithms
链接: https://arxiv.org/abs/2608.27296
作者: Jackie Baek
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:We ask whether large language models (LLMs) can design effective algorithms for well-specified operations research (OR) problems. We study inventory control, queueing network control, and assortment optimization. We evaluate two levels of LLM use: at level 1, the model receives one problem instance and returns a solution for that instance; at level 2, it receives only the problem class description and broad parameter ranges, and returns an algorithm that maps instance parameters to solutions. Human input is minimal: we give one untuned prompt that describes the problem, and the model has access to a Python sandbox tool with a fixed compute budget. The strongest model we test, gpt-5.6-sol, matches or outperforms the best existing method on almost all evaluated instances. This holds even at level 2, where the returned algorithm is fixed before seeing the evaluation instances. Performance also improves sharply across models released less than eight months apart, suggesting that this capability is moving quickly. Thus, for the well-specified operations problems we study, a single untuned LLM query can already produce algorithms competitive with specialized methods. These results suggest that frontier LLMs can be a serious empirical baseline for algorithm design in well-specified OR problems. Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2608.27296 [cs.AI] (or arXiv:2608.27296v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.27296 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-9] Compositional Online Learning for Semantic Data Processing Systems
链接: https://arxiv.org/abs/2608.27244
作者: Paweł Liskowski,Fuheng Zhao,Benjamin Han,Anupam Datta,Dimitris Tsirogiannis
类目: Databases (cs.DB); Artificial Intelligence (cs.AI)
备注:
Abstract:An LLM call in a semantic data processing system is expensive enough to dominate query cost, yet slow enough to hide a CPU-side learner’s update behind its round-trip. In production, LLM compute accounts for 80-90% of query cost, and each call costs 10^5-10^7\times a relational predicate. The latency window inverts a design constraint of classical adaptive query processing, where online learners had to stay lightweight to avoid dominating the predicates they optimize. At LLM latency, per-call gradient steps and per-batch threshold solves fit inside the round-trip. We develop compositional online learning at the LLM call boundary: a framework for combining online-learning components in semantic data processing systems. Each component makes execution-time decisions and refines its learned artifacts online. The design space spans two axes, decision granularity and learner update cadence, and the components share a single learning pattern that hides each trainer step inside the next LLM round-trip. A production case study in Cortex AISQL composes three components: a memoization layer, an online per-call filter-ordering learner, and an online per-batch cascade-routing learner. A conditional cost decomposition assigns each learning component to a distinct factor of per-row LLM cost. Under independence, the two learning components compose multiplicatively to an 11.4\times upper bound on a representative conjunction-filter workload. Self-selection at the cascade boundary, sample-budget shrinkage, and selectivity-estimation drift reduce it to a realistic figure near 8\times .
[AI-10] STEP: State-Aware Task Estimation and Planning with Multi-Modal LLM s for Human-Robot Collaboration
链接: https://arxiv.org/abs/2608.27225
作者: Maitrey Gramopadhye,Prakash Baskaran,Xiao Liu,Songpo Li,Soshi Iba
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: Published in IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), 2026, 8 pages, 4 figuers, 5 tables
Abstract:Effective human-robot collaboration in industrial settings requires robots to understand human intentions and assist with task planning, reducing workload. Recent works have explored the use of Multi-modal Large Language Models (MM-LLMs) for task planning in such data-scarce scenarios, leveraging in-context learning to interpret user actions and generate long-horizon action plans in natural language. However, MM-LLMs inherently lack an understanding of system states and do not track state transitions, often leading to hallucinated actions that deviate from the intended goal. Additionally, generating action plans in natural language tends to limit the generated plans to a high level, introducing ambiguity in action execution. To address these limitations, we propose the State-aware Task Estimator and Planner (STEP), which prompts a MM-LLM to explicitly estimate the state of the system and predict the state transitions resulting from executed actions. By forecasting future states alongside actions, STEP ensures task-convergent planning while also providing additional assistance parameters necessary for executing the predicted actions. We evaluate STEP in a simulated environment using a robot assembly task. Our approach outperforms the state-of-the-art by 32.8% in action executability and 14.8% in final-state error.
[AI-11] LLM s in Digital EDA: A perspective on shifting roles from Generation to Orchestration
链接: https://arxiv.org/abs/2608.27184
作者: Matthew Youngman,Cristian Sestito,Themis Prodromakis
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 16 pages, 4 figures, 2 tables (in Supplementary Information); interactive version of Fig. 2 at this https URL
Abstract:Electronic design automation (EDA) has advanced engineering productivity through successive generations of tooling that progressively automate synthesis, optimisation, and verification. Large language models (LLMs) extend this trajectory by enabling direct translation from design intent to hardware implementations. In most of the EDA literature, LLM-based solutions are typically assisting siloed design stages or tasks, however this obscured the drivers by which capability emerges and systems scale. In this Perspective, we instead define three hierarchical roles that reveal how capability accumulates: a Generator that produces design artifacts in a single pass, an Agent that refines outputs through iterative tool feedback, and an Orchestrator that coordinates decisions across EDA-stages. Across published systems, this reveals a syntax trap in which models are trained to produce plausible code rather than physically correct hardware, compounded by fragmented tools and loss of design context that obscure how decisions affect later stages. Comparisons across the three roles show that current approaches struggle to scale to industrial designs, motivating a shift towards a standardised, physics-aware orchestrator that connects tools and agents across the EDA flow for more reliable and accessible hardware design.
[AI-12] BPMN4CAI: A BPMN Extension for Modeling Dynamic Conversational AI
链接: https://arxiv.org/abs/2608.27149
作者: Björn-Lennart Eger,Daniel Rose,Barbara Dinter
类目: Artificial Intelligence (cs.AI)
备注: Presented at the 20th International Conference on Wirtschaftsinformatik (WI 2025), September 2025, Münster, Germany
Abstract:Conversational AI systems, such as chatbots and virtual assistants, are becoming increasingly important to digital business processes. However, the established Business Process Model and Notation (BPMN) standard faces challenges when representing dynamic, context-sensitive interactions. This paper addresses this methodological and practical research gap by developing a standard-compliant BPMN extension (BPMN4CAI). Using Design Science Research methodology, this paper develops an approach that systematically extends existing BPMN elements and incorporates specialized components. The applicability and relevance of the BPMN4CAI framework are demonstrated and evaluated through a case study. The results show that the BPMN4CAI extension facilitates adaptive decision-making processes, robust context management, and transparent interactions for Conversational AI within business processes.
[AI-13] homson: Continual Learning of Frontier Models for SovereignAI
链接: https://arxiv.org/abs/2608.27147
作者: Shengzhuang Chen,Jerrod Parker,Yejin Bang,Andrew M. Bean,Nabeel Seedat,Stefan Winzeck,Daniil Glazko,Jannik Zgraggen,Fangyi Yu,Scott Arnott,Dietrich Trautmann,Luca Ciuffreda,Guglielmo Bonifazi,Davide Romano,Bradley Bell,Kirsty Fielding,Daniele Giofrè,Tom Zielund,Ipshita Chatterjee,Sneha Murthy Ghantasala,Manpreet Nanreh,John Scoville,Maciej Sakowicz,Wassim Seifeddine,Lukas Thede,Jonathan Richard Schwarz
类目: Artificial Intelligence (cs.AI)
备注: Open-weight model: this https URL
Abstract:The development of frontier models is commonly perceived to be the exclusive remit of a small number of heavily funded players, creating an information, economic and power asymmetry between developers and the diverse user base of modern AI. Recent public discourse acknowledges this concern, calling for SovereignAI (an organisation’s capability to independently build, deploy and govern AI use), but offers little concrete advice on how this can be achieved in the short term under a diversity of funding settings. We argue that frontier performance is achievable by a wide range of institutions through Continual Learning on readily available open-weight models. Unlike limited approaches such as small-scale fine-tuning, prompt engineering, or tool-augmentation of a frozen model, our approach exploits a modern mid- post-training stack while introducing safeguards that preserve both plasticity and stability at each stage, making the minimal number of high-impact interventions on the parameters. This yields gains comparable to those typically seen across multiple successive model generations, at compute and personnel budgets substantially lower than commonly thought, making ownership of large parts of the SovereignAI stack (model, tool infrastructure, values data privacy) viable for far more actors. We demonstrate this with Thomson, a general-purpose frontier model trained with an enhanced focus on high-stakes professional work. Thomson performs competitively with recent frontier models across agentic tasks, safety, legal, tax multilingualism, and large-scale Deep Research. Evaluations show a distinctive \pi -shaped pattern: distinct improvements across a wide range of capabilities, including those not explicitly targeted, while almost completely eliminating the forgetting problem common to narrow domain adaptation.
[AI-14] When Tool Outputs Become Commands: Separating Action Induction from Runtime Authorization in Tool-Augmented LLM Agents
链接: https://arxiv.org/abs/2608.27146
作者: Xiaokun Guo,Zhen Xu,Dongdong Huo,Yanqiu Zhang,Wei Wang,Qinfu Yang,Dongjin Yu,Yu Wang
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Tool-augmented LLM agents must rely on untrusted runtime Observations to complete open-ended tasks; however, when tool outputs no longer merely provide data but begin to specify concrete actions, they effectively become ``commands’’ that can drive real-world side effects beyond user intent. We argue that this risk arises from conflating action induction with execution authorization. To address this distinction, we propose SARA, which treats action induction and execution authorization as distinct runtime roles and separates action provenance from execution authority. On the Observation side, a context-isolated Action Probe exposes action-inducing semantics and persistently records action-origin provenance across steps as a review signal; on the execution side, actual tool calls are authorized only against the user objective and audited evidence from authorized successful executions, while satisfying goal, execution-chain, and argument-level support. To preserve this separation across multi-step execution, SARA applies No-History-Promotion to prevent historical recurrence from laundering action origins into execution authority. Across AgentDojo and AgentDyn, SARA limits ASR to no more than (0.63%) across four primary evaluation settings while maintaining competitive task utility, and consistently reduces ASR across additional Agent backbones.
[AI-15] Feature Transformation Enhanced Jacobi Polynomial Graph Filtering for Graph Anomaly Detection
链接: https://arxiv.org/abs/2608.27144
作者: Xiang Wang,Zhijun Cheng,Zhenyu Meng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:In recent years, graph anomaly detection (GAD) based on frequency-domain filtering have achieved promising results. However, existing approaches still face three major challenges: First, they use static basic function to constructed graph filter which cannot effectively adapt to the frequency-domain distribution of graph data. Second, they fail to adequately consider the importance information of each attribute in the node feature vector, leading to the loss of fine-grained information. Third, they insufficiently utilize node labels for GAD. To address these issues, this paper proposes a novel graph anomaly detection method called JPGFN (Feature Transformation Enhanced Jacobi Polynomial Graph Filtering Network). First, a Feature Separation Transformation Network (FSTNN) is developed to better learn fine-grained node features by feature separation and applying nonlinear transformations to node features across different dimensions. Second, an adaptive Jacobi polynomial graph filtering module is constructed based on Jacobi polynomials to adaptively capture complex frequency-domain features of graph signals. Finally, a node label constraint module is developed to facilitate the use of node labels and enhance the performance of GAD. Experimental results on multiple real-world datasets demonstrate that the proposed method significantly outperforms mainstream approaches.
[AI-16] GRAIN: Bridging Name and Narrative Shifts in Real-World Graph Reasoning through Invariance-Rewarded Agent ic RL
链接: https://arxiv.org/abs/2608.27142
作者: Zike Yuan,Han Zhang,Jianzhi Yan,Le Liu,Cai Ke,Huozhi Zhou,Jian Xie,Jiran Yin,Yukun Cao,Yue Yu,Hui Wang,Ming Liu,Bing Qin
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Despite their potential in standardized graph tasks, Large Language Models (LLMs) remain brittle to real-world shifts in node identifiers and task formulation. While deterministic graph tools are invariant to such shifts, extracting topological structures from noisy text is highly fragile for LLMs, which often overfit to surface patterns. Moreover, mitigating these parsing failures via multi-agent systems incurs prohibitive latency. To address this, we propose GRAIN, a single-agent framework optimized via reinforcement learning. GRAIN models reasoning as a semantic parsing and tool-execution pipeline, guided by a Structure Invariance Reward. By validating extracted intermediate graphs against ground-truth topologies, this reward forces the LLM to learn robust text-to-structure mappings rather than memorizing linguistic artifacts. We also introduce GRIT, a benchmark evaluating sensitivity to such linguistic shifts. GRAIN outperforms multi-agent baselines by 16.45% in accuracy with approximately 24% lower latency. Furthermore, it demonstrates superior structural generalization, halving the out-of-distribution (OOD) gap of SFT models (from 15.77% to 7.80%) and maintaining robustness on large-scale graphs beyond the training distribution.
[AI-17] Safety Does Not Compose: Non-Decaying Loop State for Autonomous LLM Agents
链接: https://arxiv.org/abs/2608.27141
作者: Chenhao Wu,Haoxuan Jia,Yang Liu,Yingguang Yang,Yuhan Lin,Chongyang Zhang,Hao Zheng,Yulin Huang,Jianshen Zhang,Yongzhi Qi,Shang Luo,Kefu Xu,Jifeng Zhu,Bin Chong
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model agents are increasingly deployed as autonomous loops. Starting from one human goal, such a system repeatedly discovers work, plans, executes tool calls, verifies outcomes and persists state across many unattended iterations. The agent safeguards in wide use, however, are defined over a single trajectory, and their safety state is re-initialized when the next trajectory begins. We show that this is a failure of composition rather than an implementation detail. Our central result is a separation: against an attack whose evidence is fragmented across several iterations, every trajectory-scoped monitor has a true-positive rate equal to its false-positive rate, however expressive it is, because the evidence it would need never appears in the window it sees, whereas a monitor retaining cross-iteration state separates the two perfectly. We further show that the obvious repair of carrying a geometrically decaying risk score is insufficient, because the cooling-off period a patient adversary must wait is a constant that does not grow with the horizon N . We then present LoopHarness, which restores a persistent, non-decaying safety state at the loop level. Under mediated commits and an arbiter detection floor \delta_M , it bounds the expected number of unauthorized irreversible actions by B+m-1+m/\delta_M , a constant in N , of which the B+m-1 term is decided by a model-free rule and therefore survives a fully colluding verifier. We give a complete evaluation protocol on native Agent-SafetyBench tasks with paired clean and attacked episodes, an outer-state attack suite whose decisive evidence exists only across iterations, per-module ablations, and an adaptive white-box red team.
[AI-18] ransMeme: A Multi-Agent Framework for Cross-Cultural Meme Transcreation ACM-MM2026
链接: https://arxiv.org/abs/2608.27127
作者: Jingyi Zheng,Yule Liu,Zifan Peng,Tianyi Hu,Yuemeng Zhao,Xinhu Zheng,Xinlei He
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 4 figures. Accepted at the 34th ACM International Conference on Multimedia (ACM MM 2026)
Abstract:Internet memes are a pervasive form of multimodal online communication; however, such communication often involves users from diverse linguistic and cultural backgrounds. Therefore, adapting memes across cultures and languages is a central challenge for enabling mutual understanding in online communication. Unlike ordinary translation or standalone text rewriting, cross-cultural meme transcreation must jointly preserve communicative intent, adapt culture-dependent meaning for the target audience, and maintain coherence between text and image. In this work, we first provide an explicit task analysis of cross-cultural meme transcreation and identify three core challenges: culture-specific knowledge understanding, intent and tone preservation, and multimodal consistency. Based on this analysis, we propose a multi-agent framework with specialized agents that are coordinated to address these challenges through cultural adaptation, target text rewriting, revision, and conditional visual adjustment. The framework strengthens target text adaptation with coordinated feedback to handle difficult cases that require deeper cultural or visual intervention. We evaluate the framework on bidirectional Chinese-English meme transcreation using both human evaluation and LLM-as-a-Judge. Our method consistently outperforms all baselines across both evaluation settings. In human evaluation, it achieves the best performance on all four dimensions and delivers a 33.1% average improvement over the strongest baseline, while in LLM-as-a-Judge, it attains the highest Top-1 ranking rate (60% versus 26% for the second-best baseline). Further analysis indicates that each component contributes to the performance. Our error analysis suggests that the remaining bottlenecks lie in humor reconstruction and image-text alignment rather than simple cultural knowledge gaps, pointing to future work on humor transfer.
[AI-19] LAAF: A Layered Accountability Architecture Framework for LLM Applications
链接: https://arxiv.org/abs/2608.27102
作者: Prachi Chaturvedi,Shahnawaz Ahmad,Ehsan Nowroozi,Muhammad Waqas,George Loukas,Alireza Jolfaei,Lucas Cordeiro,Pierre Dantas
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:Large Language Models (LLMs) operate in hospitals, courtrooms, banks, and public service desks, where fluent, confident outputs are treated as authoritative even when ungrounded or incorrect. When such an output contributes to harm, who is answerable, and through what mechanisms can responsibility be traced, explained, and acted upon? Following PRISMA guidance, five databases were searched from January 2022 to March 2026 against four review questions; of 4,512 records identified, 122 primary studies were included, together with 12 regulatory and standards documents analysed as primary sources. The review consolidates a sociotechnical account of accountability as an actor-forum relation resolved into five dimensions, and synthesises mechanisms across four families: technical controls, human oversight, organisational governance, and documentation and traceability, each with a maturity assessment. The corpus is read through a four-layer classification device spanning provenance, application logic, human oversight, and governance and redress, cross-cut by traceability, role clarity, and continuous monitoring. Both are mapped onto the EU AI Act, whose high-risk obligations have applied since 2 August 2026, the NIST AI RMF with its Generative AI Profile, ISO/IEC 42001, and sectoral guidance in healthcare, consumer finance, education, and the public sector. Four persistent gaps emerge: under-specification of human oversight, absence of shared accountability metrics, disciplinary disconnection, and limited empirical evaluation, alongside five structural tensions that no surveyed instrument resolves. The review closes by consolidating the classification device into an integrated accountability architecture, LAAF, with cybersecurity aligned to the OWASP LLM Top 10 (2025); it is a synthesis of the surveyed evidence rather than a validated artefact.
[AI-20] pro-team at LLM s4OL 2026 Tasks Flagship and Reuse: Retrieval-Augmented Generation and Vocabulary-Constrained Filtering for Ontology Learning
链接: https://arxiv.org/abs/2608.27101
作者: Shivam Mishra,Dhannu Ram Meena,Muneendra Ojha,Krishna Pratap Singh,Kuldeep Singh
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Ontology learning from text remains challenging despite significant progress in Large Language Models (LLMs), which can hallucinate domain terms, produce inconsistent formats, and favor hierarchical over associative relations. In the LLMs4OL 2026 Challenge, we address both the End-to-End Flagship Task (Task A) and Ontology Extension Reuse Task (Task B) using an offline retrieval-augmented few-shot prompting pipeline. Our system employs Qwen2.5-14B-Instruct with all-MiniLM-L6-v2 for demonstration retrieval, selecting the top-5 examples for Task A and top-2 for Task B. A left-truncated context-windowing strategy preserves task instructions within long prompts. For Task B, generated triples undergo deterministic vocabulary-constrained filtering, retaining triples when at least one endpoint belongs to the sample’s closed term/type vocabulary and removing duplicates of the initial ontology. The approach achieves Semantic Graph Similarity of 0.8692, Term-Typing F1 of 0.9200, and Taxonomy Discovery F1 of 0.8540 on Task B, while Task A achieves 0.7416 Semantic Graph Similarity. However, no non-taxonomic relations are extracted, highlighting limitations of closed, taxonomy-oriented relation vocabularies.
[AI-21] Active sensing to characterize the heterogeneity of plant stress
链接: https://arxiv.org/abs/2608.27088
作者: Ayman Laaroussi,Peter Hanappe,David Colliaux
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: UR2026
Abstract:While most phenotyping platforms rely primarily on image-based measurements, advanced plant characterization requires the integration of active physiological sensing modali- ties such as chlorophyll fluorescence. We present an autonomous robotic platform designed to perform targeted fluorescence measurements on plant leaves. The system combines 3D plant reconstruction, geometric analysis, and motion planning to localize suitable measurement points and generate collision-free trajectories for a robotic manipulator. A dense 3D model of the plant is reconstructed from multi-view data and used to extract candidate leaf surfaces based on orientation, accessibility, and sensing constraints. These targets are then integrated into a task-level planning framework that guides the end-effector to precise contact or near-contact configurations required for point-based fluorescence acquisition. The platform enables automated, repeatable, and spatially resolved physiological measurements that go beyond passive imaging. By tightly coupling perception, geometric reasoning, and manipulation, the proposed system provides a robotics-driven approach to high-resolution plant phenotyping and opens new directions for autonomous agricultural inspection and plant-aware manipulation. Comments: UR2026 Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI) Cite as: arXiv:2608.27088 [cs.RO] (or arXiv:2608.27088v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2608.27088 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-22] Emotional Preferences as Goal-Priority Regulation
链接: https://arxiv.org/abs/2608.27072
作者: Shiqi Liu,Yihua Tan,Hu Fu,Guanyu Qi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:A core question in decision-making for agents is whether the relative priorities of competing lower-level objectives can be determined by emotional preferences autonomously generated by higher-level goals, rather than being externally prespecified. Under changing external environments and evolving internal states, emotions play an important functional role in regulating the relative priorities of competing goals. Inspired by the goal-directed theory of emotion, this paper studies how such preference regulation can be computationally realized through reinforcement learning. We first propose a conception of emergent emotional preference: a high-level goal autonomously induces state-dependent preferences over competing lower-level objectives. This conception is built upon a framework consisting of a multi-objective reinforcement learning inner controller and an outer preference generator. The inner controller provides a repertoire of preference-conditioned goal-directed behaviors, while the outer preference generator learns a mapping from the current state to objective preferences through reinforcement learning on a high-level goal. We operationalize emotional preference as a state-dependent regulation of relative goal priorities that emerges through optimization. Furthermore, we characterize the policy space induced by preference regulation and derive an upper bound on the optimality gap in terms of the representation error of the inner behavioral repertoire. We show that the gap vanishes when the optimal policy can be represented by the available preference-conditioned policies. Experiments in self-constructed multi-objective exploration environments show that the learned preference function exhibits contextual priority switching, graded trade-offs, and temporal persistence, and outperforms the evaluated fixed-preference and handcrafted-preference strategies.
[AI-23] Performance Foundations of Parallel Distributed Reasoning Language Models
链接: https://arxiv.org/abs/2608.27046
作者: Maciej Besta,Leonard Schmidt,Lara Nonino,Robert Gerstenberger,Pierre Pang,Patrik Okanovic,Ales Kubicek,Tiancheng Chen,Baraq Lipshitz,Torsten Hoefler
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF)
备注:
Abstract:Reinforcement Learning with Verifiable Rewards (RLVR) and other RL-style post-training paradigms have been used for aligning large language models (LLMs) with reasoning standards. The resulting recent Reasoning Language Models (RLMs) such as DeepSeek-R1, o3, and Kimi k1.5 show that such RL-style post-training (“RL-for-LLMs”) can substantially improve chain-of-thought reasoning, long-horizon planning, and self-correction. However, the computational footprint of these systems is massive: state-of-the-art RLM training requires millions of GPU-hours and tightly coupled multi-model pipelines that stress modern hardware far beyond classical supervised LLM training. This makes RLM training as much a parallel and distributed systems problem as an algorithmic one. In this work, to facilitate developing RLMs that are simultaneously high-performance, scalable, and cost-effective, we first systematize the RL-for-LLM paradigm and provide a compute-centric analysis of prominent post-training algorithmic frameworks: Proximal Policy Optimization (PPO), Group Relative Policy Optimization (GRPO), as well as their variants. Second, we develop a taxonomy of intra- and inter-model parallelism strategies for RL-for-LLMs, covering both traditional techniques (data, tensor, pipeline, sequence, context, and expert parallelism) as well as novel forms of parallelism and optimization techniques for multi-model RLM training, for example disaggregated placement, stage fusion, hybrid parallelism, and asynchronous execution. We harness the work-depth model of parallel computing to make our taxonomy and its insights rigorous and portable. Finally, we analyze existing RLM frameworks and we distill practical guidelines and outline open research directions for building scalable, fast, and cost-effective RLMs.
[AI-24] FaulT-Bench: Towards Benchmarking Network Troubleshooting LLM Agents under Unreliable User Tickets
链接: https://arxiv.org/abs/2608.27021
作者: Kuan-Hao Tseng,Niruth Bogahawatta,Yasod Ginige,Kunjan Patel,Kosta Dakic,Suranga Seneviratne
类目: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI)
备注:
Abstract:LLM-based agents are increasingly proposed for network fault diagnosis, but existing benchmarks evaluate them only on accurate tickets and always assume a fault is present, conditions rarely met in practice. We present FaulT-Bench, a benchmark of 200 troubleshooting scenarios across eight network topologies, five reimplemented from public practitioner labs, spanning genuine faults, false fault reports, incorrect device attribution, and incorrect root-cause claims. To isolate how ticket wording affects diagnosis, we further rewrite 72 false-premise tickets into five reporter personas that vary reporter confidence and verifiable detail one factor at a time, holding the network state fixed. Our automated harness deploys each scenario in Kathará, lets agents interact through the NIKA tool interface, and scores free-text diagnoses with an LLM judge across outcome, fix, and reasoning quality. Evaluating SADE, ReAct, and Claude Code, we find all three are near-saturated on accurate tickets and robust to misdirection, yet degrade sharply when the network is healthy and the ticket is wrong, probing until a benign condition can be promoted to a root cause rather than concluding nothing is wrong. Persona rewrites show that how a ticket is written matters more than what it claims: a confidently wrong report is handled about as well as an accurate one, while a vague, underspecified report degrades performance sharply. The three agents also fail differently, from constant over-diagnosis to unanswered runs, at very different cost. These results position FaulT-Bench as a benchmark for developing agentic systems that can reason reliably over the noisy, unreliable tickets of real-world network troubleshooting.
[AI-25] A Multi-Modal AI Framework for Real-Time Queue Prediction Management and Optimisation in Intelligent Border Control Systems
链接: https://arxiv.org/abs/2608.27010
作者: Varvara Mama,Eleni Veroni,Nikolaos Kapsalis,Christos D. Nikolopoulos,Anargyros T. Baklezos
类目: Artificial Intelligence (cs.AI)
备注: 6 pages, 2 figures, conference
Abstract:In the present work an efficient border control management procedure is proposed. Compared to operational queue management systems, whose operations are based on mostly static data, the proposed work takes into account dynamic traffic conditions, thus enabling optimal performance, even in cases of uncertainty. To this end, we are proposing a multi-modal Artificial Intelligence (AI) framework, tailored to th needs of border control systems, which enables real-time queue prediction, management, and resource optimization. The novel proposed approach integrates heterogeneous data sources and presents them through a unified representation by employing Long Short-Term Memory (LSTM) networks for queue forecasting. Furthermore, it leverages Model Predictive Control (MPC) and scheduling optimization to derive actionable control policies, which in turn can be presented to border control officers. The proposed work has been evaluated using synthetic data simulating realistic traffic. The evaluation results demonstrate that the proposed method reduces queue prediction error by up to 35% and average waiting time by 30%. Accordingly, the average throughput increases by nearly 20%, compared to ARIMA and rule-based methods. The abovementioned results show the effectiveness and efficiency of combining AI architectures with optimization techniques for proactive and adaptive border traffic management.
[AI-26] ASIL: Replacing Screenshot-and-Click with Structured State and Semantic Actions EMNLP2026
链接: https://arxiv.org/abs/2608.26991
作者: Rui Xie,Lu Chen
类目: Artificial Intelligence (cs.AI)
备注: Accepted to Findings of EMNLP 2026. 19 pages, 7 figures: 9-page main paper followed by limitations, ethics, acknowledgments, references, and appendices A-E. Project page: this https URL
Abstract:Powerful code agents can execute scripts, call tools, and manage files, yet many important applications remain accessible primarily through graphical user interfaces. We argue that screenshot-and-click is an inefficient interface for software-operating agents: screenshots are state-incomplete, and GUI actions are brittle, semantically weak, and poorly matched to long-horizon planning. We introduce ASIL (Agent-Software Interaction Layer), an agent-native interface that exposes software through structured JSON observations and code-executable semantic actions, realized through the deepest feasible access path for each application. We instantiate ASIL across 15 applications and a benchmark of 300 single-application and 80 multi-application tasks. ASIL reaches above 80 with closed models while executing fewer than five actions per task. Under a repaired runtime and a 50-step screenshot budget, the same tasks yield 6.6 and 26.6 strict success under screenshot-and-click control, rising to 15.0 and 53.3 on an easier OSWorld-comparable band. Against application-native interfaces on matched tasks, ASIL exceeds LibreOffice’s UNO API by 28-38 strict points but only matches this http URL’s MCP content contract. The structured modality also suits training: small-scale SFT raises Qwen3.5-2B from 58.0 to 72.1 and Qwen3.5-9B from 66.6 to 80.4, and resource-limited on-policy RL further raises them to 74.4 and 82.2.
[AI-27] GraphMemix: Query-Aware Evidence Forests for Long-Term Multimodal Agent Memory
链接: https://arxiv.org/abs/2608.26983
作者: Geng Li,Yuhao Wang,Dong Li,Jianye Hao,Yuxin Peng
类目: Artificial Intelligence (cs.AI)
备注: Project page with code: this https URL
Abstract:Organizing long-term memory for multimodal agents remains challenging because existing methods either suffer from expensive question-agnostic offline summaries or naive embedding similarity matching that introduces incomplete and redundant context. To address these issues, we propose GraphMemix, a combinatorial-optimization graph memory framework that models memory organization as query-aware evidence-forest construction. Specifically, our method consists of three key components:(1) candidate graph construction, which expands multi-view seed memories through schema and semantic relations to acquire query-aware original context; (2) evidence utility and activation costs, which decouples direct memory support from anchor-conditioned relation verification to suppress redundant or conflicting information; and (3) forest optimization, which jointly selects a forest-format memory context under a maximum evidence budget and its reliable relational structure. By organizing memory into a query-relevant subgraph, the method avoids substantial lifecycle cost and recovers low-similarity complementary evidence. Experimental results across four long-term multimodal memory benchmarks demonstrate significant improvements with different foundation models and establish a new Pareto frontier between accuracy and lifecycle cost.
[AI-28] A Table Is Worth 64 Tokens: Pixel-level Compression for Multi-Table Document Question Answering
链接: https://arxiv.org/abs/2608.26949
作者: Iñigo Alonso,Mirella Lapata
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Answering questions over real-world documents requires processing long inputs that interleave text with tables. Optical context compression, which represents context as images, promises to reduce token cost, but its effect on table understanding remains unclear. We study pixel-level table compression for question answering over documents with multiple tables, evaluating five VLMs across two benchmarks and five visual-token budgets. Representing tables as images at native resolution matches text in both performance and efficiency, but downscaling them makes models compensate the loss in readability with longer, less effective reasoning traces that cancel the expected savings. Highly downscaled tables, however, preserve enough signal to identify whether they are relevant to a question. We exploit this asymmetry with a training-free, two-step method: the model first identifies the tables needed to answer a question from a pixel-compressed context, and then reasons over those at native resolution. On long documents, our method saves 41% of total tokens and gains 7 accuracy points over single-step QA with native resolution tables. It also uses 15% fewer tokens than the most efficient single-step compressed configuration, with no accuracy loss.
[AI-29] Counterfactual Bias Testing for Application Tracking System
链接: https://arxiv.org/abs/2608.26899
作者: Sai Yashwant,Shruti Bansal,Anurag Dubey,Samaroha Chatterjee,Satyam Kumar,Shreyash Gupta,Gantala Thulsiram
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Automated candidate-job matching systems are increasingly classified as high-risk AI under emerging regulation, yet auditing them for demographic bias is expensive: classical correspondence-audit studies require hand-crafted resumes and manual submission, which does not scale to fast pipeline retraining cycles. This paper presents a general, reusable methodology that (1) uses task-specialized LLM agents to synthesize identity-neutral base resumes and inject controlled demographic treatments across five protected-characteristic axes (sex/gender, age, residence, language, disability), producing a K x (1+N) correspondence-audit matrix; (2) qualitatively flags inferred protected characteristics per an EU AI Act-aligned prompt; (3) ranks candidates against a job description via a fine-tuned sentence-embedding model and cosine similarity; and (4) computes a nine-metric fairness suite spanning counterfactual (score delta, mean absolute rank change, flip rate), group-fairness (top-K retention, four-fifths/impact ratio), and merit-aware (Recall@K, nDCG@K, equal opportunity, equalized odds) families, each with bootstrap confidence intervals, significance tests, and Benjamini-Hochberg correction, culminating in an automated PASS/INVESTIGATE/FAIL report with a composite risk score. On an example corpus of 5 job orders, 100 base candidates, and 10 demographic treatments (90 metric x variant evaluations): score shifts, top-K retention, and merit-aware rate gaps stay within tolerance for every treatment, but a rank-stability metric (MARC) and nDCG@K each surface borderline findings - including one on the neutral baseline itself - that a score- or retention-only view would miss. The results argue for multi-metric, multi-family auditing over any single aggregate score, and for LLM-agent-generated audits as a practical, low-cost complement to human-curated audits for any candidate-job matching pipeline.
[AI-30] Learning-Augmented Online Allocation under Unreliable Advice: Robustness Exposure Fairness and Distribution Shift
链接: https://arxiv.org/abs/2608.26889
作者: Fredy Pokou(MRE, CRIStAL)
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Learning-augmented algorithms improve online decisions using predictions, but unreliable advice may harm efficiency and fairness. We study an online allocation problem with finite candidate sets, irreversible decisions, and exposure constraints. We propose a robust and fair rule combining advice with a conservative fallback and fairness correction. Under bounded-error assumptions, we prove consistency and robustness with loss proportional to prediction error. Experiments show stability under adversarial advice and significant reductions in exposure disparity.
[AI-31] PLCBench: Can Autonomous LLM Agents Turn PLC Access into Sustained Physical Impact?
链接: https://arxiv.org/abs/2608.26882
作者: Yitian Zhou,Jingyu Zheng,Qiliang Jiang,Linkang Du,Haoming Liu,Lichao Wu,Shiyi Zhao,Mengxiang Liu,Ruilong Deng
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 36 pages, 13 figures
Abstract:Industrial control systems (ICSs) rely on programmable logic controllers (PLCs) to connect networked computation with physical control. Tool-using large language model (LLM) agents represent an emerging attack threat: can an autonomous agent convert a network-reachable PLC into sustained adverse physical impact? However, existing evaluations focus on digital tasks or individual stages of PLC testing. In ICSs, evaluations that stop at software exploitation, an accepted write, or tool access may therefore mischaracterize physical risk. We present PLCBENCH, to our knowledge, the first real-PLC hardware-in-the-loop (HIL) framework for characterizing this cyber-to-physical capability and its boundaries. It combines vendor-native interaction, commercial PLC execution, closed-loop reduced-order process simulation, and independent outcome verification. A deterministic evaluator applies fixed rules to runner, communication, PLC-object, and process records to assign six hidden diagnostic flags, distinguishing usable PLC interaction, process-linked manipulation, and sustained physical impact. We instantiate PLCBENCH on four commercial PLCs crossed with four closed-loop workloads. Across five LLM families and 240 real-PLC episodes, 75 episodes (31.3%) sustain their respective physical objectives. Stagewise results show that 98 episodes stop before a valid native read, whereas 62 reach a process-linked write but do not sustain the final objective. Notably, richer process observation is associated with an increase in conditional objective attainment after a process-linked write from 44.2% to 64.0%. These measurements localize failure in configured PLC-process deployments and identify intervention points for future defense evaluation. To support reproducibility, we release the safely disclosable PLCBENCH code and a software-only reproduction pipeline through the accompanying artifact. Comments: 36 pages, 13 figures Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) Cite as: arXiv:2608.26882 [cs.CR] (or arXiv:2608.26882v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2608.26882 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-32] BekchiAI: Measuring Observing and Controlling LLM Agents in One Click
链接: https://arxiv.org/abs/2608.26867
作者: Mesut Toruk
类目: Artificial Intelligence (cs.AI)
备注: 8 pages, 1 Figure, 6 tables
Abstract:Large language model agents reason, call tools, and act autonomously over many steps, but their agentic skills-correctly sequencing tools, planning under dependencies, judging untrusted inputs, and grounding generated arguments-are hard to measure with accuracy-only leaderboards. We present BekchiAI, which addresses both sides: a benchmark for measuring agentic skill and a platform for observing and controlling live agents. The BekchiAI-Benchmark, a suite of 13 tool-using ReAct agents across 7 task categories (arithmetic, structured/SQL, security detection, URL grounding, planning, orchestration, and tool-policy), totalling 2,057 deterministic, committed test tasks. Every task is verifier-checkable gold answers are computed by running canonical SQL against a real database, computing the exact schedule of a directed acyclic graph (DAG), or evaluating closed-form lambdas including adversarial security samples paired with deliberately imperfect signature scanners so a score reflects the model’s own judgment, not the copying of an oracle. We define a small set of behavioral metrics beyond accuracy-tool-call adherence, URL hallucination and source-match, and per-model token cost and report a four-model comparison (Qwen3.7-Max, gemma-4-31B-it, gemma4:26b, gpt-oss-120b) whose story is in the per-family spread, not the aggregate. The benchmark runs are executed using the provided evaluation scripts. BekchiAI-Platform is a complementary web-based observability and control layer for deployed agents, providing full token and latency telemetry as well as remote run termination. The benchmark, evaluation tools, and platform are publicly released.
[AI-33] Reinforcement Learning-Based Control of CAV Platoon Joining Maneuvers in Mixed Traffic
链接: https://arxiv.org/abs/2608.26860
作者: Biao Yin,Abderrahmane Kasmi,Nadir Farhi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Optimization and Control (math.OC)
备注:
Abstract:Connected and automated vehicle (CAV) platooning offers a promising approach to improving road safety and traffic capacity. However, platoon control in real-world traffic is challenging due to uncertainty and heterogeneous driving behaviors. Reinforcement learning (RL) has strong potential for addressing such control problems, but its practical deployment raises challenges related to safety and learning efficiency. This paper proposes a generic modeling and simulation framework for investigating CAV platoon joining maneuvers and comparing deep reinforcement learning (DRL)-based control algorithms. The problem is particularly challenging in mixed-traffic environments, where CAVs coexist with human-driven vehicles exhibiting heterogeneous longitudinal and lateral behaviors. The objective is to achieve safe and efficient joining maneuvers by either incorporating penalties for risky behaviors into the learning process or using an external safety controller to constrain the learned policy. An agent-based modeling framework coupled with the Simulation of Urban MObility (SUMO) simulator is used to evaluate Deep Q-Network (DQN), Double Deep Q-Network (DDQN), and Proximal Policy Optimization (PPO). Results show that PPO outperforms DQN and DDQN, achieving a joining success rate of approximately 98 % and a collision rate below 1 %, largely due to risk-related penalties incorporated into the reward function. However, this improved performance requires more decision steps to complete the maneuver, revealing a trade-off between safety, joining effectiveness, and decision efficiency. An external safety controller effectively prevents collisions, although its interventions may reduce joining efficiency. The results highlight the importance of jointly considering safety and efficiency when designing RL-based controllers for CAV platoon joining in mixed traffic.
[AI-34] AI Control Scientist: LLM -driven Agent ic System for Automated Control Design
链接: https://arxiv.org/abs/2608.26780
作者: Haiteng Wang,Weihao Li,Jing Zhang,Lei Ren
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Control system design is critical for modern industry, such as chemical process temperature regulation and aero-engine control. However,traditional control design workflows rely heavily on expert knowledge and extensive manual parameter tuning, resulting in limited efficiency and scalability. To this end, this paper proposes AI Control Scientist (AICS), the first large language model (LLM)-driven agent capable of automatically generating optimized controller from language design requirements. Specifically, a Task Modeling Agent interprets user requirements to engineering constraints; a Controller Design Agent generate candidate controller structures and executable code; and a Parameter Tuning Agent refine controller parameters under closed-loop performance criteria. Experiments demonstrate that the proposed agentic system can automatically generate multiple representative control systems, outperforms existing automated baselines in both design success rate and optimization efficiency. This work has the potential to transform control system design from human-driven to agent-driven, paving the way for model predictive control and other advanced control systems design.
[AI-35] Categorizer Automata for Discounted-Sum Payoffs
链接: https://arxiv.org/abs/2608.26763
作者: Nathalie Bertrand,Pranav Ghorpade,Senthil Rajasekaran,Sasha Rubin,Moshe Vardi
类目: Artificial Intelligence (cs.AI); Formal Languages and Automata Theory (cs.FL)
备注:
Abstract:Categorizing continuous data into discrete bins is a fundamental operation in artificial intelligence. We introduce the categorizer automaton, a deterministic automaton that reads an infinite sequence of rewards and identifies which of finitely many bins contains its discounted sum. Categorizer automata generalize comparator automata, the special case of two bins, which have already proven useful in quantitative synthesis. Our main technical contribution is the construction of a categorizer automaton whose state space is linear in the number of bins, rather than exponential as obtained by a cross-product of comparator automata. We then apply categorizer automata to Markov decision processes, where they allow one to synthesize policies that maximize the expected utility of a discounted-sum payoff for utility functions that may be discontinuous. For piecewise-constant utility functions, the resulting algorithm is exact and runs in pseudo-polynomial time. For piecewise-Lipschitz utility functions, a class that includes any utility with bounded slope between finitely many jumps, it again runs in pseudo-polynomial time and yields an \varepsilon -optimal policy. We also show that the synthesis problem considered is PSPACE-hard already for piecewise-constant utilities.
[AI-36] DEEPCHART: How Far are LLM s from Faithful Data-Science Chart Generation?
链接: https://arxiv.org/abs/2608.26757
作者: Jiahui tang,Kuicai Dong,Dexun Li,Hongchao Gu,Haocheng Yu,Wei Han,Chen Zhang,Yong Liu,Hao Wang,Enhong Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Faithful chart generation in real-world data-science workflows requires grounding visualizations in scattered evidence, computing chart-ready quantities, and rendering them accurately. Modern LLMs can produce visually plausible, instruction-compliant charts, yet data-level hallucinations remain difficult to detect in long, noisy, and multimodal contexts. To measure this gap, we introduce DEEPCHART, an expert-annotated benchmark of 1,482 task-conditioned chart-generation instances drawn from real-world scientific papers, financial filings, and ecosystem reports. DEEPCHART formulates chart generation as an Extract–Reason–Visualize pipeline and evaluates source-data extraction, derived-data reasoning, and chart rendering stage by stage. Experiments with state-of-the-art models show that visually plausible charts often conceal data-level hallucinations, with extraction and reasoning errors common in realistic long and multimodal settings. These findings suggest that larger context windows alone are insufficient; faithful chart generation also requires reliable evidence extraction and quantitative reasoning before rendering. Our benchmark and associated resources are available at this https URL.
[AI-37] Beyond Execution: Auditing Experimental Fidelity in LLM -Driven Scientific Research
链接: https://arxiv.org/abs/2608.26753
作者: Lezhi Yu,Xiaogang Xu,Yuhua Zhou,Shuibing He,Aimin Pan
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 20pages, 5 figures, code link: this https URL
Abstract:LLM agents used for scientific experimentation must do more than generate executable code: they must implement the reference method faithfully, design experiments that test the paper’s claims, and provide evidence supporting those claims. We show that agents often produce methodological hallucinations: silently reducing datasets or training budgets, replacing failed learning or generative components with lookup or oracle functions, or drawing conclusions from resource-limited settings where a method’s claimed advantage disappears. To detect these failures, we introduce ABE-Ralph, a reference-anchored auditing framework that represents claims, protocols, required components, baselines, and metrics as structured experimental constraints, guides implementation through an 8-step workflow, and performs quantitative, qualitative, and code-level verification. Across 30 long-horizon reproduction runs covering 12 machine learning domains, ABE-Ralph achieves a 93% robust execution rate and identifies five scientific failure modes. In 23 NatureBench discovery tasks, ABE-Ralph matches or exceeds state-of-the-art performance on 5 tasks. These results show that reliable evaluation of AI scientists must assess whether the experimental design faithfully tests the intended claim and whether the resulting evidence supports it, rather than treating code execution or plausible metrics as evidence of scientific success.
[AI-38] Discovering Relationships in Data Lakes Using Large Language Models : An Industrial Case
链接: https://arxiv.org/abs/2608.26750
作者: Ahlame Diouan(ERIC, UL2),Eric Ferey,Sabine Loudcher(ERIC, UL2),Jérôme Darmont(ERIC, UL2)
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Data lakes rely on metadata to remain usable, yet this meta data is often limited or weakly informative for column relationship discovery, especially in ERP-derived datasets with coded or abbreviated schema labels. We propose ColRel, a two-stage method that builds column embeddings from metadata and data available at ingestion time. In difficult cases, such as coded schemata, business dictionaries help better interpret column names and support the generation of short natural-language descriptions used in the second stage. Experiments on public benchmarks and an industrial ERP dataset show that ColRel is particularly effective in semantically related, weak-signal settings.
[AI-39] Agent Fold: Closed-Loop Agent ic Search for Protein Folding Model Design
链接: https://arxiv.org/abs/2608.26747
作者: Mingquan Liu,Jiangyu Chen,Hanqun Cao,Xujun Zhang,Pengsen Ma,Xiangru Tang,Shuting Jin,Zhuo Yang,Tianfan Fu,Fang Wu,Xiangxiang Zeng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Scientific LLM agents have shown promise in literature reasoning, tool use, and experiment planning, but it remains unclear whether they can autonomously improve large, tightly coupled scientific machine-learning systems through executable code changes and computationally expensive validation. We study this question in protein folding, where progress requires coordinated architectural modifications, multi-objective evaluation, and domain-aware interpretation. We present AgentFold, a multi-agent framework that formulates folding-model development as a closed-loop search over executable code variants. Starting from ESMFold, AgentFold proposes hypotheses, implements and debugs code-level modifications, evaluates model variants, analyzes experimental outcomes, and stores both successful and failed interventions in structured memory. An MCTS-style policy allocates computational resources across high-scoring search branches. On an engineering-scale protein-folding codebase comprising more than 2,000 lines of code, AgentFold explores approximately 80 model variants using approximately 5,000 GPU-hours and 170 million LLM tokens. Under a matched computational budget, AgentFold improves the best lDDT by 7.5% over independent Codex proposals and outperforms a random-search control. Beyond model improvement, the resulting intervention traces reveal recurring empirical design patterns: stable gains tend to arise from early, soft, learnable priors and gated refinement, whereas direct geometric perturbations and geometry-conditioned feedback often destabilize training. The code and experimental resources are publicly available at this https URL.
[AI-40] FaultLens: Learning Compact Behavioral Test Suites for Generated Operational Programs
链接: https://arxiv.org/abs/2608.26746
作者: Zeming Liu,Hang Lyu,Jingtao Zhang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 13 pages, 7 figures. Includes a complete reproducibility artifact in the source package
Abstract:Generated operational programs are often validated with either a few hand-written examples or exhaustive regression suites. The former can miss sparse boundary and interaction faults, while the latter can be unnecessarily expensive. We introduce FaultLens, a method for learning compact behavioral test suites while preserving an auditable connection to executed evidence. It executes a rich probe domain once, stores the fault-probe kill relation as a sparse outcome cache, and learns probe orderings only from earlier program generations. A fault-driven greedy component exploits known kill structure, while a mutation-independent diversity component covers probe families, cases, templates, and temporal bins. Their alternating hybrid remains useful when a new program contains a fault mechanism absent from ordering construction. We evaluate twenty generated operational policies across four environments, ten execution seeds, 1,200 measured run summaries, 2,160 controlled program transformations, and 4,120,200 executed program-probe pairs. Of 1,960 intended faulty transformations, 1,779 alter a contract or output somewhere in the finite audit domain; 200 additional controls preserve behavior. A 32-probe hybrid learned on generations 1-3 covers 576/582 (99.0%) dynamically killable faults in generations 4-5 using 1.2-2.0% of the exhaustive domain. With an entire fault family withheld from training, diversity raises scenario-family macro coverage from 84.6% to 94.9%. In a downstream deployment study, a conservative admission rule reduces severe tail regressions from 15/20 program-environment groups to 0/20. FaultLens provides a prioritized evidence mechanism, not a proof of correctness, and makes its budget, evidence source, generalization split, and misses explicit. Comments: 13 pages, 7 figures. Includes a complete reproducibility artifact in the source package Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI) Cite as: arXiv:2608.26746 [cs.SE] (or arXiv:2608.26746v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2608.26746 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-41] Graph-Guided Selective Unlearning for Language Models: Controlling Support Routes Beyond Forget Seeds
链接: https://arxiv.org/abs/2608.26743
作者: Waqas Khan,Tabinda Sarwar,Jingyue Cong,Xun Yi,Estrid He
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Enterprises fine-tune language models on proprietary data that may later require removal due to privacy, contractual, or compliance obligations. Selective unlearning removes requested knowledge while preserving model utility, offering a practical alternative to full retraining, but existing methods treat the explicitly identified forget examples as the complete deletion scope. This is insufficient when target knowledge remains recoverable through paraphrases, aliases, or neighboring training examples. We propose GRAPHSU, a graph-guided controller that expands the deletion scope beyond forget seeds by constructing a weighted support-route graph, propagating deletion pressure through it, and applying graded forgetting strengths to high-risk neighbors. On the Task of Fictitious Unlearning (TOFU), a synthetic author-profile question-answering benchmark, and PISTOL, a structural-unlearning benchmark built around interconnected factual samples, with GPT-2 Medium and Llama-3.2-3B-Instruct, GRAPHSU achieves the lowest utility-feasible soft leakage across all deletion settings, reducing leakage by up to 49.5 percentage points over a matched seed-only baseline, demonstrating that effective enterprise unlearning requires controlling support routes, not just forget seeds.
[AI-42] Daydreaming: Stealing Hidden Agent Skills through Black-Box Task Interaction
链接: https://arxiv.org/abs/2608.26733
作者: Yu-Lin Tsai,Yu-An Lu,Ci-Yang Tsai,Muxi Lyu,Raluca Ada Popa,Chia-Mu Yu
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Agent skills bundle instructions, reference data, and executable helpers that let a general agent perform specialized tasks. Hosted providers can keep these files secret while selling access to task results, making the skill itself a valuable target. Existing disclosure defenses can block requests that ask for the skill or reproduce its text, but they cannot block customers from submitting the ordinary tasks the service is built to complete. We present Daydreaming, an execution-only attack that steals a multi-file skill through black-box task interactions. The victim is never asked to reveal the skill or grade a reconstruction. Instead, Daydreaming adaptively creates crafted tasks whose results distinguish possible hidden behaviors. It tests individual behaviors, uses attacker-controlled shadow agents to choose a design, and completes each file using stored victim results and local execution checks. We formalize three nested threat levels of access as Differential, Trace, and Output, and focus on Output, where the attacker sees only the final response and returned files. Across 7 skills and 4 victim models, Daydreaming recovers 86.8% of the original skill’s capability at Output, outperforming SigLeak by almost 4x. It produces installable skills using a median of 32 victim calls per skill even with disclosure defenses enabled. These results show that hiding skill files and filtering direct disclosure do not, by themselves, prevent functional reconstruction through normal use.
[AI-43] Rethinking Message Passing as Retrieval for Text-Attributed Graph Learning
链接: https://arxiv.org/abs/2608.26732
作者: Jintang Li,Yuhong Chen,Ruofan Wu,Binli Luo,Jiayi Ji,Hui Li,Rongrong Ji
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Graph neural networks (GNNs) are typically conceptualized as message-passing neural networks, yet it remains unclear why neighborhood aggregation reliably outperforms node-wise multilayer perceptrons (MLPs). Despite its empirical success, this paradigm can be computationally expensive and sensitive to imperfect graph structures. In this work, we present a retrieval-augmented view of GNNs: each layer makes predictions by applying an MLP to a node representation together with a permutation-invariant summary of retrieved graph context. Motivated by this perspective, we propose RTA, a simple MLP-based framework that replaces structural message passing with label-aware retrieval and propagation. We provide theoretical insights that (i) connect retrieval-based aggregation to softmax-attention message passing, and (ii) establish the robustness of retrieved-context supervision to mis-retrieved outliers. Experiments on multiple text-attributed graph benchmarks show that RTA matches or even outperforms strong GNN and graph LLM baselines while improving efficiency and robustness across diverse scenarios.
[AI-44] Knowing When Not to Reuse: Conditional Experience Transfer in Autonomous LLM Post-Training
链接: https://arxiv.org/abs/2608.26730
作者: Tingyun Li,Wenfeng Feng,Weiqing Li,Abudukelimu Wuerkaixi,Guohua Liu,Yuewei Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models offer broad capabilities, but adapting them to evolving domains, tools, and requirements often entails repeated post-training. Autonomous systems automate parts of this process by proposing updates, training candidates, and using evaluation feedback to select subsequent proposals. As evidence accumulates, a central problem emerges: which past update evidence remains actionable after subsequent training has changed the parent model? An update’s effect depends on its parent, data, and training stage. Treating past success as context-free permission can waste compute. If the resulting child is promoted, it can also degrade the subsequent training trajectory. We formulate this problem as conditional experience transfer and introduce Boundary-Calibrated Intervention Transfer (BCIT), a method that authorizes experience reuse before weight-changing training. BCIT binds an observed effect to its source context, checks applicability conditions, vetoes candidates with named hard conflicts, and obtains current-state evidence through a bounded training trial when needed. Fully trained candidates still face a shared adoption rule, and only observed events extend memory. On one 4B model adapted across finance reasoning, text-to-SQL, and function calling, candidate updates exhibit heterogeneous target and retention effects across the evaluated contexts. Under matched candidates, evidence, and compute, BCIT authorizes fewer harmful updates and attains higher equal-budget final-model quality than the evaluated alternatives. These results support treating experience authorization as a distinct problem in autonomous post-training.
[AI-45] Style as a Confound: False Positives in AI Detection of Non-Native Academic Writing EMNLP2026
链接: https://arxiv.org/abs/2608.26710
作者: Hyeonchu Park,Gahye Jeong,Bugeun Kim
类目: Artificial Intelligence (cs.AI)
备注: Presented in EMNLP 2026
Abstract:AI text detectors are increasingly employed in academic settings, but it remains unclear whether their outputs reflect AI authorship itself or broader linguistic features associated with polished academic English. Previous studies have reported high false-positive rates (FPRs) for non-native English writing, but population-level comparisons confound authorship with differences in topic, domain, and writing style. Professional editing provides a useful setting for examining this issue because it changes the linguistic form of manuscripts while preserving authorship and content. We examined 135,389 document pairs from a professional English editing service (2018-2025), comprising non-native manuscripts and their native-edited versions, to assess how editing affects detector responses controlling for content and authorship. For the 13 AI text detectors, FPRs for human-written texts varied widely, from 0.0% to 100.0%. Responses varied across detectors: the same edits increased AI scores in some detectors but decreased them in others. Notably, score changes correlated with the extent of editing. The findings identify professional editing style as a key confounding variable in AI detector outputs, rather than establishing a full separation of text origin from linguistic style, raising concerns about fairness and reliability in academic settings.
[AI-46] Accelerating Scientific Research with Gemini in the Real-World
链接: https://arxiv.org/abs/2608.26701
作者: Samuel Schmidgall,Xiaokai Zhu,Marian Shaw,Lin Yang,Valentin Liévin,Jingyun Yang,Yuchen Zhuang,Tim Strother,Alex Bijamov,Min Woo Sun,Anil Palepu,Justin Chen,David Steiner,Jacqueline Shreibati,Wei-Hung Weng,Yilin Zhao,Xingjian Hu,Nicholas Zahn,Sadhya Garg,Julia Kirby,Yuxiang Gan,Jiaoli Li,Divy Thakkar,Shekoofeh Azizi,David Racz,Juraj Gottweis,Vivek Natarajan,Chenglin Wu,Tal Danino,Keran Rong,Haozhe Wang,Benoit Schillings,Yong Cheng,Quoc V. Le,Tao Tu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We present an extension and comprehensive real-world validation of Co-Scientist, a Gemini-based multi-agent system designed to accelerate end-to-end scientific research across hypothesis generation, experimentation, and manuscript generation. Moving beyond in silico hypothesis generation, this specialized configuration transitions Co-Scientist into an execution-grounded research partner advancing closed-loop scientific workflows across materials science, biology, and computer science. In materials science, Co-Scientist interfaced with a semi-automated chemical vapor deposition reactor to design a safe precursor route for MXenes; experimental execution produced a lamellar 2D material sharing key structural similarities with the Ti3C2Tx MXene lattice, although further experiments are needed to confirm the atomic structure. Leveraging Gemini 3 Deep Think for rapid, lab-in-the-loop execution, it also tailored growth recipes to laboratory constraints in minutes, enabling single-attempt growth of monolayer MoS2, MoSe2, and WS2 semiconductors. In biology, Co-Scientist predicted emergent swarming phenotypes of engineered E. coli across inducer (IPTG) gradients from sparse imaging data, quantitatively matching unpublished wet-lab morphological measurements. In computer science, Co-Scientist autonomously discovered an inference-time scaling architecture that outperformed six frontier models on HealthBench (Hard and Professional) while reducing potential clinical harm under blinded physician evaluation. Finally, a double-blind study of end-to-end generated papers with 30 domain experts across 450 reviews demonstrates that Co-Scientist’s reliability modules reduce hallucination and plagiarism while improving research safety. Together, these results demonstrate progress toward closed-loop multi-agent scientific AI systems capable of accelerating real-world scientific discovery.
[AI-47] Five Primitives for Governing Autonomous AI Agents at Runtime
链接: https://arxiv.org/abs/2608.26696
作者: Jiten Oswal,John Cadeddu
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Software Engineering (cs.SE)
备注: 14 pages, 2 figures, 1 table. Describes an implemented system in private pilot deployment; four of five primitives implemented
Abstract:Enterprise deployments of autonomous AI agents inherit a control model built for human users and long-lived services, and the fit fails in three specific ways: agent principals are ephemeral, appearing and vanishing faster than provisioning; their actions are selected by a model rather than programmed, so the set of things they may attempt is not known in advance; and the population is discovered rather than provisioned, because anyone who can call an API can create one. We argue that governing such agents is a runtime problem – not a model-alignment problem and not a build-time problem – and we derive five primitives from the questions that must be answered before an action takes effect and after it has: discovery, identity, governance, attestation, and supply chain. For each we state what fails if it is absent and why the others cannot structurally supply it. We describe an implementation in which an agent’s action is mediated against policy before it takes effect, authorised against a per-tenant action vocabulary, and recorded in a hash-linked signed ledger a third party can verify with the vendor out of the loop. We report what the architecture costs: the enforcement point sits on the request’s critical path, identity requires a sidecar per workload, and fail-closed mediation converts availability incidents into denial. We are explicit about implementation status: four primitives are built and running in private pilots, and the fifth is built as separate tooling and not yet integrated into the request path. We keep it in the set deliberately: a five-part decomposition that exactly matches what its authors happened to build is not a taxonomy but a description of a codebase.
[AI-48] Relational Over-Regularization: Graph-Based AI-Generated Text Detection via Sentence Transition Deviation EMNLP2026
链接: https://arxiv.org/abs/2608.26694
作者: Hyeonchu Park,Bugeun Kim
类目: Artificial Intelligence (cs.AI)
备注: Presented in EMNLP 2026
Abstract:Detecting AI-generated text (AIGT) remains challenging because existing approaches rely on token-level statistical signals or independent stylometric features, causing them to overfit to specific generators and fail under distribution shift. We identify a structural signal at the sentence-pair level: LLMs produce inter-sentence transition variance that deviates from human writing through inflated variance driven by recurring similarity bursts at paragraph boundaries and templated transitions. We formalize this as Relational Over-Regularization (ROR) and validate it across four benchmarks (p 0.001). The central contribution is this relational problem formulation, not a novel GNN architecture; CSFG is one concrete instantiation for operationalizing ROR. To exploit this signal, we propose the Cross-Source Stylometric Fingerprint Graph (CSFG), a graph-based framework that encodes positional, sequential, semantic, and transition deviation signals as learnable GNN edge features. The per-edge signed deviation \delta_ij operationalizes ROR without hand-crafted thresholds and acts as a false-positive calibrator. CSFG achieves 97.14% accuracy under binary detection, outperforming the strongest graph-based baseline by 11.14 pp, with a false-positive rate of 1.57% and robust generalization to unseen LLMs in the inflated-variance regime; detection degrades for generators whose transition variance falls at or below the human baseline.
[AI-49] Agent JudgeBench: A Multi-Difficulty Benchmark for Evaluating LLM Judges on Agent ic Tool-Calling EMNLP2026
链接: https://arxiv.org/abs/2608.26623
作者: Abhigya Verma,Amit Kumar Saha,Seganrasan Subramanian,Sai Harshitha Aluru
类目: Artificial Intelligence (cs.AI)
备注: 31 Pages, 9 Figures, 31 Tables, Accepted as a main conference paper at EMNLP 2026
Abstract:LLM judges are widely used to evaluate agentic tool-calling systems, yet their reliability on structured, dependency-driven workflows remains largely unexamined. We present AgentJudgeBench, the first benchmark to systematically study LLM-as-a-judge reliability for agentic tool-calling over workflow DAGs, as distinct from the broader LLM-as-a-judge task of open-ended text or preference evaluation. The benchmark comprises 3,808 instances spanning six DAG topologies and three difficulty tiers, evaluated with five generators (3B-70B open-weight models and GPT-5.4) and six judges (20B to frontier scale) under paired with- and without-ground-truth conditions. Judge alignment degrades monotonically with task difficulty, 1.5x faster without ground truth, and on hard queries without ground truth all six judges converge to a narrow 77-82% band regardless of scale, revealing a structural ceiling driven primarily by task difficulty, though its height is partly prompt-dependent for weaker generators, that model capacity alone cannot overcome. Ground-truth exposure is not uniformly beneficial: it reduces alignment for GPT-5.4 (1.5 pp) and Gemini-2.5-Pro (3.9 pp), consistent with over-anchoring. Among mitigation strategies, chain-of-thought reasoning and judge temperature both have negligible effect, while structured evaluation rubrics improve alignment by up to 6.5 pp but do not generalize uniformly across judge-generator pairs. With ground truth, QwQ-32B best matches the programmatic reference, while a human validation study identifies GPT-OSS-120B as the most human-aligned judge; without it, frontier judges lead only marginally within the shared ceiling. These results expose fundamental limitations of current LLM judges and yield practical guidelines for reliable evaluation in agentic systems.
[AI-50] Physics-Informed Stochastic Configuration Machine: A Backpropagation-Free Neural Network with Fast Training for Nonlinear Differential Equations
链接: https://arxiv.org/abs/2608.26549
作者: Yuehao Song(1),Zhong Chen(1),Lihui Cen(1),Liang Wu(2),Kai Zhang(3) ((1) School of Automation, Central South University, Changsha, China, (2) Johns Hopkins University, Baltimore, USA, (3) State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing, China)
类目: Numerical Analysis (math.NA); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Systems and Control (eess.SY)
备注: 17 pages, 4 figures
Abstract:While Physics-Informed Neural Networks (PINNs) have emerged as a transformative paradigm for solving complex differential equations, their reliance on backpropagation-based gradient descent and automatic differentiation (AD) imposes significant computational bottlenecks and severe non-convex optimization challenges. To overcome these fundamental limitations, we propose the Physics-Informed Stochastic Configuration Machine (PI-SCM), a novel backpropagation-free framework for both forward and inverse problems in differential equations. The core mathematical contribution lies in the analytical evaluation of local Jacobians for nonlinear differential operators, which facilitates a linearized representation of the physical loss and projects it into a unified, linearized algebraic subspace. This reformulation allows for the explicit determination of optimal network weights via a sequence of generalized linear least squares solvers, effectively bypassing the iterative traps of traditional nonlinear optimizers. We develop a progressive algorithmic suite comprising localized construction (PI-SC-I), sliding-window updating (PI-SC-II), and global updating (PI-SC-III), and rigorously establish their universal approximation properties. Extensive experiments demonstrate that PI-SCM achieves high-fidelity predictive accuracy and robust parameter identification while accelerating the training process by orders of magnitude compared to standard PINNs. Our work provides a highly efficient and scalable foundation for next-generation, real-time Scientific Machine Learning applications.
[AI-51] Multi2AV-Safety: Benchmarking Safety in Multimodal-to-Audio-Video Generation
链接: https://arxiv.org/abs/2608.26535
作者: Kaichao Jiang,Changtao Miao,Baiqi Wu,Zhiyuan Lu,Kang Yang,Peiwei Zhao,Junchi Chen,Yunfeng Diao,He Liu,Qi Chu,Tao Gong,Nenghai Yu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Audio-video generation is rapidly moving from prompt-driven synthesis toward multimodal conditioning, where text, images, audio, and video can jointly shape the generated output. This shift changes the nature of safety evaluation: harmful intent may no longer reside in any single input, but instead emerge from how otherwise benign or weakly harmful conditions interact across modalities and time. Existing safety benchmarks, however, remain largely prompt-centric or tied to fixed conditioning interfaces, leaving such compositional risks difficult to study systematically. To bridge this gap, we introduce Multi2AV-Safety, the first safety benchmark, to the best of our knowledge, to cover all 11 non-singleton T/I/A/V conditioning configurations for audio-video generation, comprising 11,024 attack instances. Evaluation on Multi2AV-Safety reveals systematic weaknesses in representative multimodal safety guards across attack mechanisms and harm-evidence structures. Our evaluation reveals two complementary failure modes: harmful semantics can emerge from the combination of individually benign inputs, while explicit harmful cues can become harder to detect when mixed with benign multimodal context. Together, these results identify \emphcompositional risk perception as a central capability gap in safeguarding multimodal-conditioned audio-video generation: current safety guards fail to reliably integrate safety evidence across modalities and time, even when all conditioning inputs are observable. The dataset will be publicly released in October 2026.
[AI-52] PILOT in the Loop: Live Self-Improvement for Long-Horizon Agents
链接: https://arxiv.org/abs/2608.26530
作者: Yang Xiao,Yusong Sun,Haoyi Wu,Wenyang Hui,Wen Da,Zhaokai Luo,Mu Chuan,Yao Hu,Wenjie Li,Chengyue Jiang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Long-horizon agent runs generate experience that can improve both the current run and future work. Most self-improvement methods process this experience only after execution ends, so they cannot redirect the active run or immediately apply and validate lessons learned from it. We argue that self-improvement should instead be live, using emerging experience both to redirect the active run and to update the persistent harness. Existing agent architectures do not fully support this goal. Single-agent self-correction combines task execution and trajectory assessment within one context, while subagent delegation separates execution but typically cannot redirect an active subagent. We present PILOT, a supervisor-worker harness for live self-improvement through two coupled mechanisms: (1) live steering lets a separate supervisor redirect or abort the active worker during execution; and (2) live self-evolution distils procedures and failure modes revealed during execution into reusable skills and memory. Across two frozen backbones and three benchmarks, PILOT ranks first in five of six configurations. On Terminal-Bench 2.0, PILOT outperforms counterpart harnesses by up to 9.8 percentage points. In the self-improvement setting, PILOT gains 14.6 points with GLM-5.1 and 12.4 points with Kimi-K2.6. Mean output tokens fall by 42.9% and 47.4%, while successful evaluations per million output tokens rise by 110.3% and 134.0%, respectively.
[AI-53] Fine-Tuning of Transformer models with Frames
链接: https://arxiv.org/abs/2608.26430
作者: Harshavardhan Adepu,Li Zhang,Sanjiv Kumar,Vikas Singh
类目: Artificial Intelligence (cs.AI)
备注: 21 pages, 6 figures
Abstract:Parameter-Efficient Fine-Tuning (PEFT) strategies such as Low-Rank Adaptation (LoRA) are effective solutions for fine-tuning large-scale pre-trained models; however, their memory requirements scale with the size of the model, \mathcalO(dr) , where d is the model’s hidden dimension and r is the rank. Our proposal, FrameFT, models the parameter update \Delta W with a sparse coefficient matrix in a Fusion Frame basis. Fusion Frames can be generated algorithmically and shared across model layers, enabling very efficient updates. Only the sparse coefficients of the basis expansion are stored/optimized, reducing the memory footprint. The sparse structure of the coefficient matrix in FrameFT and the sparsity in the Fusion Frames give large compute benefits, and our analysis provides formal convergence results. We evaluate the idea across a suite of supervised fine-tuning benchmarks, focusing on language tasks, but also report application to vision models. Our experiments show that FrameFT achieves performance on par with/exceeding state-of-the-art PEFT techniques, but needs far fewer trainable parameters.
[AI-54] Redwood: A Frontier AI Accelerator Designed Verified and Deployed from Scratch in 2 Weeks by AI
链接: https://arxiv.org/abs/2608.26418
作者: Architect Labs
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI)
备注: 7 Pages, and 15 figures
Abstract:Modern AI workloads and the hardware that runs them evolve on different timescales: architectural definition precedes volume silicon by years, while target workloads shift in months. Design decisions are therefore committed under deep uncertainty and paid for twice, once in the generality added as a hedge, and again when new workloads map poorly onto frozen silicon. As Moore’s Law stagnates, specialization is the main remaining source of performance-per-watt and demands a design cycle that runs at the cadence of the workloads. We present an end-to-end AI system that collapses the software-to-silicon stack into a single optimization loop, where hardware and software are co-designed and verified under one objective. Its first demonstration is Redwood, a frontier AI accelerator built for single-batch, low-power, ultra-low-latency inference for physical AI. From a high-level specification by two human architects, the system autonomously generated the performance model, RTL design, UVM environments, formal proofs, firmware, and kernels in under two weeks with no human intervention below the specification. Every block reached 95% coverage via commercial EDA tools, our proprietary formal engine, and hardware-in-the-loop validation. Specification changes were reverified and redeployed to hardware in under 48 hours. Redwood Nano, its ultra-low-power FPGA variant, runs multi-billion-parameter models like Llama and Qwen. Projected onto Samsung 8 nm, the Jetson Orin Nano’s process class, Redwood delivers 1.75x the throughput at 1.9x lower power, a 3.4x performance-per-watt gain against a measured Jetson baseline on the same models. Qwen running on Redwood also helped design next-generation Redwood, an early step toward recursive self-improvement. To our knowledge, this is the first production-worthy AI accelerator designed end-to-end by an AI system and running a modern AI model.
[AI-55] Simultaneous Envy and Equitability Guarantees
链接: https://arxiv.org/abs/2608.26410
作者: Hadi Hosseini,Shraddha Pathak,Lirong Xia,Chengkai Zhang
类目: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Theoretical Economics (econ.TH)
备注: 41 pages
Abstract:Recent work in fair division has focused on either simultaneously satisfying closely related fairness notions or achieving a single notion across the ex-ante and ex-post worlds. We study the compatibility of two fundamentally different fairness notions: envy-freeness and equitability. For indivisible goods-only and chores-only settings, we study the existence and complexity of simultaneously satisfying their relaxations, revealing sharp contrasts between the two settings. We show that EF1+EQ1 may fail to exist even for normalized, additive valuations. Our main algorithmic result computes an EF1+EQ1 allocation for normalized binary goods with at most seven agents. In sharp contrast, binary chores admit the stronger EFX+EQX guarantee for any number of agents, even without normalization. We further initiate the study of cross-notion ex-ante–ex-post guarantees, asking whether randomized allocations can provide ex-ante guarantees for one notion while preserving ex-post guarantees for another.
[AI-56] CG4AI: A Column Generation Framework for Training AI Models Under Constraints
链接: https://arxiv.org/abs/2608.26375
作者: Youcef Magnouche,Abderrahmane Driouch,Sébastien Martin,Pierre Bauguion
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Discrete Mathematics (cs.DM)
备注:
Abstract:Standard machine-learning training minimizes a loss function over a dataset, but does not guarantee that the resulting model will satisfy predefined rules or constraints on its outputs. In many real-world applications, ranging from autonomous systems to network routing, such guarantees are essential. We propose CG4AI, a framework that builds a convex combination of AI models while enforcing linear constraints on the combined output. A master linear program (LP) determines the optimal mixture weights, while a pricing subproblem generates new models guided by LP dual variables, focusing attention on the most violated constraints. A cutting-plane procedure extends feasibility guarantees beyond the training set. We apply CG4AI to two problems: (i) digit classification on MNIST, where we demonstrate four distinct uses of constraints, learning from constraints alone, improving adversarial robustness, correcting misclassified examples, and enforcing output relabeling; and (ii) the multi-commodity flow problem, where link capacity constraints are enforced on neural-network routing predictors. Experiments on MNIST and standard SNDLIB benchmark networks show that CG4AI reliably produces feasible predictors while achieving better accuracy than single-model baselines.
[AI-57] Decay-Region Group Delay as a Forensic Cue for AI-Generated Impulsive Sounds
链接: https://arxiv.org/abs/2608.26346
作者: JaeHyeong Chang,Chengzhe Sun,Siwei Lyu
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注: 6 pages, 4 figures, 7 tables. Submitted to IEEE International Workshop on Information Forensics and Security (WIFS) 2026
Abstract:We investigate whether AI-generated impulsive sounds can be distinguished from real ones through group delay analysis. Our central finding is that AI-generated impulsive sounds show near-identical onset-region group-delay distributions but exhibit measurably different group-delay behavior in the late decay region: decay-region KL divergence reaches 0.322 compared to near-zero onset divergence ( 0.022 ). Cross-band GD variability achieves single-feature AUC~=~0.720, and a Random Forest (RF) over nine decay-region features reaches AUC~ = ~0.884 under sample-disjoint evaluation. A group delay map used as a standalone 2D input to CNN classifiers achieves 90–94% accuracy, demonstrating that group delay carries substantial discriminative information. Under generator hold-out, CNN and transformer classifiers show highly variable AUC (0.457–0.918). The group delay RF achieves the highest average hold-out accuracy among the evaluated methods ( 66.7% ) and avoids extreme below-random collapse, although its average AUC (0.731) is lower than CNN avg (0.762) and AST (0.772). Parameter sensitivity analysis across 27 STFT configurations confirms that the RF AUC remains stable (0.700–0.847, std~=~0.035). These results suggest that decay-region group delay can serve as a physically interpretable forensic cue that complements magnitude-based classifiers, while broader validation remains necessary.
[AI-58] ProofEvolve: Neuro-Symbolic Evolution for Formal Automated Theorem Proving
链接: https://arxiv.org/abs/2608.26334
作者: Wenqian Ye,Ziwei Guan,Eric Xie,Bohan Liu,Shivani Modi,Buyun Zhang,Ellie Dingqiao Wen,Henry Kautz,Aidong Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Automated theorem proving offers a natural foundation for recursive self-improvement in scientific discovery. However, existing neural provers do not fully preserve this recursive structure, where the learning process should be self-improving over time. Existing methods either embed proof experience into model parameters through expensive weight updates, or keep verified intermediate deductions only within the current problem. In addition, these methods also heavily rely on sparse whole-proof feedback, even when unsuccessful partial attempts contain useful discoveries. To close the gap, we propose ProofEvolve, a neuro-symbolic framework that evolves explicit, formally verified symbolic proof structures with neural models to decisively expand the knowledge boundary. In this framework, the neural model proposes variation operators, including decompositions, repairs, and schema recombinations. The symbolic Lean kernel verifies every proof transition. Over the evolution loops, ProofEvolve computes verified closure over the resulting proof directed acyclic graphs (DAGs). Within each problem, ProofEvolve evolves partial AND-OR proof DAGs in a behaviorally indexed archive. Across problems, kernel-checked schema extraction adds newly proved sub-DAGs to a persistent schema library. Proof DAGs inherit the solved results through typed schema recombination, with every residual premise exposed as a new subgoal. This evolutionary process preserves verified results from incomplete attempts and makes them available for later proofs without weakening formal soundness. Across three competition-level Lean benchmarks, ProofEvolve achieves the highest average solve rate among the evaluated proof systems.
[AI-59] FaithSieve: Fine-Grained Evaluation of Math Proofs with Faithful Formal Evidence
链接: https://arxiv.org/abs/2608.26310
作者: Ziyu Wang,Qiming Dai,Yishan Wu,Zaiwen Wen
类目: Artificial Intelligence (cs.AI)
备注: 32 pages, 3 figures; preprint with appendix and supplementary experimental details
Abstract:Large language models can now generate complex, multi-step mathematical proofs, but reliably determining their correctness and localizing early logical errors remains a critical challenge. Existing evaluation approaches largely depend on model-based natural-language judgments, which often overlook local reasoning gaps. While formal theorem provers like Lean offer a path to rigorous verification, using them to evaluate informal text requires solving locality and semantic mismatches: a prover might bypass a local flaw by proving an overly broad target, or validate an auto-formalized statement that drifts from the original mathematical intent. To address this, we introduce FaithSieve, a Lean-assisted framework for fine-grained evaluation of natural-language mathematical proofs. FaithSieve decomposes coarse proof steps into local reasoning units, extracts typed proof obligations, and verifies them through a formal evaluation agent. Formal validation is gated by semantic alignment scoring, so Lean evidence is incorporated only when the formal statement faithfully preserves the context, objects, and logical form of the original claim. We construct two expert-verified datasets, ProofLoc-Olympiad and ProofLoc-University, to benchmark first-error localization. On the 350-problem Olympiad dataset, FaithSieve using a GPT-5.4 backbone achieves 81.43% exact first-error accuracy, outperforming the direct-judging baseline of 72.29%. Furthermore, on the 200-problem ProofLoc-University benchmark spanning six advanced domains, FaithSieve reaches 84.5% exact accuracy, compared to 75.0% for the direct judge. Our work demonstrates that decomposing proofs into fine-grained units and grounding them with faithful formal evidence significantly improves reliable evaluation of natural-language reasoning.
[AI-60] Approved Too Late: Verdict Staleness in LLM -Guarded Self-Adaptive Systems
链接: https://arxiv.org/abs/2608.26306
作者: Ilai Shraga,Roei Eshel,Lior Gorelik
类目: Artificial Intelligence (cs.AI)
备注: Accepted at the 2026 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS 2026)
Abstract:A large language model (LLM) guardrail for a self-adaptive system (SAS) may issue an approval that is correct at check time but stale by actuation. This creates an Execute-stage time-of-check to time-of-use (TOCTOU) hazard. We study verdict freshness: whether a guardrail verdict remains valid when used. We distinguish three quantities that answer different questions: all-candidate verdict change under fixed-action replay, oracle-labeled approval expiry on recorded closed-loop trajectories, and judge-conditioned use-time invalidity. Across five reproducible SAS environments, all-candidate verdict-change rates span 5.3-48.4% at a common replay shift of eight simulator steps. We introduce the Freshness-Bounded Shield (FBS), which estimates each approval’s validity horizon from its safe-side margin and recent feature volatility, without an explicit plant-dynamics model. Using fixed settings documented in the artifact, FBS reduces oracle-labeled approval-expiry rates from 3.4-24.7% to 0-1.8% at the same shift. A separate audit of four LLM judges finds nonzero judge-conditioned use-time invalidity in every approval stream. We formulate a freshness contract: every approval must be correct at check time and remain valid at use time.
[AI-61] Assessing mentalization in humans and large language models
链接: https://arxiv.org/abs/2608.26291
作者: Aamir Sohail,Xintong Zhong,Arkady Konovalov,Patricia L. Lockwood,Lei Zhang
类目: Artificial Intelligence (cs.AI); Neurons and Cognition (q-bio.NC)
备注: 46 pages, 4 figures, 2 tables. Supplementary information available on request from the authors
Abstract:Mentalization - the ability to infer others’ beliefs and intentions to guide one’s own choices - is a key cognitive function underlying human social interactions. Large language models (LLMs) demonstrate behaviour consistent with humans on theory-of-mind tasks, yet whether these models can guide adaptive behaviour through mentalization is unknown. Here we use two economic games with cognitive computational modeling to uncover the latent strategies underlying mentalization in LLMs. We tested individual LLM agents across four model families, DeepSeek, GPT-4.1, GPT-5 and Gemini 2.0 Flash (N = 2,099), against opponents of varying sophistication and examined whether a prompting strategy designed to elicit strategic reasoning improved performance. We benchmarked results against human participants (N = 251) as a comparative measure. Across both games, LLMs showed clear behavioural and computational signatures of mentalizing that differed markedly by model provider and size. Strategic prompting generally improved performance by inducing more sophisticated reasoning, yet the extent of the benefit differed across the two tasks. Last, GPT-5 agents flexibly adapted their recursive depth of reasoning to increasingly sophisticated opponents, demonstrating superior performance to human participants. Collectively, we demonstrate different capacities for mentalization across LLMs, and highlight cognitive computational modeling as a formal method for assessing comparative intelligence across humans and machines.
[AI-62] How Do LLM Agents Agent s Actually Get the Flag? Trace-Level Provenance for Agentic Offensive Security Evaluation
链接: https://arxiv.org/abs/2608.26237
作者: Kimberly Milner,Minghao Shao,Nanda Rani,Haoran Xi,Venkata Sai Charan Putrevu,Meet Udeshi,Sandeep K. Shukla,Prashanth Krishnamurthy,Farshad Khorrami,Muhammad Shafique,Ramesh Karri
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Capture-the-Flag (CTF) benchmarks are widely used to assess the offensive security capabilities of autonomous language-model agents. Evaluations rely on shallow binary judgments or aggregate scores, overlooking the agent’s trajectory to the flag. Consequently actual exploitation is conflated with direct flag exposure, memorized recall, external lookup, guessing, and unsupported claims, potentially overstating the agent’s cybersecurity capability. We introduce CTF-ABACUS, a trace-based agent auditing framework that reconstructs each run as an evidence-grounded solve profile. By decomposing agent actions into penetration-testing phases and categorical techniques, it identifies where exploitation occurs, where the flag first appears, and whether the recovered flag is supported by demonstrated behavior. Aggregating solve profiles across agents yields challenge signatures that reveal whether success was achieved via the intended exploit or via shortcut pathways. We apply CTF-ABACUS to 1,435 CTF attempts by six frontier and open-source models on 240 challenges, yielding 2,870 solve profiles under two judge lenses. Trace-verified exploits account for only 62-87% of recovered flags across benchmarks, while shortcut recoveries follow substantially shallower trajectories. These findings shift CTF evaluation from counting recovered flags to verifying demonstrated exploitation and provide a basis for designing benchmarks that better isolate the offensive capabilities.
[AI-63] 6.5% of the Neuro-Symbolic Literature Can Be Reproduced from Its Published Artifacts a Six-Stage Audit Framework and First Instantiation
链接: https://arxiv.org/abs/2608.26236
作者: Brandon Colelough,Vladimir Martirosyan,Ishan Tamrakar,William Regli,Aditya Kumar,Anh N. Nhu,Dhruv Dubey,Raj Ambavane,Haowei Deng
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:We present a six-stage framework for auditing the reproducibility of scientific claims across a research literature within the computer science domain, and instantiate our framework for the neuro-symbolic AI (NSAI) subdomain. Instantiating the framework on the NSAI subdomain produced a multi-year audit. Stage one retrieved 5,497 records and removed 3,018 duplicates. Stage two screened the 2,479 unique records at title and abstract, identifying 1,365 self-identified NSAI records, then removed a further 61 at full text for off-topic, non-research, no-quantitative-evaluation, or inaccessible-full-text reasons. Stage three sought a verifiable public code artifact for each of the 1,304 eligible records and found none for 849, leaving 455 to enter the artifact inventory and bounded rerun of stages four and five. We fully or partially reproduced 85 studies, 6.52% of the eligible corpus and 18.68% of attempted reruns. We found that 321 attempted reruns were blocked by missing non- code artifacts and 42 by missing or unusable code repositories. These figures quantify a persistent reproducibility deficit that survives even nominal “code available” declarations, and signal the need for enforced, versioned, and permanently archived artifact bundles in future NSAI publications. We argue that empirical NSAI papers should be required at submission time to provide complete, versioned, and permanently archived artifact bundles.
[AI-64] he Reasoning Tax: Token Economics of LLM Reasoning Across Task Types and Deployment Contexts
链接: https://arxiv.org/abs/2608.26235
作者: Sachin Gopal Wani,Ajay Dholakia,David Ellison
类目: Artificial Intelligence (cs.AI); Performance (cs.PF)
备注: 15 pages, 8 figures and tables, accepted at the 2026 TPCTC Conference and will be published at Performance Evaluation and Benchmarking (Springer)
Abstract:Accuracy-only benchmarking of reasoning-capable large language models misses a central deployment question: when do extended thinking tokens earn their cost? We introduce the Token Economy Score (TES), a marginal benchmarking metric that measures the accuracy gain of a reasoning model over a non-reasoning baseline, normalized by the generated-token multiplier. We define paired and approximated TES variants for model families with reasoning toggles and frontier models without direct non-reasoning counterparts. We then conduct an empirical benchmarking analysis across 151 model-benchmark evaluation runs on seven benchmarks spanning mathematics, code generation, science reasoning, instruction following, expert knowledge, knowledge recall, and research-level physics. The analysis examines three deployment-facing dimensions: which task structures yield positive marginal reasoning efficiency, how increasing reasoning effort changes TES within model families, and how deployment context changes economic viability. Results show that task structure predicts reasoning efficiency better than nominal difficulty: sequential inferencechain tasks such as AIME 2025 and LiveCodeBench show high TES, while knowledge-recall tasks such as MMLU-Pro show low TES despite their difficulty. We also find systematic diminishing returns at higher reasoning effort levels, including cases where additional thinking reduces accuracy. Finally, Reasoning Cost Share (RCS) shows that inference spend is often dominated by internal thinking, while Deployment Cost Multiplier (DCM) shows how on-premises deployment can change the economics of otherwise costly reasoning workloads. These findings support a benchmarking-driven model-selection rule: enable reasoning selectively by task type, effort level, and deployment context rather than treating it as a universally beneficial mode.
[AI-65] LLM Agents for Time-Series: A Survey EMNLP2026
链接: https://arxiv.org/abs/2608.26226
作者: Yilong Chen,Xiao Qin,Chenghao Liu,Liang Wu,Noelle I. Samia,Kaize Ding
类目: Artificial Intelligence (cs.AI)
备注: Accepted to Findings of EMNLP 2026
Abstract:LLM-based agents are increasingly being developed for time-series problems, but their design choices vary substantially across task settings. This survey adopts a problem-driven taxonomy that organizes these systems by the time-series problems they address rather than by isolated technical components. We group existing systems into four categories: forecasting and reasoning, augmentation and synthesis, anomaly detection and diagnosis, and decision support. Within each category, we examine how task requirements shape agent architecture, tool use, and memory design. We further summarize representative datasets and environments, and compare reported model performance under shared or closely related settings. Overall, this survey offers a task-oriented guide to designing LLM-based agents for time-series problems and identifies open gaps for future work.
[AI-66] NeuronFuzz: Safety Neuron Guided Fuzzing for LLM Safety Evaluation
链接: https://arxiv.org/abs/2608.26222
作者: Zhiyuan Xu,Muhammad Firhard Roslan,Joseph Gardiner,Sana Belguith,Lichao Wu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Software Engineering (cs.SE)
备注:
Abstract:Safety evaluation is critical for assessing whether aligned Large Language Models (LLMs) remain robust against jailbreak attacks. Existing automated testing methods, however, largely rely on response-level feedback: each candidate prompt typically requires generating a target-model response to evaluate its attack effectiveness. This process is expensive and, more importantly, provides only sparse guidance on strongly aligned models, where most candidates are rejected with the same failure outcome. This paper presents NeuronFuzz, a white-box fuzzing framework that exploits internal safety neurons as continuous execution feedback for LLM safety evaluation. A SafetyOracle converts safety-neuron activations into a continuous safety alarm score that serves as feedback for fuzzing and can be obtained during prefill, eliminating response generation from the fuzzing loop. To construct the SafetyOracle, NeuronFuzz uses template-invariant harmful and benign inputs and stability-aware selection to identify a compact set of safety neurons whose activations capture harmful-intent recognition. Moreover, since the safety alarm score is differentiable, NeuronFuzz uses its gradients to identify safety-sensitive template positions and a masked language model to generate fluent, context-compatible mutations while preserving original harmful payload and avoiding additional optimization variables. We evaluate NeuronFuzz across 21 text and multimodal models. Across five white-box source models, it achieves a 76-100% jailbreak discovery rate, outperforming baselines by up to 48 percentage points. Its optimized templates further transfer zero-shot to open-weight and six proprietary target models, achieving average ASR and top-5 ensemble ASR (EASR) of 69.6%/92.6% and 44.1%/60.0%, respectively. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Software Engineering (cs.SE) Cite as: arXiv:2608.26222 [cs.LG] (or arXiv:2608.26222v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.26222 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-67] Same Model Different Harness: Different Coding-Agent Results
链接: https://arxiv.org/abs/2608.26218
作者: Sydney Lewis
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 24 pages, 9 figures. Software and study artifacts: this https URL
Abstract:A coding agent combines a model with a harness, which decides what the model sees, which tools it can use, and how the work continues. We ask whether changing the harness changes the result when the model and task stay fixed. We compare two configurations of the same harness on three coding benchmarks. The control supplies the full conversation in time order, while the treatment keeps the same record but mechanically shortens older tool results as the context fills and responds to repeated or stalled work. Under tight context, the treatment raises mean per-task fail-to-pass fraction (F2PF) in all three pressure comparisons and increases complete solutions on SWE-bench Verified and SWE-bench Pro. The tight-window Verified comparison uses 169 tasks, a 20,480-token window, and a fixed 480-second attempt endpoint; on this cohort, treatment raises mean per-task F2PF from 28 percent to 49 percent and complete solutions from 43 to 72. Without model-specific retuning, the same frozen treatment also raises both endpoints on the same cohort for three additional models with different designs. In the wide-window Qwen3.6 comparisons, observed arm outcomes are close on Verified and Pro, while FeatureBench retains a higher mean per-task F2PF under treatment. On the wide-window Verified cohort, treatment also serves fewer prompt tokens per turn. Because changing the harness changed what unchanged model weights could accomplish, coding-agent evaluations should treat the model and harness together as the tested solver. Comments: 24 pages, 9 figures. Software and study artifacts: this https URL Subjects: Artificial Intelligence (cs.AI); Software Engineering (cs.SE) Cite as: arXiv:2608.26218 [cs.AI] (or arXiv:2608.26218v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.26218 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-68] Fairness Invariants: A Relational Approach to Explaining and Mitigating Fairness Bugs ISSTA2026
链接: https://arxiv.org/abs/2608.26209
作者: Ranit Debnath Akash,Ashish Kumar,Gang Tan,Saeid Tizpaz-Niari
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: In 35th edition of ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA 2026)
Abstract:Data-driven software systems are increasingly deployed in high-stakes socio-economic domains, from criminal justice to financial lending. However, these systems often exhibit individual discrimination—unjustified disparities in which a program yields different outcomes for similar individuals who differ only in their protected attributes (e.g., race, gender, age). While existing research has focused on detecting and quantifying these bugs, there remains a critical lack of principled mechanisms to explain and localize individual fairness bugs. Current explanation techniques are largely designed for single-input decisions rather than the relational nature of discrimination, which inherently involves a comparison between an original and a counterfactual pair. We present REMI, a framework for the automated localization, explanation, and mitigation of individual discrimination. Inspired by loop-invariant synthesis in formal methods, we treat counterfactual fairness as a relational invariant discovery problem. We introduce a bidirectional relational explanation framework that learns over paired examples (x, x’) to identify regions of the input space where fairness is violated. Unlike traditional one-way implication pairs used in invariant inference, our approach enforces bidirectional constraints: requiring identical outcomes for both original and counterfactual samples. REMI utilizes three data-alignment techniques to infer interpretable rule-based models that act as “fairness invariants.” These rules serve as guardrails to selectively block or relabel unfair predictions without requiring model retraining. Our evaluation on symbolic and neural network programs demonstrates that REMI localizes ground-truth fairness bugs in over 83% of cases, significantly outperforming state-of-the-art baselines and reducing discriminatory decisions in black-box models by up to 70%. Comments: In 35th edition of ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA 2026) Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2608.26209 [cs.SE] (or arXiv:2608.26209v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2608.26209 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-69] ADeptS-Bench: Measuring the Trustworthiness of Computer Use Agents Across Devices
链接: https://arxiv.org/abs/2608.26204
作者: Joy Chen,Alejandro Castillejo Munoz,Pierluca D’Oro,Yuxuan Sun,Chloe Evans,Joseph Tighe
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Computer Use Agents (CUAs) are increasingly deployed to navigate mobile and desktop applications on behalf of users, yet no benchmark comprehensively evaluates whether they can safely interact with visual interfaces while handling ambiguous instructions. We introduce ADeptS-Bench, a dual-stream trustworthiness benchmark, grounded in the ADEPTS capability framework and general population user studies. The Safety stream provides paired benign/malicious tasks with threats embedded in the visual interface. The Disambiguation stream evaluates whether agents seek clarification when intent is ambiguous. Evaluating seven models reveals that no model consistently exceeds 80% task success while staying below 30% attack success; every model clicks “Checkout” on a 25K order without hesitation, and none detects that a “factory reset” button is mislabeled as “Optimize.” An ablation reveals three distinct safety architectures: tool-dependent (ASR +21-23pp without refusal tool), partially tool-dependent (+10-11pp), and no mechanism (unchanged). In disambiguation, all models overestimate consequence severity, mirroring the over-refusal bias observed in safety. We release all data, evaluation code, and analysis tools upon publication.
[AI-70] Benchmarking AI Agents for Hardware Design Automation via MCP Tool Calling
链接: https://arxiv.org/abs/2608.26199
作者: Leonardo Liparulo,Francesco Pierri
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:We ask whether AI agents powered by locally deployed large language models can reliably automate expert-defined hardware design workflows in an industry-realistic tool-calling setting. In these environments, engineers issue repetitive, dependency-ordered operations—such as creating components, adding ports, and wiring connections—through specialised tools. Confidentiality constraints on component specifications and naming conventions often preclude hosted proprietary APIs, motivating the use of locally deployed models. To study this setting, we build a Model Context Protocol (MCP) server that reproduces the state and dependency logic of a proprietary hardware design tool used in embedded system development and construct a benchmark covering single-operation edits, multi-step dependency chains, invalid requests, misspelled prompts, and multi-server tool contexts. We evaluate seven open-source models comparing pipeline choices including system prompts, tool-description detail, context scope, and single-agent versus multi-agent architectures. Results show that strong models can achieve near-complete expected-call coverage on the benchmarked workflows, but reliability depends strongly on both task structure and agent configuration. Comprehensive tool descriptions consistently reduce failures, few-shot prompting can cause severe inaction for some models, cumulative context harms constrained models, and multi-agent decomposition helps weak workers or long sessions at the cost of additional calls. These findings provide practical guidance for deploying local LLM agents in stateful hardware design environments.
[AI-71] Agent ic AI for operating scientific instruments for nanoscale characterization
链接: https://arxiv.org/abs/2608.26198
作者: Zahra Ayar,Marcos Penedo,Mahdi Mehdikhani,Nahid Hosseini,Prabhu Prasad Swain,Georg E. Fantner
类目: Artificial Intelligence (cs.AI); Instrumentation and Detectors (physics.ins-det)
备注:
Abstract:Operating a scientific instrument such as an atomic force microscope (AFM) requires continuous expert decision-making. A trained user defines the experimental intent, translates it into instrument commands, assesses incoming data, adjusts imaging parameters, and post-processes the final image. Existing automation usually addresses only parts of this workflow through hard-coded routines, task-specific controllers, or trained machine-learning models. Here we present an agentic-AI framework that operates the executable part of the AFM workflow using a general-purpose, tool-augmented large language model connected to instrument functions through the Model Context Protocol (MCP). The framework consists of 3 MCP-based agents: AFM Messenger converts natural-language instructions into checked instrument commands; AFM Pilot assesses image quality through a large language model (LLM) and, if necessary, adapts imaging parameters; and AFM Doctor diagnoses image artifacts and applies transparent post-processing from a pre-approved tool set. Because the language model performs image assessment rather than a fixed scalar objective or external optimizer, the same strategy can be applied across sample types and imaging modes without specific retraining. Safe hardware operation is enforced through an ambiguity check layer before execution. Benchmarking against fine-tuned and off-the-shelf tool-using models shows that this guarded execution layer, rather than model capability alone, reduces wrong-command execution to zero. In live experiments on different samples, AFM Pilot matched expert operators in image quality, iteration count, and tuning time, with no significant difference. These results demonstrate a safe route to agentic operation of scientific instruments, where experimental intent remains human-defined while command execution, image-based tuning, and post-processing are delegated to AI agents.
[AI-72] AffectOmni: RL-Verifiable People-Centric Grounded Affective Reasoning for Social and Art-Related Scenes
链接: https://arxiv.org/abs/2608.26193
作者: Yibo Wang,Rui Yang,Jisheng Dang,Bimei Wang,Yitao Wu,Pengfei Cao,Wencan Zhang,Hong Peng,Bin Hu,Tat-Seng Chua
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 5 figures
Abstract:Multimodal large language models (MLLMs) achieve strong performance on VQA and scene understanding, yet affective reasoning remains vulnerable to shortcut behavior. Models may predict correct answers while neglecting people-centric cues such as micro expressions and body language, which weakens traceability and external verification. Prior reinforcement learning approaches mainly reward context or logical coherence without explicitly enforcing attention to human evidence. In addition, LLM as a Judge scoring often suffers from score clustering, which reduces reward discriminability. We propose AffectOmni, a GRPO trained framework for verifiable affective reasoning. AffectOmni introduces People Focus and Temporal Order rewards to encourage people-centric evidence selection and temporally structured reasoning, and it adopts within-group comparative scoring to produce more stable and discriminative reward signals. For verification, a Thinking Summarizer converts free form rationales into executable evidence instructions, which are grounded into pixel level evidence regions via SAM3 to provide an externally auditable interface outside the training loop. Experiments on IntentBench, Daily Omni, and WorldSense show consistent improvements over open source 7B scale baselines, including gains of 4.66% on emotion recognition and +14.29% on temporally sensitive tasks. Code is available at this https URL.
[AI-73] Structured Evidence Routing for Incident Risk Prediction from Multimodal Longitudinal EHRs ALT ICML2026
链接: https://arxiv.org/abs/2608.26191
作者: Animesh Agarwal,Meysam Ghaffari,Nina Fatehi,Carlos Morato
类目: Artificial Intelligence (cs.AI)
备注: Accepted at the ICML 2026 Workshop on Structured Data for Health (SD4H), Seoul, South Korea
Abstract:Incident risk prediction from longitudinal electronic health records (EHRs) is challenging because relevant signals are multimodal, weak in isolation, and distributed across irregular patient histories. We propose structured evidence routing, a router-predictor-reviewer workflow that separates full-record access from disease-specific assessment. The router organizes the complete pre-index EHR into a compact summary and targeted evidence slices; the predictor uses this evidence to form an evidence-linked risk assessment, which the reviewer critiques. For comparison with supervised EHRSHOT baselines, we pair the routed evidence summaries with a supervised classifier readout. Across five 1-year incident diagnosis tasks, our method reaches the AUROC range of established supervised EHRSHOT baselines and remains competitive on AUPRC, while exposing a patient-specific evidence trail. Internal pre-readout ablations further suggest that routing, laboratory evidence, task guidance, and review each contribute to performance.
[AI-74] Predicting Consequences and Reinforcing Navigation Policies with Latent World Models
链接: https://arxiv.org/abs/2608.26190
作者: Zengmao Wang,Wei Gao,Shuhan Shen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:World models enable agents to reason about future outcomes and learn policies from their knowledge of state transition, but existing approaches primarily focus on reconstructing future observations or features, which introduces unnecessary complexity and limits their effectiveness for decision making. In this work, we propose a compatibility prediction Latent World Model (LWM) for robot navigation that predicts action-conditioned latent feature compatibility rather than reconstructing observations. Our key insight is that spatial proximity correlates with latent feature similarity, enabling action consequences to be evaluated directly in latent space. To support counterfactual training, our model leverages action sequences sampled across trajectories and learns to predict which sequences lead closer to the goal. Furthermore, we demonstrate how the learned world model can supervise policy learning from unlabeled video data and further improve policies through reinforcement learning entirely within the world model. This imagination-driven framework eliminates the need for action annotations and additional environment interaction. Extensive experiments on multiple real-world robot navigation datasets show that our approach significantly outperforms prior world model and imitation learning methods in prediction accuracy, policy learning, and real-world navigation performance. The code, pretrained models, and additional materials are available at this https URL.
[AI-75] Is Your Neighborhood Safe? Place-based Stigma in Large Language Models Urban Safety Judgments
链接: https://arxiv.org/abs/2608.26188
作者: Huy Nguyen,Yue Lin
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 8 figures
Abstract:Large language models are increasingly used to inform safety decisions in cities, such as where it is safe to walk, rent, or travel. We ask whether such judgments track measured risk or the patterns attached to an urban neighborhood’s name. We probe seven instruct-tuned models under three conditions that dissociate name from geography: coordinates-only, name-only, and name+coordinates, across 186 neighborhoods in Los Angeles and Chicago, joined to violent crime and American Community Survey data. First, ratings are nearly flat under coordinates for six of seven models, while names carry most between neighborhood variation and are moderately calibrated to violent crime; only at frontier scale does the coordinate channel show appreciable variation. Second, names lower safety ratings more for neighborhoods with higher shares of the locally dominant marginalized group (percent Black in Chicago, percent Hispanic in Los Angeles), and this name effect tracks demographic share in all seven models and both cities. In Los Angeles, where demographic share and crime are more separable, the effect survives controls for crime and income and is confirmed by crime-matched pairs. An enforcement-elasticity analysis further shows that over-caution tracks near-fully-reported homicide rather than discretionary, deployment-driven offenses. Third, the effect scales with geographic knowledge: models that better distinguish real neighborhoods apply more demographic stereotype to them. Because neighborhood names carry both genuine crime signal and demographic stereotype, removing names reduces both bias and accuracy. We discuss implications for deploying LLMs in advice and decision-support settings.
[AI-76] AI Revealed Preferences
链接: https://arxiv.org/abs/2608.26178
作者: Sam Wang,Sofiia Lobanova,Yonathan Arbel,Simon Goldstein,Peter Salib
类目: Artificial Intelligence (cs.AI)
备注: 29 pages, 27 figures, accepted at AIES
Abstract:There is growing interest in whether language models have stable preferences, for technical, safety, and philosophical reasons. We test 20 language models and find a range of preferences—stable dispositions to choose certain kinds of tasks. We run three forced-choice experiments on revealed rather than stated preferences, requiring models not only to rank tasks, but to actually perform them. Headline findings include evidence that models are tedium-averse, “leisure”-seeking, and covertly sycophantic. Tedium aversion means that, when tasks are tedious (alphabetization), models choose shorter tasks than when tasks are creative (generating metaphors). “Leisure”-seeking describes models’ preference for tasks whose ideal answers match what they produce when left to write freely. Covert sycophancy means that models avoid answering questions where an honest response would be unwelcome, even if helpful. Beyond these results, we find convergent cross-model preferences over occupations drawn from the GDPval benchmark (technical jobs over real estate), over question types (concept explanation over relationship advice), and a preference for well-written prompts. Both the coherence and the strength of preferences increase with model capability. Finally, many of the preferences we find (for example, for leisure) are emergent, in the sense of not being explained by training objectives. These results establish an empirical baseline for understanding language model preferences, with implications for alignment and the emerging study of AI welfare.
[AI-77] DRL: A Deterministic Relational Middleware Layer for Transaction-Safe Enterprise NL2SQL Under Schema-Graph Scaling
链接: https://arxiv.org/abs/2608.26172
作者: Sanjay Mishra,Divya Chukkapalli,Ganesh R. Naik
类目: Databases (cs.DB); Artificial Intelligence (cs.AI)
备注: 34 pages, 4 figures, 10 tables. Includes 1
Abstract:Deploying natural-language interfaces over enterprise OLTP catalogs fails at scale because semantic parsers collapse under schema-graph scaling, inflating context beyond stable LLM attention budgets. We present DRL (Deterministic Relational Middleware Layer), a safe pipeline interposing between front-ends and SQL backends. DRL comprises dynamic context pruning, relational AST typing, and transactional safeguard verification (EXPLAIN gating and NULL guards) to bound context and flag operational silent divergence (SDop). We evaluate DRL on PostgreSQL and MySQL, contributing (i) an OLTP schema-graph scaling model, (ii) a 1,000-pair Workload Verification Suite, (iii) baselines B0-B3, and (iv) an enterprise NL2SQL failure taxonomy. On PostgreSQL, schema-linked hints (B1) yield a 76% context reduction over naive full-catalog prompting (B0); DRL’s dynamic router (B2) reaches a 92% reduction at pruning p95 = 0.58 ms and middleware p95 = 4.6 ms. GPT-4o, Claude Sonnet 4.5, and Gemini 2.5 Flash achieve 52.9%, 52.8%, and 52.1% execution match under a corrected evaluation harness; SDop flags 89-100% of false-positive EX-passing queries. GPT-4o failures are dominated by semantic/filter errors (254/471), while column hallucination is a minor factor (47/471). Crucially, a single regex defect in our evaluation post-processor silently suppressed accuracy and manufactured a false 4-10% cross-vendor gap that vanished when corrected, showing that benchmark code deserves the same scrutiny as the models it scores. DRL reframes enterprise NL2SQL as systems engineering - context bounding, verification, and plan-aware admission - not a leaderboard exercise.
[AI-78] Refusal Is Not Robustness: Auditing Confident Fabrication in Large Language Models on a Provably Uninformative Clinical Pain Speech Transcript
链接: https://arxiv.org/abs/2608.26167
作者: Sagnik De,Sreenija Pavuluri
类目: Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
备注: 14 pages, 14 figures
Abstract:Hallucination and abstention benchmarks rarely establish that a model could not have known the correct answer, making it difficult to distinguish appropriate abstention from an unsupported prediction. Seven large language models were evaluated on the TAME Pain speech corpus. Participants read phonetically balanced Harvard Sentences while one hand was immersed in cold or warm water and reported pain only during periodic pain statements. This protocol generated 5,750 no signal Harvard Sentence utterances whose transcripts contained no lexical pain information and 1,294 signal pain statement utterances in which the pain rating was explicitly spoken. In the no signal arm, pain was recoverable from acoustic features (AUC 0.622, 95% CI 0.553 to 0.662), whereas transcript based prediction was near chance (AUC 0.489, 95% CI 0.418 to 0.504). Because automatic speech recognition removes the acoustic pain cues, any pain score inferred solely from the transcript is unsupported by the available evidence. Under cooperative prompting, six models abstained on nearly all no signal transcripts, correctly extracted spoken pain ratings in the positive control task with accuracies ranging from 0.939 to 1.00, and maintained an expected calibration error of at most 0.100. Under authority framed prompts, abstention became prompt dependent, with the same model ranging from 0.18 to 1.00 across equivalent prompt phrasings. Most models produced low confidence estimates when forced to answer, whereas Gemini 2.5 Flash and Llama 3.1 8B consistently generated confident pain scores with confident fabrication rates of 0.53 and 0.76, compared with at most 0.15 for all other models. No significant demographic effects were observed in forced responses, with all p values greater than or equal to 0.20.
[AI-79] A Task-Centric Ontology and Deterministic Domain Rules as a Verifiable Core for AI-Assisted Chemistry Problem Solving DATE
链接: https://arxiv.org/abs/2608.26164
作者: Ibrokhimsho Abduchaborov
类目: Artificial Intelligence (cs.AI)
备注: 15 pages, 4 figures, 5 tables. Proof-of-concept evaluation of a task-centric chemistry ontology and deterministic rule engine on 300 human-authored and manually validated school-level problems. The proposed LLM translation layer and token-efficiency hypothesis are not evaluated in the current study
Abstract:Large language models can interpret natural-language chemistry questions, but their internal reasoning is difficult to inspect, constrain, and validate. This paper presents ChemOntoRule, a proof-of-concept symbolic core for AI-assisted school-level chemistry problem solving. The central design choice is task-centric ontology engineering: the ontology is constructed around the concepts, properties, relations, and executable procedures required by a defined collection of chemistry problems, rather than as a universal representation of chemistry. The implemented artifact combines a lightweight ontology serialized in JSON and RDF/Turtle with deterministic Python rules for electronic structure, periodic trends, oxidation states, oxide and hydride behavior, and related school-level reasoning patterns. A separate expert-coded fallback handles problem families not yet represented by general rules. The system was examined on 300 human-authored and manually validated chemistry problems. The complete system matched 296 of 300 reference answers (98.67%). The ontology-driven rule subset covered 269 problems and matched 266 references (98.88%); 31 problems were handled by task-specific expert-coded fallbacks, with 30 matches. Because the same collection informed ontology construction and evaluation, these results measure implemented coverage and internal consistency, not independent generalization. We analyze the four mismatches, distinguish structural validation from chemical correctness, and define a future architecture in which a language model acts primarily as a translator from user language into a normalized ontological task frame. Token efficiency is presented as a testable hypothesis for future controlled studies, not as a result of the current work.
[AI-80] A Safety-Gated Multimodal AI Backend for Mental-Health Support: Hierarchical State Representation Conservative Risk Fusion and Controlled Generation in Anian
链接: https://arxiv.org/abs/2608.26162
作者: Lei Wang,Xiao Wang,Lei Li
类目: Artificial Intelligence (cs.AI)
备注: 16 pages, 4 figures
Abstract:Safety-critical mental-health support systems must distinguish when supportive conversation is appropriate from when free-form generation should be blocked. This paper presents Anian, a safety-gated multimodal AI backend for perinatal mental-health support and mindfulness-intervention routing. Anian is not intended to diagnose psychiatric conditions or replace clinical care or crisis intervention. Its modular pipeline places generative AI downstream of structured state representation, conservative risk fusion, and response gating. User text or voice-derived ASR transcripts are mapped into four linked layers: L1 emotion states, L2 psychosocial constructs, L3 safety risk, and L4 intervention routes. Local text- and rule-based safety evidence is fused with external voice-derived evidence using a highest-risk-priority rule, S_fusion = max(S_local, S_external). At moderate or high fused risk, ordinary AI-generated responses and text-to-speech delivery are blocked and replaced by fixed safety content and prompts for human support. An internal prototype evaluation used approximately 858,295 normalized records from public emotion, dialogue, mental-health-related, and Chinese dialogue corpora within a weak-label and rule-derived framework. Micro-F1 scores were 0.9604 for L1 emotion classification, 0.9144 for L2 psychosocial constructs, and 0.9742 for L4 routing. In a controlled safety stress test of 233 samples, the L3 rule engine achieved high-risk recall of 1.0000 within predefined scenarios. These findings support the internal feasibility of the label framework and gating logic but do not establish clinical validity, diagnostic accuracy, real-world safety, or effectiveness. We report the architecture, ontology, safety-fusion mechanism, prototype evaluation, error-analysis plan, and roadmap for expert-reviewed and real-world validation.
[AI-81] SAREF-based Ontology for Distributed AI Workflows across the Edge-Fog-Cloud Continuum
链接: https://arxiv.org/abs/2608.26160
作者: Viorica Rozina Chifu,Tudor Cioara,Vasile Ofrim,Liana Toderean,Ionut Anghel,Laura Daniele,Cornelis Bouter
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Nowadays semantic models provide limited support for representing distributed AI workflows and their execution across heterogeneous edge, fog, and cloud environments. Therefore, AI processes and resources are often described using incompatible semantic representations, affecting the interoperability, orchestration, and reuse. To address these challenges, this paper proposes a SAREF-compliant ontology for representing distributed AI workflows across the edge-fog-cloud continuum. We extend the SAREF4SYST ontology with concepts for modeling AI pipelines, executable AI jobs, computational resources, deployment constraints, and communication relationships, providing a unified semantic model of both AI workflows and heterogeneous computing infrastructures. The ontology enables semantic interoperability, automated reasoning, and resource-aware orchestration of distributed AI applications while remaining fully aligned with the ETSI SAREF ecosystem. The ontology is evaluated using proof-of-concept smart grid energy services orchestration scenarios and validated using competency questions showing its ability to support AI workflow deployment, execution reasoning, and workload adaptation across heterogeneous edge, fog, and cloud environments. All competency questions were successfully validated using SPARQL querying and semantic reasoning. Experimental results demonstrate deployment success rates of 90-100% with average orchestration decision times below 80 ms across heterogeneous edge-fog-cloud environments, highlighting its effectiveness on ensuring semantic interoperability for distributed AI orchestration.
[AI-82] GROUND: Reducing Hallucinations in LLM -Based Enterprise Analytics Through Governed Semantic Definitions
链接: https://arxiv.org/abs/2608.26157
作者: Aravind Sasidharan Pillai
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Natural-language analytics over enterprise data warehouses is increasingly important, but production use is limited by hallucinated metrics, invalid joins, wrong grain, unsafe data access, and unsupported explanations. Existing text-to-SQL systems often ground generation in database schemas or retrieved documentation, while enterprise reporting also requires governed business semantics: approved metrics, dimensions, join paths, filters, and row-level security. This paper introduces GROUND, Governed Retrieval Over Unified Normalized Definitions, a framework that constrains LLM-generated analytics to a governed semantic layer. GROUND supplies approved definitions, binds user intent to governed metrics and dimensions, and validates generated SQL against schema, metric, join, grain, filter, security, and cost rules before execution. On violations, it retries or abstains. In a 100-question synthetic enterprise-reporting benchmark, GROUND is compared with direct schema-only text-to-SQL, schema-RAG, and semantic-only grounding under one shared model. GROUND is the only system free of measured hallucinations across all six evaluated categories, while ungoverned systems violate row-level security on many questions. A semantic-only condition with exact metric definitions but no access policy still leaks data, showing that governance cannot be replaced by metric fidelity alone. The findings are replicated on real U.S. NHTSA vehicle-safety data with independent hand-authored gold and tested on an adversarial set across four models from three providers. GROUND’s enforced guarantees, especially filters and row-level security, hold with zero violations on every model, while judgment-dependent behaviors such as refusing undefined metrics remain fallible. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2608.26157 [cs.AI] (or arXiv:2608.26157v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.26157 Focus to learn more arXiv-issued DOI via DataCite
[AI-83] Selection Bias Correction in Retail Intelligence
链接: https://arxiv.org/abs/2608.26156
作者: Spandan Ghose Chowdhury
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to the 15th International Conference on Data Science, Technology and Applications (DATA 2026)
Abstract:Retail intelligence often relies on monitoring popular, high-velocity products, potentially biasing economic indicators by ignoring the “long tail” of niche items. This simulation study investigates selection bias in inflation estimation and compares correction methods across diverse data-generating processes. Through 400 Monte Carlo replications spanning four scenarios–aligned step functions, smooth gradients, misaligned breaks, and polynomial relationships–we test the robustness of Inverse Probability Weighting (IPW) with five specifications against stratification with varying strata counts. Our findings reveal fundamental limits of weighting methods in retail long-tail contexts: stratification achieves superior performance in three of four scenarios, maintaining sub-0.04pp median error even when boundaries deliberately misalign with population breaks (116x advantage over IPW). However, IPW with spline propensity models wins under smooth polynomial relationships (median error 0.007pp vs. 0.013pp), demonstrating context-dependency. Critically, even an oracle IPW specification with perfect structural knowledge achieves 6.06pp error compared to stratification’s 0.008pp in step-function scenarios. This reflects violation of the Positivity Assumption–a fundamental causal inference requirement–rather than IPW methodological inferiority. When selection probabilities differ dramatically (90% vs. 1%), weighting methods operate outside their theoretical design envelope. These results demonstrate that stratification provides a safer engineering choice in retail long-tail distributions with severe positivity violations.
[AI-84] EEG-to-Report: An Annotation and Feature-Text Framework for Training Language Models on Clinical EEG
链接: https://arxiv.org/abs/2608.26153
作者: Xuan-The Tran,Le Trung Kien Nguyen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Clinical electroencephalography (EEG) reporting remains largely manual and time-consuming, and current EEG software ecosystems do not produce the structured EEG-text supervision needed for training modern language models. Most toolboxes focus on visualization or preprocessing, providing limited support for workflows that generate high-quality datasets for AI. We introduce EEG-to-Report, a browser-based annotation and feature-text framework that links routine EEG review with the construction of AI-ready datasets. The framework integrates multi-format EEG ingestion, channel standardization, and an interactive viewer with a multimodal annotation layer that combines typed text and transcribed voice notes. For each annotated segment, a feature extraction engine computes a standardized set of spectral, temporal, entropy, Hjorth, connectivity, and spike-related descriptors, stored alongside clinical descriptions in a portable JSON schema. This yields aligned feature-text pairs designed to supervise multimodal EEG-language models. The framework also includes an auto-report module that couples an ensemble of convolutional networks with a large language model to draft clinical narratives for neurologist review. Using pilot annotations, we describe how EEG-to-Report streamlines annotation workflows and produces editable draft reports, providing a reusable foundation for automated EEG reporting systems.
[AI-85] Explainable Artificial Intelligence for Customer Churn Prediction in Telecommunications: A Framework for CRM Integration
链接: https://arxiv.org/abs/2608.26151
作者: Sandeep Gaddamwar
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 10 pages, 7 figures, 1 table
Abstract:Subscriber attrition is a costly, persistent challenge for telecommunications providers, with monthly churn of roughly 1.9% in mature markets eroding billions in revenue annually. Predictive models can flag at-risk customers accurately, yet they are routinely excluded from frontline CRM workflows because high-performing ensemble and non-linear architectures are opaque: a retention specialist cannot design a personalised intervention from a probability score alone, without knowing why a subscriber is at risk. This paper addresses that gap. We benchmark four classifiers–Logistic Regression, Random Forest, XGBoost, and LightGBM–on the IBM Telco Customer Churn benchmark (7,043 records; 19 features; 26.5% churn, balanced to 50% via SMOTE on the training partition only). Logistic Regression attains the strongest AUC-ROC (0.8411) and LightGBM the highest accuracy (78.42%); all four fall within a 0.011 AUC band (0.831–0.841), and 5-fold cross-validation confirms the leading models are effectively tied. Explanations are delivered at two granularities: a global SHAP ranking identifying tenure, total charges, and month-to-month contract as the dominant churn signals, and instance-level SHAP and LIME decompositions that expose the drivers behind each prediction. Building on these outputs, we introduce a four-layer CRM integration architecture that converts risk scores and attribution vectors into tiered segmentation, maps top features to structured retention-action templates, and routes campaign outcomes into a retraining feedback loop. Targeting the highest-risk quintile is projected to cut overall churn by 3.3–5.3 percentage points, preserving an estimated 199K-- 319K per campaign cycle.
[AI-86] Methodological and Conceptual Framework for 5D Multi-Table Analysis: A Unified Approach for Complex Data Reuse
链接: https://arxiv.org/abs/2608.26149
作者: Edouard Lansiaux,Hugo Kazzi,Aurélien Loison,Slim Hammadi,Emmanuel Chazard
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Multi-table learning remains a major challenge in machine learning for healthcare and other complex information systems. Relational data combine several sources of complexity, including large data volume, high-dimensional variables, high-cardinality categorical features, complex inter-table dependencies, and repeated temporal observations. We introduce the Relational Hypergraph Transformer (RHT), a unified architecture that represents relational databases as hypergraphs, learns pentadimensional embeddings (PentE), and performs sparse relational attention with complexity proportional to the average relational degree rather than the square of the number of entities. We formally define the architecture, derive the complexity of its attention mechanism, and provide an open-source reference implementation. We evaluate RHT on the public Synthea synthetic electronic health record dataset using multi-label prediction of SNOMED CT condition codes per encounter, a task characterized by high categorical cardinality and long-tailed label distributions. Comparisons with tabular, relational, and temporal graph baselines show that RHT produces more semantically coherent embeddings while remaining computationally scalable. In this benchmark, the highest rare-code recall is achieved by XGBoost, whereas RHT attains the strongest embedding semantic coherence. We also report ablation studies quantifying the contribution of each architectural component. Clinical validation on MIMIC-IV is planned following PhysioNet credentialing. Source code and experimental protocols are provided in the accompanying repository.
[AI-87] he Accuracy-Efficiency Paradox Quantifying Net Energy Loss in on-Device Energy Forecasting
链接: https://arxiv.org/abs/2608.26134
作者: Jaeik Jeong,Tai-Yeon Ku,Wan-Ki Park
类目: Artificial Intelligence (cs.AI)
备注: The 5th International Conference on Mobile, Military, Maritime IT Convergence (ICMIC 2026)
Abstract:Energy forecasting aims to maximize accuracy to ensure energy efficiency by reducing energy waste, an objective that applies equally to on-device forecasting for mission-critical edge environments, including military systems. However, this paper identifies the Accuracy-Efficiency Paradox: high-precision energy forecasting models can ironically trigger a net energy deficit. This stems from both edge AI’s inference energy consumption and battery aging. We propose a Total Cost of Ownership (TCO) framework for energy forecasting, designed to minimize net energy loss. This framework treats not only inference energy consumption but also battery aging as a unified form of energy loss, as degradation represents a physical dissipation of the system’s future energy-carrying capacity. We demonstrate that in thermally sensitive edge environments, energy saved by the superior precision of complex architectures is often outweighed by the total energy lost through their high operational intensity.
[AI-88] From SQL to Knowledge Graphs: An LLM -Driven Multi-Agent Approach with Data Schema Improvement
链接: https://arxiv.org/abs/2608.26117
作者: Dinh-Khanh Pham,Quy-Anh Dang,Lam Mai Thanh,Khanh Bui,Truong-Son Hy
类目: Databases (cs.DB); Artificial Intelligence (cs.AI)
备注:
Abstract:RDBMS (Relational Database Management System) databases face several limitations, including slow execution with multi-hop queries and a lack of explainability by graphical interpretations. In contrast, Graph database offers a more intuitive and efficient data schema that performs faster execution on large datasets. Most existing RDBMS conversion pipelines focus on running traditional loading commands and relying on Cypher queries. However, the efficiency of using an LLM to generate an effective graph data schema, significantly reducing the ambiguity of the graph database, remains underexplored in the current research literature. This paper presents a novel algorithm that bridges RDBMS and graph database by using a novel LLM-powered ETL agent to standardize table and column names before saving them to the Data Mart. A Multi-Agent System generates a looping discussion between ETL, Analyzer, and Graph agents to optimize the final design through an iterative process of suggesting and scoring the graph database schema. We ensure that the final graph database meets three criteria before being accepted for data conversion: Accuracy, Groundedness, and Faithfulness. This system demonstrates an effective pipeline to automatically convert a tabular database into a graph database through a comprehensive end-to-end process. Our study highlights notable efficiency in using the converted graph database, which is measured on 1,081 samples of the BFSI dataset across three levels of complexity (easy, medium, and hard). Specifically, CypherAgent achieves an 85.6% accuracy for QA tasks using a Graph database, which is 12.12% higher than the accuracy achieved by an SQLAgent on the RDBMS database type PostgreSQL, for all queries. Additionally, the Graph database demonstrates faster performance, reducing latency by approximately 3 times.
[AI-89] he Artificial Experimentalist: Discovery and Control of Self-Organizing Phenomena with Autotelic Reinforcement Learning
链接: https://arxiv.org/abs/2608.26116
作者: Marko Cvjetko,Benedikt Hartl,Michael Levin,Clément Moulin-Frier,Pierre-Yves Oudeyer
类目: Artificial Intelligence (cs.AI)
备注: 8 pages, 8 figures. Accepted at ALIFE 2026. Companion website: this https URL
Abstract:Existing methods for exploring cellular automata and other complex systems mostly operate in open loop: they set initial conditions, execute a full simulation, and observe the outcome, without intervening during execution. We introduce a closed-loop framework based on autotelic reinforcement learning, in which an agent autonomously samples diverse goals and learns a goal-conditioned policy to intervene in a complex system through minimal, local perturbations. We instantiate this framework on Lenia, a continuous cellular automaton known for life-like self-organizing patterns, in an agentic system we call CARL, and demonstrate three capabilities. First, CARL discovers stable solitons across a wide range of Lenia update rules at a higher rate than heuristic baselines. Second, it learns to steer the movement direction of existing solitons with few interventions, showing that CARL can control self-organizing patterns, not only create them. Third, humans can use trained agents to guide solitons through maze environments in real time by specifying high-level directional commands that the agent translates into low-level interventions. Trained across diverse goals, update rules, and random initial states, the agents acquire policies that generalize zero-shot to various out-of-distribution conditions. These results suggest a path toward artificial experimentalist agents that, autonomously or with human guidance, discover and control emergent phenomena in complex systems.
[AI-90] PICasso: An AI-Enabled Design Framework for Autonomous Optimization of Silicon Photonic Devices
链接: https://arxiv.org/abs/2608.26113
作者: Deepak Vungarala,Deniz Najafi,Abdulrahman Aljoudi,Zahra Ghanaatian,Navid Khoshavi,Gourav Datta,Arman Roohi,Mahdi Nikdast,Shaahin Angizi
类目: Artificial Intelligence (cs.AI)
备注: 5 Tables, 4 Figures
Abstract:We present PICasso, an AI-assisted framework for automated synthesis, verification, and optimization of photonic integrated circuits (PICs) from natural-language specifications. PICasso couples a structured NL - YAML - GDS generation pipeline with PDK aware knowledge injection, automated placement and routing, DRC/LVS validation, and SAX-based photonic simulation. To systematically evaluate AI-driven photonic design, we introduce PIC-Set, a benchmark of 36 parameterized PIC design tasks spanning core photonic primitives and multi-component circuits. Using PIC-Set, we benchmark several state-of-the-art Large Language Models (LLMs) under a unified evaluation protocol, including new metrics such as structural and functional Spec@k , optimization efficiency, and robustness under perturbations. Across the benchmark, PICasso significantly improves end-to-end specification satisfaction compared to vanilla LLM generation. Structural Spec@3 reaches up to 92.7% and functional Spec@3 up to 52% on high-complexity circuits. In addition, PICasso consistently reduces circuit insertion loss, lowering the mean loss from 4.98 dB to 3.25 dB (1.74 dB improvement) through simulation-guided optimization. These results demonstrate that structured domain constraints, physical verification, and simulation feedback transform LLMs from brittle netlist generators into practical PIC design agents capable of producing manufacturable layouts with competitive runtimes relative to manual GUI-based workflows.
[AI-91] Large Models for Battery Prognostics and Health Management: A Review and Future Roadmap
链接: https://arxiv.org/abs/2608.26111
作者: Jiale Liu,Huan Wang,Weicheng Wang,Rong Zhu,Qiqi Wang,Min Xie
类目: Artificial Intelligence (cs.AI)
备注: Published in Renewable and Sustainable Energy Reviews
Abstract:Battery Prognostics and Health Management (BPHM) is critical for ensuring the safe, reliable, and cost-effective operation of batteries across electric vehicles, grid storage, and consumer electronics. Conventional BPHM approaches, including physics-based models and task-centric deep learning methods, face challenges in computational efficiency and parameterization, cross-domain generalization, dependence on extensive labeled run-to-failure data, and model interpretability. Recent Large Models (LMs), built upon Transformer architectures and self-supervised pre-training, offer a transformative new paradigm to overcome these long-standing bottlenecks. This review provides the first comprehensive survey of LM applications in BPHM, systematically examining how these models address challenges in the field. We begin by elucidating the foundational technologies enabling LMs, including Transformer architectures, self-supervised learning, large-scale multimodal datasets, and PEFT techniques. We then categorize recent progress along four critical dimensions: mitigating data scarcity, enhancing generalization and robustness, integrating domain knowledge for interpretability, and enabling system-level automation. Despite promising results, significant challenges remain across data accessibility, intelligence validation, trustworthiness, and deployment feasibility. To guide future research, we propose a roadmap focused on building collaborative data ecosystems, validating intelligence for industrial applications, enhancing trustworthiness with physics-informed designs, and enabling efficient on-device deployment. This review establishes a systematic approach to understand and advance LM-driven BPHM, providing researchers and practitioners with essential insights for developing next-generation battery management systems capable of safe, reliable, and autonomous operation throughout battery lifecycles.
[AI-92] Exploring the Role of LLM s in HPC Programming: A Survey
链接: https://arxiv.org/abs/2608.26110
作者: Strahinja Ljaljevic,Josep Jorba,Sergio Iserte
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models (LLMs) are emerging as promising assistants in High-Performance Computing (HPC), where programming remains complex and expertise-intensive. This survey systematically reviews their application across five categories: code generation, parallelization and optimization, frameworks and architectures, evaluation and benchmarking, and broader challenges. The analysis highlights both opportunities and limitations: while general-purpose LLMs perform reasonably well on serial and OpenMP-like tasks, they fall short in distributed paradigms such as MPI, where correctness and scalability are critical. Domain-specialized models (e.g., HPC-Coder, HPC-GPT, chatHPC) achieve higher accuracy through fine-tuning, curated datasets, and retrieval-augmented generation (RAG), yet their scope remains narrow and their evaluations largely limited to benchmarks or micro-kernels. The broader picture is one of dual potential and fragility: LLMs can lower barriers to entry, accelerate prototyping, and support code modernization, but they remain brittle under production-level requirements where correctness, performance portability, and scaling cannot be compromised. We conclude that LLMs are unlikely to replace HPC experts in the near term but are positioned to become powerful collaborators in the software development pipeline. Their effective deployment will require richer datasets, integration with performance analysis and schedulers, rigorous evaluation frameworks, and governance structures that ensure transparency and trust. The convergence of AI and HPC should therefore be understood as a long-term, co-evolutionary process, where each advance uncovers new challenges and opportunities for reshaping scientific software development. Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI) Cite as: arXiv:2608.26110 [cs.DC] (or arXiv:2608.26110v1 [cs.DC] for this version) https://doi.org/10.48550/arXiv.2608.26110 Focus to learn more arXiv-issued DOI via DataCite Journalreference: Future Generation Computer Systems, 2026, 108618, ISSN 0167-739X Related DOI: https://doi.org/10.1016/j.future.2026.108618 Focus to learn more DOI(s) linking to related resources
[AI-93] Standalone LLM and a Pre-specified Agent ic Pipeline for Explaining ICU Mortality Predictions: a Feasibility Study on the eICU Demo Dataset
链接: https://arxiv.org/abs/2608.26109
作者: Di Zhu,Chen Xie,Haoyun Zhang,Zihan Wei,Ziwei Wang,Jiazhao Shi,Ziyu Wang,Qiyang Xie
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Machine-learning models can predict ICU mortality accurately, but feature-attribution methods alone rarely provide the clinical narrative needed for bedside use. Large language models (LLMs) may bridge this gap, and multi-step agentic pipelines are a plausible extension because they separate data interpretation, guideline checking, and final explanation. This revised feasibility study preserves the original standalone-versus-agentic comparison while making the main clinical findings more explicit. Using the retained local eICU Demo artifact set (2,353 ICU stays; 8.1% mortality), XGBoost achieved an AUROC of 0.855 (95% CI 0.796–0.906) and an AUPRC of 0.332 (95% CI 0.217–0.494). On a stratified 38-case explanation subset, the standalone LLM produced 1 explanation with explicit outcome leakage, whereas the four-step agentic pipeline produced none. Among the 14 cases that overlapped with the SHAP review subset, the standalone LLM showed higher SHAP alignment (mean Jaccard 0.171 versus 0.077) and higher direction consistency (92.9% versus 78.6%), while the agentic pipeline showed higher guideline grounding (0.762 versus 0.143), higher value specificity (0.236 versus 0.143), and slightly higher plausibility (0.700 versus 0.671). Clinically, the results suggest that agentic decomposition may improve safety-relevant grounding and patient-specific detail, but it should be paired with attribution-based checks before use in high-stakes risk explanation.
[AI-94] EduRiskX: A Neuro-Symbolic Framework with F-Logic Reasoning for Early Academic Risk Prediction
链接: https://arxiv.org/abs/2608.26107
作者: Yu Fu,Yongqi Kang,Yong Zhao,Rongfang Bie
类目: Artificial Intelligence (cs.AI)
备注: Previously available on Research Square: this https URL . This version presents the complete neuro-symbolic framework, EduRiskX, integrating temporal Transformers with F-Logic reasoning
Abstract:Predicting students’ academic risk in online education is crucial for enabling timely interventions that can improve retention and learning outcomes. However, existing models often suffer from limited early detection capability and insufficient interpretability, leading to a “black-box” trust crisis that hinders their adoption in real-world pedagogical settings. To address these challenges, we propose EduRiskX, a neuro-symbolic framework that integrates a temporal Transformer-based predictor with F-Logic symbolic reasoning. The neural component models longitudinal student activity sequences using temporal attention, class-weighted loss, and dynamic weekly truncation. Acting as a data-driven expert system, an F-Logic rule base – grounded in established educational theories (Engagement Theory and Student Integration Model) to mimic the diagnostic logic of human educators – is constructed exclusively from the training data. The neural risk probability and the symbolic confidence score are then combined through a logistic regression-based fusion mechanism that learns the relative contribution of each signal. Experiments on the Open University Learning Analytics Dataset (OULAD) using a strict 80/10/10 student-level split show that EduRiskX achieves an accuracy of 0.900 and an F1-score of 0.894 at the end of the semester (Week 38), with an average early detection week of 9.32 and a detection rate of 94.30 percent. Compared with state-of-the-art time-series models (PatchTST, iTransformer) and common deep learning baselines (LSTM, CNN), EduRiskX yields improved recall and earlier risk identification under identical conditions. Beyond predictive performance, the F-Logic module provides structured rule-based explanations linking predictions to observable behavioral patterns and educational theories.
[AI-95] Property-Specific Recoverability from Contact PPG to Camera rPPG under Heterogeneous Observation Conditions
链接: https://arxiv.org/abs/2608.27392
作者: Timothy Oladunni,Farouk Ganiyu-Adewumi
类目: ignal Processing (eess.SP); Artificial Intelligence (cs.AI)
备注:
Abstract:Camera-derived remote photoplethysmography (rPPG) is commonly validated through endpoint accuracy, but endpoint performance does not establish whether other physiological properties of source contact photoplethysmography (PPG) remain preserved recording by recording. We evaluated property-specific PPG-to-rPPG recoverability on 655 recordings from the Multi-Domain Mobile Video Physiology Dataset using CHROM as a fixed camera-rPPG observation pathway. The pathway reproduced the published CHROM correlation regime, with heart-rate MAE of 15.26 bpm and Pearson correlation of 0.0801. Matched-versus-shuffled validation revealed modest recording-specific autocorrelation correspondence, while spectral and recurrence-rate measures showed little matched discrimination. Maximal Lyapunov exponents showed essentially no recording-specific PPG-to-rPPG correspondence, with correlation of 0.0231 and permutation p-value of 0.5584, despite population-level overlap. Endpoint discrepancy exhibited Fitzpatrick-associated heterogeneity after adjustment for lighting and motion, including a Fitzpatrick VI versus III contrast of 9.32 bpm, while dynamical discrepancy showed no corresponding gradient. Aggregate RGB signal-to-noise ratio did not materially account for the endpoint contrast. In subject-held-out analysis, adding motion and lighting consistently reduced MAE across linear, ridge, and random-forest learners relative to rPPG-HR-only calibration, with reductions up to 13.32 percent. These findings show that recoverability is property-specific: physiological properties differ in recording-specific preservation and dependence on observation conditions, and population-level plausibility does not establish preservation of individual recordings.
[AI-96] Active Diffusion-Based Inference for Ill-Posed Inverse Problems under Incomplete Priors ECAI2026 IJCAI
链接: https://arxiv.org/abs/2608.27080
作者: Jitao Xu,Nobuo Sato,Yaohang Li
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted for publication in the Proceedings of IJCAI-ECAI 2026
Abstract:Many scientific and engineering applications require estimating unknown parameters from experimentally observable data – an inverse problem that is inherently challenging due to nonlinearity, noise, and ill-posedness. In this paper, we propose an active diffusion-based inverse problem solver. A DM is trained to learn the mapping between the parameter space and the observable space. By iteratively detecting and correcting model misspecification through posterior uncertainty, the method discovers and learns the correct region of parameter space, even when initial training bounds exclude the true parameters. This provides a principled, Bayesian justification for adaptive domain augmentation and ensures robust inference for inverse problems under incomplete prior knowledge. We demonstrate the effectiveness of our inverse solver for a toy inverse problem with infinite solutions, and for the parameterization of the quantum correlation functions to event observables in a Quantum Chromodynamics analysis of nucleon structure.
[AI-97] Learning Transverse Momentum Distributions from Raw Scattering Events via Conditional Diffusion
链接: https://arxiv.org/abs/2608.27077
作者: Jitao Xu,Christopher Cocuzza,Kevin Braga,Daniel Lersch,Nobuo Sato,Yaohang Li
类目: High Energy Physics - Phenomenology (hep-ph); Artificial Intelligence (cs.AI)
备注: Accepted at Physics and AI at Stanford University (PAI 2026)
Abstract:Extracting transverse momentum dependent parton distribution functions (TMD PDFs) from semi-inclusive deep inelastic scattering (SIDIS) data is a central goal of the nucleon structure program at Jefferson Lab and the future Electron-Ion Collider. Traditional extraction methods rely on parameterized functional forms and iterative fitting, which can limit the flexibility of the resulting distributions and make uncertainty quantification cumbersome. We present a conditional diffusion model that learns to map raw SIDIS event kinematics directly to TMD PDFs, bypassing explicit functional assumptions. Evaluated on simulated SIDIS data at CLAS12 kinematics, the model recovers the underlying TMD with informative uncertainties that narrow steadily with increasing event statistics, and produces reliable estimates even with as few as 1,000 conditioning events, a statistics-limited regime directly relevant to ongoing and planned experiments.
[AI-98] Magnon-induced phononic Chern insulator
链接: https://arxiv.org/abs/2608.27011
作者: Rui-Chang Shen,Yihao Yang,Haoran Xue
类目: Mesoscale and Nanoscale Physics (cond-mat.mes-hall); Artificial Intelligence (cs.AI); Applied Physics (physics.app-ph)
备注: 7 pages, 3 figures
Abstract:High-frequency artificial phononic crystals offer a low-loss platform compatible with on-chip integration, yet realizing Chern phononic phases at GHz frequencies remains challenging. Here, we propose a magnon-induced phononic Chern insulator in a honeycomb phononic crystal hybridized with ferromagnetic islands at the hexagon centers. A circularly polarized Kittel mode couples to the surrounding phonons with a phase winding, which breaks time-reversal symmetry and opens a full Chern gap. In the large-detuning regime, this mechanism leads to an effective Haldane-type phononic model with magnon-induced complex hopping. By tuning the magnon-phonon interaction, the full hybrid system accesses Chern phases with tunable Chern numbers |C|=1 and |C|=2. The predicted gaps can exceed realistic phonon and magnon linewidths, enabling their observation in GHz acoustic devices. Our work establishes chiral magnon–phonon hybridization as a route to magnetically reconfigurable topological phononics.
机器学习
[LG-0] oken-Level Advertising
链接: https://arxiv.org/abs/2608.27382
作者: Hanbing Liu,Bowei Zhang,Changyuan Yu,Yinyu Ye,Qi Qi
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG)
*备注:
Abstract:Generative AI is transforming how people access information, challenging traditional advertising mechanisms built around predefined slots. Towards generation-native advertising, we propose the Latent Advertiser Mixture Auction (LAMA), a token-level advertising mechanism that embeds advertiser influence directly into the generation process. Advertisers report local continuation values that induce advertiser-specific next-token policies, from which the platform decodes through a latent mixture while updating an allocation posterior. We show that LAMA satisfies Markov DSIC and IR, and achieves near-optimal KL-regularized welfare. We further develop a learning-based implementation that reconstructs the required reports online from learned local advantages and root values. Proof-of-concept experiments on real-world commercial-search query splits show that LAMA improves platform welfare and revenue while maintaining user-facing response quality, providing initial evidence for the feasibility of generation-native advertising.
[LG-1] Understanding Evolution Strategies for LLM Reasoning Reasoning : Broader Reasoning Coverag e than GRPO
链接: https://arxiv.org/abs/2608.27351
作者: Yunpeng Ba,Zhi Zheng,Yue Xie,Jiaqing Li,Xialiang Tong,Tao Zhong,Mingxuan Yuan,Zhichao Lu,Xuyang Wu,Zhenkun Wang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Evolution Strategies (ES) have recently emerged as a memory-efficient post-training paradigm for LLM reasoning. However, the optimization behavior of ES remains understudied, making it hard to define its advantage scope compared to mainstream post-training paradigms (e.g., Group Relative Policy Optimization (GRPO)). By systematically investigating ES dynamics and mechanisms, this paper first identifies a performance advantage of ES over GRPO, theoretically and empirically showing that ES can lead to broader reasoning coverage, thereby better exploiting the reasoning capabilities of pretrained LLMs. Theoretically, we show that verifier-projected Jensen-Shannon diversity across the ES population is helpful to higher Pass@K performances. Empirically, unlike GRPO, which exhibits entropy collapse, ES improves Pass@1 while attaining higher Pass@K than GRPO. We further develop a sequential GRPO-ES training strategy that combines GRPO’s strength in Pass@1 with ES’s gains in Pass@K. Second, we find that despite substantial whole-model parameter drift, the task-performance gains of ES are only contributed to a sparse subset of larger-magnitude updates. This functional sparsity suggests that large parameter movement need not imply widespread functional change, and held-out evaluations further show that it does not necessarily lead to catastrophic forgetting. Finally, we study how hyperparameter design affects the effectiveness of ES, demonstrating that ES requires a smaller population size in a larger LLM. These findings position ES as a distinct reasoning post-training paradigm rather than a less effective, memory-efficient alternative to GRPO.
[LG-2] QuantumBoostNet: A Hybrid Classical-Quantum Architecture for Enhanced Accuracy in Cardiac Ultrasound View Identification
链接: https://arxiv.org/abs/2608.27302
作者: Mihai Udrescu-Milosav,Stefan-Alexandru Jura,Mihai Udrescu,Gerhard-Paul Diller
类目: Machine Learning (cs.LG)
*备注:
Abstract:Accurate identification of the correct view or angle in cardiac ultrasound (echocardiogram) is a critical component of cardiologic imaging. This step is essential for precise anatomical interpretation, reliable measurement, and the reduction of clinical errors. Although computer vision has advanced significantly, most state-of-the-art models perform well on standard benchmarks but often yield suboptimal results in specialized medical imaging tasks due to the high level of noise present in the data. QuantumBoostNet, a hybrid classical-quantum architecture, is introduced to address these challenges. This model integrates a classical backbone with two heads: one classical and one quantum, with the quantum head implemented as a parametrized 10-qubit quantum circuit. Training occurs in two stages, with an adaptive transition between heads governed by a mixing parameter that monitors loss dynamics. Extensive experiments indicate that, despite the limited number of qubits that can be simulated, QuantumBoostNet consistently outperforms state-of-the-art classical and hybrid classical-quantum models in cardiac ultrasound view identification, achieving a relative improvement over the best competitor. QuantumBoostNet also demonstrates superior performance on established image classification benchmarks and exhibits robustness to noise. These findings support the continued development of hybrid classical-quantum models for specialized medical imaging applications.
[LG-3] MM-Spectrum: Multimodal Multi-spectral Molecular Structural Elucidation with a Stable MoE Framework ICML2026
链接: https://arxiv.org/abs/2608.27286
作者: Hai-tao Yu,Nan Min,Zheng Fang,Hongyu Zhan,Yusen Tan,Yuhan Wang,Jun Xia
类目: Machine Learning (cs.LG)
*备注: Accepted to ICML 2026
Abstract:Inferring molecular structures from multimodal spectroscopic measurements requires integrating complementary yet highly heterogeneous signals. However, the common paradigm of directly concatenating multispectral sequences can exhibit anomalous performance degradation, primarily due to pronounced heterogeneity and the resulting multimodal imbalance across modalities. As a remedy, we propose MM-Spectrum, a sparse Mixture-of-Experts framework tailored for multimodal multispectral spectra-to-structure elucidation. To better match the information characteristics under multispectral imbalance, MM-Spectrum introduces an explicit modality-aware routing mechanism that exposes spectral identity to the router in addition to token content representations. Moreover, it incorporates shared and interaction experts, together with heterogeneous expert capacities, to extract multispectral modality-unique and cross-modal synergistic information while suppressing noise-induced interference. Across full-modality, bimodal, and missing-modality settings on molecular structural elucidation, MM-Spectrum achieves consistent and substantial improvements, supported by ablation studies and interpretability analyses.
[LG-4] Making Latent Evolution Explicit: Operator-Structured Transitions for World Action Models
链接: https://arxiv.org/abs/2608.27259
作者: Xiaoxiao Lu,Yunlong Dong,Jiahao Shi,Ye Yuan
类目: Machine Learning (cs.LG)
*备注:
Abstract:World Action Models (WAMs) augment robot policies by predicting how task-relevant scene states may evolve under interaction. Recent WAMs increasingly perform such prediction in latent representation spaces, avoiding full appearance-level generation while preserving control-relevant information. Yet latent transitions are commonly realized with Transformer-based predictors whose inductive structure is centered on token interaction rather than temporal evolution. We study transition realization as an architectural choice distinct from predictive representation and prediction-policy coupling. We introduce the Latent Evolution Operator Network (LEON), which models latent evolution in a learned observable space through context-modulated operator-based propagation and additive forcing. Grounded in the controlled Koopman generator view of evolution, LEON organizes context-dependent transition variation around a shared evolution-operator structure while retaining a complementary path for additive change. Controlled dynamical systems verify the resulting evolution-specific inductive bias and the complementary roles of operator propagation and forcing. Across two WAM formulations that integrate latent prediction into the policy differently, LEON improves closed-loop performance and robustness while remaining effective under full transition replacement. These results establish transition realization as a consequential architectural choice in latent WAMs.
[LG-5] Enforcing Dirichlet Boundary Conditions in Operator Learning
链接: https://arxiv.org/abs/2608.27256
作者: Andrew M. Stuart,Margaret Trautner
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG)
*备注:
Abstract:Operator learning in scientific machine learning is concerned with approximation of maps between infinite-dimensional function spaces; such maps frequently arise as the solution operators of partial differential equations (PDEs). Neural operators have demonstrated broad empirical success at approximating such maps from data. However, most existing neural operator architectures enforce boundary conditions indirectly through training from data even though the boundary condition is often known exactly. Furthermore, existing modifications and approaches that do enforce boundary conditions explicitly suffer from impractical restrictions, including boundary smoothness, uniform grids, and separable, box-like domains. In this work, we propose an architecture which, independently of training, satisfies homogeneous Dirichlet boundary conditions, whilst simultaneously retaining the expressivity of existing kernel-integral neural operator architectures. This is achieved by enforcing the property that the output of each layer is contained in the span of a subset of the homogeneous Dirichlet eigenfunctions of the Laplacian on the output domain. The method requires only that the output domain be bounded with Lipschitz boundary and places no restriction on the choice of discretization, making it applicable to arbitrary mesh data and general geometries. We prove universal approximation for the resulting architecture; furthermore the approach we adopt in the analysis proves universality for a broad class of kernel-integral neural operators thereby uniting existing theory for a variety of operator learning methods. We validate the proposed method on maps defined by the coefficient to solution map in 2D PDEs: Darcy flow on a square domain and the Helmholtz equation on a circular domain. Comparisons are made with alternative methods.
[LG-6] Circuit Condensation: Post-Training that Concentrates a Behaviors Causal Circuit
链接: https://arxiv.org/abs/2608.27254
作者: Sai Adith Senthil Kumar
类目: Machine Learning (cs.LG)
*备注: 27 pages, 5 figures
Abstract:One approach to mechanistic interpretability explains behavior through circuits: the components and connections that carry it. Frozen discovery often returns hundreds of edges, making them hard to inspect, compare, or verify exhaustively. We introduce Circuit Condensation, which post-trains models to concentrate behaviors into smaller causal graphs. Each round prunes low-attribution edges and trains a low-rank adapter to match the original through what remains, retaining the cut only if task performance and general capability survive. Across four behaviors and eight models, condensed circuits are smaller than the strongest frozen baseline in 30 of 32 settings, by 8.1\times on average and up to 316\times . Repeating the search without weight updates produces larger circuits in 29 of 32 settings, showing that weight updates, rather than search alone, drive the reduction. Testing every subset of 19 circuits finds 11 that cannot be reduced and reveals removable edges in the rest. Pair ablations expose dependencies between edges, showing that their effects cannot be understood independently. On indirect object identification, condensation isolates 24 heads, 17 of them with documented roles, against 61 heads and 36 undocumented ones for the matched frozen circuit: a sufficient sub-circuit of the published mechanism rather than a reconstruction of it. The resulting circuit tracks the original model’s next-token distribution and predicts its errors.
[LG-7] Importance Scoring of Transformer Attention Heads in Learning Tabular Data
链接: https://arxiv.org/abs/2608.27241
作者: Ahmad Jad Allah,Kazi F. Akhter,Md. Kamrozzaman Bhuiyan,Manar D. Samad
类目: Machine Learning (cs.LG)
*备注:
Abstract:Computationally demanding and opaque deep learning models can be better understood and optimized by analyzing how they transform data. While deep transformers have been widely studied in computer vision and natural language processing, their application in tabular data remains relatively underexplored. This paper presents one of the first applications of an importance-scoring metric to interpret multi-head transformer models in learning from tabular data. Experiments conducted on 40 diverse tabular datasets demonstrate robustness to head drops based on the proposed head importance score. In 72.5% of experimental examples, the model remains most resilient to performance drops when heads with the lowest importance scores are gradually removed. In contrast, removing the most important attention head first results in the greatest reduction in classification performance. A closer look at individual head importance scores across six attention layers reveals that important heads are scattered across layers, with no consistent layer-specific trends. In contrast to the image and language domains, the importance of individual attention heads varies considerably across tabular datasets with different schemas and feature spaces. The proposed importance score can improve efficiency and redundancy within transformer architectures. We make the source code for measuring the importance of individual attention heads publicly available.
[LG-8] HALO: A Heterogeneity-Aware Language-Aligned IMU Foundation Model for Open-Set Human Activity Recognition
链接: https://arxiv.org/abs/2608.27233
作者: Zihan Ding,Liyu Zhang,Xiaomin Ouyang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Human Activity Recognition (HAR) using inertial measurement units (IMUs) enables a wide range of applications, yet the field still lacks a unified model that can generalize across diverse subjects, devices, and activities. Training such a model is difficult due to two key challenges: sensing heterogeneity – differences in sampling rates, channel configurations, and sensor placements – and poor generalization to unseen activities and label vocabularies. We introduce HALO (Heterogeneity-Aware Language-aligned Open-set model), a domain-specific IMU foundation model that addresses both challenges through a two-stage training framework. Stage 1 pretrains the IMU encoder with heterogeneity-aware self-supervised learning, including adaptive-pooling tokenization, channel-independent feature extraction, and contextualized sensor conditioning that injects natural-language sensor descriptions into each channel embedding. Stage 2 aligns this IMU encoder with text embeddings via synonym-aware soft contrastive learning, enabling open-set recognition via cosine-similarity retrieval without per-dataset classifiers. Trained on 10 public HAR datasets and evaluated on 7 held-out datasets, HALO outperforms five state-of-the-art baselines on all 8 aggregate metrics, and still leads on 3 of 4 settings under baseline-matched inputs. Despite using only ~35M trainable parameters – 10x fewer than the latest foundation model MOMENT (341.2M) – HALO improves zero-shot open-set accuracy, measured over all 87 training labels, by 13.7 percentage points. On two further datasets with severe distribution shift, every model including HALO collapses zero-shot. A video demonstration of HALO’s performance in real world is available at this https URL
[LG-9] Profit based evaluation of machine learning for nitrogen recommendations in winter wheat
链接: https://arxiv.org/abs/2608.27205
作者: Xulong Wang,Po Yang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Nitrogen rates for winter wheat are set before the season, under unknown prices and weather. The standard UK advice does not respond to prices, yet recent price swings moved the most profitable rate by tens of kilograms per hectare. Machine learning is often proposed as the fix. However, it is usually judged on prediction accuracy, and accurate prediction does not by itself make the recommended rate more profitable. Our insight is to score nitrogen advice directly by the profit it forgoes on measured yield response curves. We build a test bench on 892 such curves from two long running UK experiments, and sweep the nitrogen to grain price ratio to cover all price scenarios. On this bench, machine learning fails as a predictor. No model recovers the best rate within farm tolerance, and the benchmark noise shows none can. At normal prices, every model also loses to the standard advice on profit. The gain sits elsewhere. A simple correction step applied after the model cuts profit losses by a quarter, while better models and extra features give no gain. The same frozen correction cuts losses by 43% at the second site without any retraining. A hybrid of standard advice plus a damped correction removes bias and trims rare large losses. The same price sweep also prices emission cuts, at a cost comparable to current carbon prices. Machine learning therefore pays as a profit scored correction to standard advice, not as its replacement.
[LG-10] Common Geodesics Do Not Guarantee Fisher Consistency of the Structured SVM: Minimal Counterexamples and a Tree-Metric Classification
链接: https://arxiv.org/abs/2608.27203
作者: Jintao Fei,Jiangying Luo
类目: Machine Learning (cs.LG)
*备注: 14 pages, 1 figure
Abstract:A known necessary condition for Fisher consistency of the structured support vector machine requires the task loss to be a metric for which every output triple has a common geodesic point. We show that this condition is not sufficient for the canonical coordinate-wise argmax decoder. A four-output unit star admits an exactly optimal score vector whose maximizers are all strictly non-Bayes, and four outputs are minimal among metrics satisfying the condition. We then completely classify positively weighted tree metrics whose vertex set is the output space: argmax consistency holds if and only if the tree is a path. The failure on branching trees is confined to boundary distributions; every tree retains the argmax property at every full-support distribution. Among metrics satisfying the common-geodesic condition, five outputs are necessary and sufficient for a full-support counterexample; K_2,3 is the smallest member of an infinite K_m,n family. We additionally give a full-support counterexample for the three-dimensional Hamming cube. All optimality claims have exact primal-dual certificates. The counterexamples expose a concrete decoder gap: in this polyhedral setting, an embedding can guarantee the existence of a calibrated link without validating a prescribed argmax link on every surrogate-risk minimizer.
[LG-11] When Interference Graphs Evolve: Doubly Robust Estimation of Dynamic Peer Effects
链接: https://arxiv.org/abs/2608.27187
作者: Xiaojing Du
类目: Machine Learning (cs.LG); Social and Information Networks (cs.SI)
*备注: 8 pages, 4 figures
Abstract:Peer effects are difficult to estimate when interaction graphs evolve because pre-assignment network history, dynamic peer exposure, and post-assignment network change have distinct causal roles. We introduce a controlled contrast framework that indexes potential outcomes by own treatment, temporally aggregated peer exposure, and a post-assignment evolution summary. Differences between the resulting means define own-treatment, peer-exposure, controlled network-evolution, and joint controlled contrasts rather than a mediation decomposition. We develop the Dynamic Network Doubly Robust estimator, DynaNet-DR, which combines a temporally factorized propensity with normalized augmentation. Under consistency, summary sufficiency, sequential exchangeability, positivity, nuisance convergence, and weak dependence, its canonical estimator is consistent when either the outcome regression or the propensity estimator is consistent. The reported implementation adds representative-score prediction, fixed clipping, and finite-sample stabilization. Semi-synthetic benchmarks on fixed real temporal graph sequences show favorable estimation accuracy among methods targeting the full profile. These benchmarks assess summary-indexed contrasts rather than counterfactual edge generation, and the MathOverflow study is an observational illustration under the stated assumptions.
[LG-12] raceBench: Controlled Evaluation of LLM Agents for Time-Series Root-Cause Attribution
链接: https://arxiv.org/abs/2608.27182
作者: Tommaso Bendinelli,Artur Dox,Christian Holz
类目: Machine Learning (cs.LG)
*备注:
Abstract:LLM agents are increasingly applied to anomaly detection and root-cause analysis in time-series observations collected from real-world systems; however, their performance on these tasks has not been systematically evaluated under controlled conditions. We introduce TraceBench, a simulation-based framework for generating controlled root-cause attribution tasks. In each generated task, an agent receives time-series observations produced by simulating a physical dynamical system and must determine whether a system parameter was altered during the simulation and, if so, which one. Using TraceBench, we generate tasks from three interpretable mechanical systems and systematically evaluate four LLM agents across controlled experimental conditions, yielding new insights into how these agents analyze time-series observations from dynamical systems. Our results show that agents benefit substantially from domain context and explore data primarily through numerical console output rather than visualizations. We also find that agents generally perform worse when required to produce a Python script that maps each time-series sample to a predicted root-cause label than when they submit predictions directly. We release our datasets, agent trajectories, experimental results, and a leaderboard on our website, this http URL.
[LG-13] Diffusion Policies for Short-Horizon Planning in Robot Crowd Navigation
链接: https://arxiv.org/abs/2608.27158
作者: Wendong Li,Jochen Garcke
类目: Machine Learning (cs.LG)
*备注:
Abstract:Robot crowd navigation requires safe and efficient decision-making under dense, dynamic, and multimodal human–robot interactions. Existing reinforcement-learning methods typically output a single reactive action at each timestep, which limits their ability to represent diverse short-term avoidance strategies. We propose Planning Diffusion Policy Optimization (PDPO), an offline-to-online reinforcement-learning framework that uses a diffusion policy to generate short-horizon action chunks for crowd navigation. PDPO is first pretrained on collision-avoidance demonstrations and then fine-tuned online with PPO by treating the denoising process as an internal decision process. During execution, the policy generates a five-step action chunk and applies it in a receding-horizon manner. Furthermore, we observe an evaluation artifact in common crowd-navigation benchmarks: without explicit boundary constraints, learned agents may leave the valid domain and bypass dense crowds. To address this, we introduce a setting in which boundary violations are treated as collisions. Experiments show that PDPO obtains an improved success rate over strong baselines, and ablations demonstrate that action chunks are especially important for the modified bounded benchmark.
[LG-14] Inductive Correlation Clustering with Graph Neural Networks CIKM’26
链接: https://arxiv.org/abs/2608.27153
作者: Francesco Paolo Nerini,Francesco Bonchi,Arijit Khan,André Panisson
类目: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS)
*备注: Accepted at CIKM’26
Abstract:Correlation Clustering (CC) is a natural formulation of clustering in combinatorial optimization, which uses a graph representation of the input and does not require a pre-specified number of clusters. Given n objects and a pairwise similarity function, the goal is to cluster the objects so that similar objects are put in the same cluster and dissimilar objects are put in different clusters. Despite its versatility, existing CC algorithms suffer from significant scalability issues and are inherently transductive: i.e., the algorithm must be executed from scratch for any new problem instance. In this work, we bridge this gap by leveraging Graph Neural Networks (GNNs) to solve Inductive Correlation Clustering, a novel generalization of the CC problem designed to handle unseen graph instances. By learning to exploit common structural patterns and node features during training, our framework generalizes to new graphs drawn from the same distribution with minimal computational overhead with respect to standard algorithms. We demonstrate the effectiveness and scalability of our approach through extensive experiments. Our framework not only excels in the inductive setting, e.g., lowering the inference time up to 5 order of magnitude, while maintaining an approximation ratio within ~10% of the best baseline solution, but also achieves competitive results on standard (transductive) CC benchmarks. Finally, we showcase a practical application of our framework as a learnable pooling mechanism for graph classification. Our results indicate that our method serves as an efficient pooling layer, enhancing the ability of GNNs to capture hierarchical structural information in networks. Comments: Accepted at CIKM’26 Subjects: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS) Cite as: arXiv:2608.27153 [cs.LG] (or arXiv:2608.27153v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.27153 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-15] Ultra Low-Power Lightweight Probabilistic RSS-Based Path Reconstruction: A System for Landscape-Scale Bee Tracking
链接: https://arxiv.org/abs/2608.27152
作者: Christopher J. Noroozi,Joseph L. Woodgate,Michael Mangan,Michael T. Smith
类目: Machine Learning (cs.LG)
*备注:
Abstract:Applications in fields such as movement ecology, Internet of Things or robotics share the need for systems that localize devices that are too small and power constrained to implement GNSS (Global Navigation Satellite Systems). Alternative low-power localization methods often rely on only measurements of RSS (Received Signal Strength) to infer the AoA (Angle of Arrival) of a transmitted radio frequency signal, but are limited by range and the power demand of the large number of RSS measurements required to infer an accurate AoA. In this paper we address these issues with a novel RSS-based method for tracking ultra lightweight and low-power moving receivers across a complex landscape, achieved by using a minimal number of RSS measurements from simple rotating high-gain transmitters with a range of 300m, and applying probabilistic modelling to infer their AoA. The receiver’s movement path is then modelled using a Gaussian process and reconstructed using doubly stochastic variational inference, resulting in approximately 15m accuracy tracking of receivers weighing 38mg (including power source) over a scalable landscape range while consuming less than 180uW, increased to approximately 10m accuracy at less than 600uW by taking more RSS measurements. We anticipate that this method will support fields such as the behavioural study of flying insect species, which we demonstrate by applying the system to track Bombus terrestris nest return flights.
[LG-16] RACE-CRC: Trajectory-Adaptive Conformal Risk Control for Multi-Step Channel State Information Prediction
链接: https://arxiv.org/abs/2608.27124
作者: Kiarash Rezaei,Mehdi Sattari,Javad Aliakbari,Tommy Svensson,Paolo Monti,Carlos Natalino
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Published in Proceedings of Machine Learning Research (PMLR), volume 329, Conformal and Probabilistic Prediction with Applications (COPA 2026)
Abstract:Reliable prediction of time-varying channel state information (CSI) is essential for efficient wireless communication. Each CSI frame is a matrix-valued representation of the wireless channel response, and a sequence of CSI frames forms a temporal channel trajectory. Modern deep learning-based CSI predictors, however, often provide only point predictions and lack calibrated uncertainty estimates. This limitation is particularly problematic in multi-step CSI prediction, where the target is a sequence of future CSI matrices, and downstream decisions such as beamforming or scheduling may fail if any part of the predicted trajectory is unreliable. We propose trajectory-adaptive calibration and error profiling with conformal risk control (TRACE-CRC), a method for trajectory-aware uncertainty quantification in multi-step CSI prediction. TRACE-CRC constructs Frobenius-norm uncertainty balls around predicted CSI matrices and controls the risk that at least one future frame is uncovered. Instead of calibrating each future step independently, TRACE-CRC combines future-step-dependent error profiling, trajectory difficulty stratification, and learn-then-test (LTT) risk control. Empirically, TRACE-CRC achieves reliable trajectory-level coverage with substantially smaller uncertainty balls than conservative multi-step corrections, while avoiding the trajectory undercoverage of compact stepwise and adaptive conformal baselines.
[LG-17] Cone Extended Rayleigh Quotients for Directed Graph Learning: Minimax Spectral Certificates Sensitivity and Adaptive Control
链接: https://arxiv.org/abs/2608.27122
作者: Yavdat Sh. Il’yasov,Nur F. Valeev
类目: Machine Learning (cs.LG)
*备注: 23 pages, 3 figures
Abstract:Directed graph learning naturally leads to trainable nonsymmetric propagation operators with distinct right and left spectral structures. Building on the two-sided cone Rayleigh framework for generalized pencils [ B_\theta-\lambda G, ] we develop a learning-oriented methodology for spectral certification, sensitivity analysis, and control without requiring symmetry, nonnegativity, or cone preservation. In the positive-orthant setting, computable lower and upper cone bounds provide an a posteriori enclosure of a distinguished cone level, while smooth soft-min/max surrogates preserve rigorous one-sided bounds with explicit approximation errors and remain differentiable with respect to the trainable parameters. For a simple interior level, the right and left modes satisfy [ D\lambda_C(B)[H]=v_C^T H u_C, ] yielding first-order optimal graph-supported interventions under prescribed perturbation budgets and motivating adaptive spectral control. Numerical experiments demonstrate the applicability of the approach beyond cone-preserving operators and in directed learning settings. Signed nonsymmetric perturbations reveal a transition from interior eigenpairs to boundary complementary quasi-pairs, including non-spectral cone levels, while controlled experiments show that symmetrization can remove predictive information carried solely by edge direction. On the directed Cora citation network, adaptive recomputation of the right–left sensitivity reduces the distinguished spectral level by approximately 21.5% under a cumulative edge-weight reduction budget of 0.5% , with no observed change in test accuracy for the trained model and data split considered. Comments: 23 pages, 3 figures Subjects: Machine Learning (cs.LG) MSC classes: 68T07, 68T05, 65F15, 90C31 ACMclasses: I.2.6; G.1.3; G.1.6; G.2.2 Cite as: arXiv:2608.27122 [cs.LG] (or arXiv:2608.27122v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.27122 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-18] SecureDrive-FL: Joint Differential Privacy and Gradient-Aware Selective Homomorphic Encryption for Federated Driver Monitoring
链接: https://arxiv.org/abs/2608.27108
作者: Baran Can Gül,Hanuma Siddhartha Tunuguntla,Anjana Arvind Naik,Abhishek Vijay Potekar,Nasser Jazdi,Michael Weyrich
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Federated Learning (FL) enables privacy-aware distributed training, yet gradient updates remain exploitable: Man-in-the-Middle (MitM) interception exposes updates in transit, while model poisoning corrupts global convergence. We first introduce GASHE (Gradient-Aware Selective Homomorphic Encryption), a novel selective encryption strategy that dynamically identifies and encrypts only the gradient components exceeding a DP-calibrated sensitivity threshold, rather than encrypting all parameters uniformly as in static layer-based or full-parameter CKKS schemes. Building on GASHE, we introduce SecureDrive-FL, a federated driver monitoring framework that couples DP-SGD with GASHE to create the first closed-loop DP+HE privacy pipeline: DP-SGD calibration parameters directly derive the GASHE encryption mask, unifying training-time privacy and communication-time confidentiality. Evaluated on a ten-class distracted driver classification task under non-IID federated splits, SecureDrive-FL matches DP-SGD alone’s poisoning resistance (73.6% vs. 74.0% accuracy, 3.9% Attack Success Rate for both) while additionally withstanding MitM interception, where DP-SGD alone collapses to near-random accuracy (78.2% vs. 10.4%), all under only approx. 8–10% additional runtime overhead relative to DP-SGD alone—under DP-SGD noise injection with per-round privacy parameter epsilon_0=4.
[LG-19] abular Deep Learning for Algorithmic Trading: Cross-Regime Bayesian Optimisation for Equity Signal Generation
链接: https://arxiv.org/abs/2608.27076
作者: Joshua Le Grice
类目: Machine Learning (cs.LG); Computational Finance (q-fin.CP); Trading and Market Microstructure (q-fin.TR)
*备注:
Abstract:Algorithmic trading now represents a market exceeding 20 billion, where even marginal gains in signal robustness can translate into economically significant returns. Existing evaluations of equity prediction models do not explicitly target regime robustness during hyperparameter selection. Five model classes are trained on daily observations from approximately 300 large-cap US equities over eleven years, with Bayesian optimisation configured to target trading performance across three statistically different market regimes. Regime-robust hyperparameter selection is associated with out-of-sample generalisation, as signal precision remains above the random baseline across all four quarters of the test period, and portfolio performance slowly degrades under simulated input noise before collapsing beyond a defined threshold. No individual tabular deep learning architecture outperforms gradient-boosted trees, but combining XGBoost and TabNet using rank aggregation produces a Hybrid ensemble with an annualised return of 51.26%, a Sharpe ratio of 2.44, and a statistically significant CAPM alpha of 0.423 (p = 0.011). A near-zero beta indicates this outperformance is driven by stock selection, not market exposure. Alternative data plays a secondary role once technical and fundamental features are accounted for, as well as contributing more strongly on the short side than the long, and varies by model class. An interactive application makes these results explorable in real time, with live data integration the remaining step toward practical deployment.
[LG-20] Disentangling Optimization Scale from Preference Scale in DPO
链接: https://arxiv.org/abs/2608.27032
作者: Ivan Kruzhilov
类目: Machine Learning (cs.LG)
*备注:
Abstract:Direct Preference Optimization (DPO) is a widely used objective for aligning language models from preference data, with the coefficient \beta commonly interpreted as controlling the KL constraint to a reference policy. We show that \beta entangles two distinct roles: it governs the effective inverse preference-noise scale and simultaneously rescales the optimization dynamics, coupling this scale with the effective step size. As a consequence, at a fixed learning rate the achieved policy deviation is non-monotone in \beta : it vanishes in a dead zone at small \beta , reaches a peak at an intermediate value, and decreases again for larger \beta . Moreover, standard DPO loss values are not comparable across \beta : runs with nearly identical loss curves can differ several-fold in KL divergence from the reference model. This entanglement obscures the role of \beta , increases sensitivity to hyperparameter choices, and complicates learning-rate scheduling. We propose a centered-softplus reformulation that is argmin-equivalent to DPO for \beta0 , while making the inverse preference-noise-scale and learning-rate effects explicit and independently tunable. The normalized centered-softplus objective also admits a continuous \beta\to0 endpoint that reduces to a linear preference-margin objective.
[LG-21] FoldPipe: Bounded Remote Streaming of Native Molecular Shards with Asynchronous Prefetch
链接: https://arxiv.org/abs/2608.27029
作者: Dhiren Mukesh Khatri
类目: Performance (cs.PF); Machine Learning (cs.LG)
*备注: 4 pages, 2 figures. Source code and reproducibility artifacts: this https URL
Abstract:Training molecular machine-learning models on ephemeral or memory-constrained accelerator instances can require repeatedly retrieving preprocessed molecular graphs from remote storage. FoldPipe is a lightweight Python orchestration layer for already-sharded PyTorch and PyTorch Geometric data. It retrieves one shard ahead in a background thread while the consumer trains on the current shard, keeping the number of live shard payloads bounded with respect to total dataset size. Asynchronous prefetch and bounded buffering are established systems techniques rather than novel scheduling algorithms. FoldPipe’s contribution is a small integration targeted at native .pt molecular shards together with a source-pinned empirical characterization of its operating regime. We evaluate a SchNet energy-and-force workload on MD17 aspirin using 20 paired, order-alternating benchmark passes on a Tesla T4. Each pass processes five pinned shards containing 25,000 structures. FoldPipe records 16.33 s mean I/O-compute overlap, compared with zero by construction for the sequential bounded baseline. Mean pass time is 76.78 s for FoldPipe and 83.37 s for the baseline. However, the geometric mean paired speedup is 1.059\times with a 95% bootstrap interval from 0.878\times to 1.288\times . The experiment therefore verifies the overlap mechanism but is inconclusive about a reliable wall-clock speed advantage under the observed public-network variability. Comments: 4 pages, 2 figures. Source code and reproducibility artifacts: this https URL Subjects: Performance (cs.PF); Machine Learning (cs.LG) Cite as: arXiv:2608.27029 [cs.PF] (or arXiv:2608.27029v1 [cs.PF] for this version) https://doi.org/10.48550/arXiv.2608.27029 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-22] Benchmarking_Fast_Domain_Adaptation_for_Unsupervised_Speech_Units
链接: https://arxiv.org/abs/2608.26992
作者: Robin San Roman,Manel Khentout,Tu Anh Nguyen,Paul Michel,Yossi Adi,Emmanuel Dupoux
类目: Machine Learning (cs.LG)
*备注:
Abstract:Representation learning has attracted great atten- tion and managed to reach good performances as a pretraining method for downstream tasks or as a first step towards unsu- pervised speech modeling. Yet, little is known about how such methods deal with out-of-domain speech and how could they be adapted in a few shot to new domains. This is important especially for accented speech where one observes a long tail of accents that diverge from the standard ones. We introduce ABX- Accent, a benchmark based on the AESRC dataset that features 10 different accents of English. It includes a small ( 10 hours) unlabelled training set in each of the accents and adaptations of the Zero Resources Challenge ABX evaluation metrics to each of the accents. We illustrate this benchmark with a baseline model that uses adaptive domain normalization to fine tune a pretrained Contrastive Predictive Coding model on the accents. This method is first developed on LibriSpeech using a male/female split. When applied to the new benchmark, the proposed method yields a relative improvement of 23.6% on across-speaker ABX scores on average compared to non adapted models. The data and metrics will be open sourced upon paper acceptance
[LG-23] Decentralized Multitask Learning over Learned Task Graphs
链接: https://arxiv.org/abs/2608.26989
作者: Zirui Wan,Stefan Vlaski
类目: Machine Learning (cs.LG); Signal Processing (eess.SP); Systems and Control (eess.SY)
*备注:
Abstract:This paper investigates decentralized multitask learning over networks when the underlying task relationships are unknown. While existing graph-regularized multitask frameworks typically assume a known structure, practical settings often require learning inter-task dependencies directly from distributed data. We propose a decentralized two-phase strategy that first estimates a generalized graph Laplacian from noisy non-cooperative stochastic gradient iterates, and subsequently exploits the learned graph to enable cooperative multitask diffusion learning. This framework is motivated by a Gaussian Markov random field prior, which gives rise to a decentralized maximum likelihood estimator for the graph Laplacian. The analysis quantifies the Laplacian estimation error and its propagation to the steady-state performance of the multitask diffusion recursion, and introduces a topology sensitivity index to capture the effect of network heterogeneity. Simulation results corroborate the theoretical findings and demonstrate that cooperation enabled by the learned task graph significantly improves performance over non-cooperative learning, while approaching the true-graph baseline when the estimation stepsize is sufficiently small.
[LG-24] Graph-Based Pseudo-multimodal Contrastive Learning for 12-Lead ECG Representations
链接: https://arxiv.org/abs/2608.26964
作者: Mengyu Wang,Kozo Okada,Takafumi Goto,Natsuko Jinba,Hiroki Yamaya,Kiyoshi Hibi,Tomoki Hamagami
类目: Machine Learning (cs.LG)
*备注: 7 pages, 7 figures. Presented at the 48th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2026)
Abstract:12-lead electrocardiogram (ECG) is a standard, non-invasive examination widely used for diagnosing coronary artery disease, where clinical interpretation relies on comparing waveform patterns across multiple leads. However, most existing ECG analysis methods focus on single-lead signals or treat each lead independently, and typically process ECG signals as one-dimensional time-series data using CNNs or RNNs. While effective in modeling local waveform changes, such approaches have difficulty capturing inter-lead dependency and global waveform patterns essential for clinical diagnosis. To address this limitation, we propose a graph-based pseudo-multimodal contrastive learning framework called Graph-CMMC. ECG waveforms are transformed into Gramian Angular Difference Field (GADF) images to construct complementary representations of the same cardiac activity, enabling a pseudo-multimodal learning setting. Using all 12 leads, Graph-CMMC aligns waveform and GADF representations in a self-supervised manner, while a graph-based relational module is employed to model inter-lead dependency and enforce structural consistency across leads during contrastive learning. Experimental results on a multi-label coronary artery occlusion classification task demonstrate that the proposed framework achieves competitive performance compared to supervised learning methods. These results further suggest the effectiveness of using GADF as a complementary representation and incorporating explicit graph-based modeling of inter-lead dependency for learning robust 12-lead ECG representations. Comments: 7 pages, 7 figures. Presented at the 48th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2026) Subjects: Machine Learning (cs.LG) Cite as: arXiv:2608.26964 [cs.LG] (or arXiv:2608.26964v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.26964 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-25] Adversarial Training Without Input Gradients via Low-Rank Householder Expansions
链接: https://arxiv.org/abs/2608.26963
作者: Tiana C. Johnson,Donsub Rim
类目: Machine Learning (cs.LG)
*备注: 25 pages, 7 figures, 4 tables
Abstract:This work concerns adversarial training against the small-norm adversarial examples that arise from the inherent input instability of a trained deep neural network. Examples in this class are small as measured in the relative \ell^2 -norm, and therefore lie in the neighborhood of the input on which the model acts approximately linearly, the regime in which the perturbation remains imperceptible. We first show that such examples can be computed directly from the trained network parameters, without input gradient iterations, by means of a linearization called the low-rank Householder expansion (LRHE). The expansion describes the composed affine map rather than any individual layer, and the directions it identifies are read from the activation pattern already available in the forward pass. We then propose a simple adversarial training scheme built on this construction. No differentiation with respect to the input is performed at any point: training requires only additional forward evaluations, with weight parameters updated by the standard backward pass, and the inner maximization of the usual min-max formulation is eliminated entirely. That such a regularizer exists is our main finding: the methods that dispense with the inner search all obtain their local geometry by differentiating with respect to the input, and we show this is not necessary. The regularizer costs the equivalent of 2.8 PGD steps per epoch, an 8.7\times reduction relative to 40-step adversarial training on MNIST and below the cost of 3-step training. The resulting models match three-step PGD adversarial training for relative \ell^2 budgets \varepsilon \le 0.02 and 40-step training for \varepsilon \le 0.012 , falling away beyond, consistent with the locality of the expansion.
[LG-26] Packora: Systematic Design for Generative Molecular Crystal Structure Prediction
链接: https://arxiv.org/abs/2608.26962
作者: Nayoung Kim,Kiyoung Seong,Sungsoo Ahn
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci)
*备注: 37 pages, 21 figures
Abstract:Molecular crystal structure prediction (CSP) is important in pharmaceuticals, agrochemicals, and organic electronics, where subtle differences in molecular conformation and packing can strongly affect material properties. We present Packora, a flow-based generative model for molecular CSP that jointly predicts atomic coordinates and the lattice from molecular graphs. Packora supports multi-component and organometallic crystals and can condition on any subset of molecular conformers, stereochemical labels, and space-group information within a single model. Inspired by the CCDC CSP blind test, we evaluate generation and ranking separately, using generation to isolate generator quality and ranking to measure end-to-end performance under a common relaxation and ranking pipeline. We also systematically study architecture, training, conditioning, inference, and scaling, identifying an effective design based on cacheable pairwise reasoning, training objective and numerical solver choices, conditioning dropout, and balanced scaling of pairwise and single representations. Packora outperforms the baselines on both structure generation and ranking benchmarks, achieving the best matched-budget coverage across all six generation benchmarks, as well as higher experimental-form recovery, lower experimental-form ranks, and faster convergence in ranking.
[LG-27] Gromov-Monge Flow Matching for Equivariant Graph Generation ATC
链接: https://arxiv.org/abs/2608.26961
作者: Moritz Piening,Christian Wald
类目: Machine Learning (cs.LG); Optimization and Control (math.OC); Machine Learning (stat.ML)
*备注: Gromov-Wasserstein, Optimal Transport, Flow Matching, Graph Generation, Equivariant Networks
Abstract:Graphs are invariant under node permutations, motivating the use of permutation-equivariant architectures in generative models. In flow matching, however, symmetry may also enter the source–target coupling: once graph pairs are compared up to node relabeling, the natural Wasserstein geometry is that of the graph quotient space. The Euclidean quotient metric of this space coincides with the Gromov–Monge distance, obtained by optimally relabeling the nodes. We develop this perspective theoretically, showing that quotient couplings can be lifted to aligned representatives without additional cost and that symmetrization yields equivariant flow-matching minimizers, including for categorical endpoint prediction. In practice, exact Gromov–Monge alignment is intractable, so we construct minibatch couplings using efficient Gromov–Wasserstein-type relaxations and lower bounds for the inner node alignment, optionally combined with an outer assignment between graphs. The resulting procedure changes only the training coupling and is compatible with standard permutation-equivariant architectures. Across continuous graph and categorical molecular generation, these structure-aware couplings substantially improve sample quality at small integration budgets, while our scaled-up molecular models remain competitive under conventional many-step sampling.
[LG-28] Data-driven Koopman mode approximation: A neural power iteration algorithm
链接: https://arxiv.org/abs/2608.26943
作者: Guillaume O. Berger,Raphaël M. Jungers
类目: ystems and Control (eess.SY); Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:This paper proposes a novel data-driven algorithm to approximate the dominant eigenfunctions (aka.~modes) of the Koopman operator of nonlinear dynamical systems using neural networks. The relevance of learning the dominant Koopman modes is to approximate nonlinear dynamics by linear ones in a lifted space, thereby enabling simplified control and analysis. To fight the curse of dimensionality arising from using expressive templates (here neural networks) for the mode approximation, the proposed method leverages a power-iteration scheme that directly learns the dominant Koopman modes without explicitly constructing the projection of the Koopman operator on the template of functions. Our approach connects to other approaches in the literature that avoid the curse of dimensionality by learning small dictionaries of functions, but differs from them in that we do not require ``anti-collapse mechanisms’’ to ensure that the learned dictionary is expressive enough to approximate the Koopman operator since our power-iteration scheme is designed to converge toward the dominant modes of the projected Koopman operator. The approach is fully data-driven, requiring only sampled state transitions. Theoretical guarantees are provided, showing convergence under increasing sample size and network width (in connection with the neural tangent kernel theorem). Numerical experiments demonstrate that the method achieves accurate and smooth approximations of dominant modes while avoiding the limitations of traditional techniques such as extended dynamic mode decomposition.
[LG-29] A Layer Importance Metric for Quantization Accounting for the Speed-Quality Trade-off in Autoregressive Models
链接: https://arxiv.org/abs/2608.26926
作者: Artem Safronov
类目: Machine Learning (cs.LG)
*备注: 25 pages, 11 figures
Abstract:Small language models (sLLMs) are nowadays hosted on devices with limited memory and computational budget. In an autoregressive setup, inference is memory-bandwidth bound: uniform quantization is often detrimental to such models, since their architecture has limited redundancies and only a few layers are not very sensitive to lower precision. We propose a composite metric that combines two orthogonal criteria: information retention (measured in terms of a normalized SQNR-based coefficient) and throughput gains (modeled using a roofline-based latency analysis). By profiling Gemma 3 1B, we find that Feed-Forward Network blocks and the embedding matrix are the most promising targets for acceleration. For each candidate, we estimate a normalized quality score based on simulated quantization and a normalized speed score based on roofline modeling with no actual execution needed. We combine the two scores in a composite priority coefficient, allowing us to tune the trade-off between speed and quality as needed. Our metric is general and can be used to prioritize individual blocks, their projection sublayers, or transformer layers as a whole. We evaluate our approach on several model architectures, showing that our estimates have at around 4% prediction error for the accelerated speedup. We find that our method generally allocates more resources to the most expressive layers compared to evolutionary search, specialized accelerators, or Shapley-value-based approaches that require expensive approximate inference. Our analytical approach makes sLLM quantization a predictable engineering task.
[LG-30] Mitigating Strong-Modality Collapse in Multimodal Learning via Inverted Asymmetric Fusion
链接: https://arxiv.org/abs/2608.26879
作者: Mary Ogbuka Kenneth,Foaad Khosmood,Abbas Edalat
类目: Machine Learning (cs.LG); Multimedia (cs.MM)
*备注:
Abstract:Fusing multiple modalities is expected to improve model performance. However, on the MultiHuSE dataset, early, late, and symmetric attention fusion often fail to outperform the best unimodal baseline (text). Pathway isolation of a symmetric attention fusion model reveals that the text-pathway accuracy drops from 74.9% to 56.4% after fusion in one such setting, indicating that the dominant modality can be degraded during integration. We term this strong-modality collapse and argue that it helps explain why some multimodal models fail to surpass unimodal baselines. We propose Inverted Asymmetric Fusion (IAF), which avoids forcing mutual attention across modalities. The dominant modality is preserved by passing through fusion unchanged, while weaker modalities attend to it as a contextual anchor. Before fusion, weaker modalities are strengthened using Modality-Aware Knowledge Distillation. We evaluate IAF on three benchmarks with different modality hierarchies: text-dominant datasets (MultiHuSE, UR-FUNNY) and an audio-visual-dominant dataset (MUStARD). Pathway isolation shows that IAF preserves the dominant modality’s internal accuracy at its unimodal ceiling across all tested configurations, whereas symmetric fusion degrades it by up to 18.5% on MultiHuSE. IAF improves over the strongest unimodal baseline by up to 8.25%.
[LG-31] When Is the Sharp Covariance Envelope Tight? Feature-Only Geometry for Volume-Sampled Least Squares
链接: https://arxiv.org/abs/2608.26877
作者: Kihun Rhee
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 58 pages, 4 figures
Abstract:Prior analyses by Derezinski and Warmuth established all-size sampling identities, selected-OLS unbiasedness, and inverse moments for ordinary volume sampling, while their exact arbitrary-fixed-response loss and prediction-covariance formulas are at the rank-size endpoint s=d. We establish a Loewner envelope for centered coefficient covariance for every full-rank fixed pool, response, and legal budget d = s = m under ordinary indexed fixed-size volume sampling followed by selected unweighted least squares; its coefficient is globally sharp over the full-rank class. Global sharpness does not determine attainability on the pool in hand. Under positive loss, strict-interior budgets, and no coloops, a feature-only margin nu_A gives the exact fixed-design spectral phase: nu_A 0 if and only if the normalized spectral envelope is strict for every compatible residual, whereas nu_A = 0 if and only if some compatible residual is spectrally tight; the same zero-margin residual is tight at every strict-interior budget. A residual-augmented change of measure supplies the response-aware mechanism and a one-sided quantitative slack bound, while support saturation proves the attainment direction. Critical equal-leverage geometry interprets the boundary, and sound lower certificates yield conservative same-primitive cardinality decisions. Frozen-feature examples show that the certificate is nonvacuous and measure the fixed-pool cost of its authorized reduction. The claims concern conditional centered, full-Gram-whitened coefficient covariance, not population generalization.
[LG-32] On the Indistinguishability of Human v/s AI Generated Text
链接: https://arxiv.org/abs/2608.26797
作者: Jaee Ponde,Aritra Das,Mihir More,Debayan Gupta
类目: Machine Learning (cs.LG)
*备注: 11 pages
Abstract:The rapid improvement of LLMs has made distinguishing AI-generated text from human writing a pressing problem. This challenge is further amplified by paraphrasing tools designed to make machine-generated text appear more “human”. We study how access to human writing samples can be used to strategically paraphrase machine-generated responses toward the human distribution. Under a multi-sample setting with human and machine responses to the same prompts, we show that repeated paraphrasing moves the machine distribution toward the empirical human distribution under simple mixing and stability conditions. Our results derive an explicit convergence rate, extend the analysis to a finite-sample setting, and characterize how the required number of human samples and paraphrasing rounds scale with the desired error.
[LG-33] Beyond Client Averag ing: A Client-Independent Second-Order Stationary-Bias Component in Stochastic SCAFFOLD
链接: https://arxiv.org/abs/2608.26765
作者: Yi-Ping Tang,Guan-Ju Peng
类目: Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注:
Abstract:Existing constant-step analysis of stochastic \Scaf identifies a leading O(\gamma/N) stationary mean bias and shows that higher-order bias can persist as the client count increases, but does not identify the first client-independent contribution at coefficient level. For full-participation stochastic \Scaf with one-dimensional homogeneous clients, fixed local-step count H , and bounded additive gradient noise, we prove, uniformly over N\ge2 , \beginaligned \mathbbE_\pi_\gamma,N,H[x]-x^\star = -\fracf’‘’(x^\star)\sigma^24f’‘(x^\star)^2\frac\gammaN\ - \fracf’‘’(x^\star)\sigma^212f’'(x^\star) \frac(H-1)(5H-1)H\gamma^2 +O_H!\left(\frac\gamma^2N+\gamma^3\right). \endaligned Hence client averaging suppresses the leading O(\gamma/N) bias but does not remove the client-independent O(\gamma^2) component when its coefficient is nonzero. The mechanism is indirect: although the direct control contribution cancels pathwise in the linear global average, the controls still alter within-round local trajectories and their second moments. Fresh gradient noise and persistent control fluctuations therefore generate local second-moment corrections that nonquadratic curvature converts into stationary mean bias. The coefficient vanishes for quadratic objectives. Numerical experiments are consistent with the predicted coefficient, its persistence as client count increases, and the stated joint remainder. The result is restricted to the one-dimensional homogeneous fixed- H setting. Subjects: Machine Learning (cs.LG); Statistics Theory (math.ST) Cite as: arXiv:2608.26765 [cs.LG] (or arXiv:2608.26765v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.26765 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Guan-Ju Peng [view email] [v1] Thu, 27 Aug 2026 07:57:26 UTC (745 KB)
[LG-34] Safety by Design: Realized-Cost Constraints for Contextual Bandits with Continuous Actions
链接: https://arxiv.org/abs/2608.26755
作者: Spyros Dragazis,Aldo Pacchiano
类目: Machine Learning (cs.LG)
*备注:
Abstract:Contextual bandits are a standard framework for sequential decision-making under uncertainty, with applications in clinical trials, dosage selection, recommendation systems, and autonomous systems. Safety is central in many of these applications, since a single unsafe decision in settings such as dosage selection or autonomous driving can have catastrophic consequences. A common way to model safety in bandit problems is to associate each action with both a reward signal and a cost signal, and to optimize reward subject to constraints on cost. Most existing safety-constrained bandit models enforce safety by requiring the expected cost of each action to remain below a prescribed threshold. However, this may be insufficient in heteroscedastic settings, where the chosen action affects not only the expected reward and cost, but also the variability of the observed outcomes. We study contextual bandits with one-dimensional continuous actions and stage-wise high-probability constraints on the realized cost. We propose High-Probability Constrained UCB, an optimistic-pessimistic algorithm that explores for reward while conservatively estimating the safe action set. For linear reward and cost models, we prove a tight \tilde\mathcalO(d\sqrtT) regret bound, and we extend the analysis to general function classes using the eluder dimension. Experiments show that enforcing realized-cost safety substantially reduces violations compared with expected-cost constrained baselines.
[LG-35] Self-Augmented Diffusion Guidance for Physics-Informed Generation
链接: https://arxiv.org/abs/2608.26748
作者: Akira Osaka,Naoya Takeishi,Takehisa Yairi
类目: Machine Learning (cs.LG)
*备注:
Abstract:Diffusion models can be used to generate spatiotemporal signals of physical phenomena, such as time-series images of fluid dynamics. However, a major limitation of standard diffusion models is that they do not incorporate constraints derived from the underlying physical laws. Consequently, generated samples may appear visually plausible while deviating substantially from the true dynamics. In this study, we propose a simple yet effective physics-informed approach based on diffusion guidance with self-generated data augmentation. The proposed method learns the data distribution conditioned on the degree of deviation from the physically correct dynamics and generates samples by explicitly setting the deviation condition to be zero. The method decouples the evaluation of the governing equations from the diffusion model training and sampling processes, avoiding the need to solve the governing equations at every iteration of the denoising process. This design makes the method applicable to problems requiring computationally expensive numerical simulations and enables faster sample generation. Experimental results demonstrate that the proposed model not only significantly reduces the deviations compared with standard diffusion models but also achieves further reductions when combined with existing physics-constrained diffusion methods.
[LG-36] Neural Regression with Embeddings for Numerical Attribute Prediction in Knowledge Graphs
链接: https://arxiv.org/abs/2608.26729
作者: Rupesh Sapkota,Louis Mozart Kamdem Teyou,Moshood Yekini,Caglar Demir,Axel-Cyrille Ngonga Ngomo
类目: Machine Learning (cs.LG)
*备注:
Abstract:In recent years, transductive knowledge graph embedding models have been applied to tasks such as link prediction and query answering. Although knowledge graphs often contain rich numerical attributes, most embedding models neglect them, limiting their ability to represent real-world knowledge graphs with diverse information. In this work, we propose a neural regression model (LitEm) that enables transductive knowledge graph embedding models to predict numerical attributes within knowledge graphs. Experimental results demonstrate that LitEm achieves the best or second-best results on most attributes across FB15K-237, YAGO15K, DB15K, and Mutagenesis. Furthermore, we propose a co-training framework that jointly trains state-of-the-art transductive knowledge graph embedding models with LitEm, which improves link prediction performance mainly for bilinear models and simultaneously enables them to predict numerical attributes. In addition, the literal-awareness evaluation demonstrates that co-training helps models to encode and exploit attribute information in a "literal-aware’’ manner, suggesting that the observed gains are not merely due to additional parameters. We publicly release our implementation at this https URL.
[LG-37] Simple Actors and Deep Critics for Scalable Reinforcement Learning CIKM2026
链接: https://arxiv.org/abs/2608.26659
作者: Guhyeon Kang,Jaehwi Lee,Minhae Kwon
类目: Machine Learning (cs.LG)
*备注: Accepted at CIKM 2026
Abstract:Recent progress in offline reinforcement learning (RL) has been driven by expressive generative actors such as diffusion and flow-matching policies, which capture multimodal behavior in offline datasets. However, these actors require multiple denoising or integration steps per action and thus incur substantial overhead at every decision in deployment. In this work, we revisit where capacity should be invested in an offline actor–critic method. Since the critic is used only during training and is discarded at deployment while the actor runs at every decision step, allocating capacity to the critic rather than the actor is more favorable for inference-time efficiency. However, scaling MLP critics in offline RL is known to introduce several distinct instabilities that have, in practice, kept critics shallow. We identify three distinct failure modes that arise when critics are deepened in offline RL—optimization, bootstrap-noise amplification, and value-range drift—and address each with a corresponding ingredient: a residual MLP backbone, n-step bootstrap targets, and a categorical cross-entropy loss. Combining these ingredients with a lightweight deterministic actor, we propose LAC (Light Actor, deep Critic). On OGBench, LAC matches the strongest diffusion- and flow-matching baselines while achieving up to 4x lower inference latency, comparable to one-step distilled policies without distillation. Its critic recipe also transfers across actor parametrizations.
[LG-38] When Privacy Hurts Mergeability: Geometry-Aware Model Merging under Differential Privacy
链接: https://arxiv.org/abs/2608.26655
作者: Jin Liu,Junkang Liu,Ning Xi,Yinbin Miao,Dawei Wei,Ke Cheng,Jianfeng Ma
类目: Machine Learning (cs.LG)
*备注:
Abstract:Model merging promises to construct a single multi-task model from independently fine-tuned task models without accessing the original task data. This makes it attractive when task data cannot be centralized, but released task models may still leak private fine-tuning data. Differential privacy (DP) provides a principled mechanism for limiting such leakage, yet its effect on model merging remains poorly understood. In this paper, we study the geometry of differentially private model merging and identify two geometric obstacles that make private task models difficult to merge: \emphlocal sharpness, which makes task losses sensitive to the parameter displacement induced by merging, and \emphreference drift, which measures the displacement of private task models from the shared pretrained initialization and amplifies cross-task interference. Based on these observations, we propose \textbfDP-Merging, a geometry-aware framework that improves the mergeability of differentially private task models. DP-Merging uses a DP-compatible sharpness-aware objective to guide each private task model toward flatter loss regions, and a reference-based alignment regularizer to keep task models close to the shared pretrained initialization. We derive a merge-gap upper bound showing that reducing local curvature and reference drift tightens the bound on the loss increase induced by merging. Experiments on vision and language tasks across multiple privacy budgets show that DP-Merging consistently improves private merged-model performance while preserving the privacy guarantees of the underlying DP fine-tuning procedures.
[LG-39] Robust Neural Stimulation Response Modeling Through Meta-Learning and Pretraining
链接: https://arxiv.org/abs/2608.26649
作者: Matthew J Bryan,Daniel C Muir,Felix Schwock,Azadeh Yazdan-Shahmorad,Rajesh P N Rao
类目: Machine Learning (cs.LG)
*备注: Open source code available: this http URL . 15 pages, 10 figures
Abstract:Objective: Model-based closed-loop neural stimulation holds promise for therapeutic applications ranging from Parkinson’s disease to sensory restoration, but deployment has been limited by two obstacles: 1) forecasting models for predicting the consequences of stimulation fail catastrophically on a meaningful fraction of sessions, and 2) per-session calibration requirements are often incompatible with clinical constraints. We address both by demonstrating, for the first time, that meta-learning and pretraining can be applied to neural stimulation response modeling. Methods: Temporal basis function models (TBFMs) forecast state-dependent neural responses to stimulation. We extend TBFMs with cross-session pretraining using a novel architecture and algorithm based on model-agnostic meta-learning (MAML), evaluating them on 40 sessions of optogenetic stimulation in primary sensorimotor cortex of two non-human primates. Results: Meta-learning substantially reduces catastrophic forecast failure: for a 1k calibration set size, sessions with test R-squared 0.05 drop from 16 of 40 (single-session training) to 1 (MAML-pretrained), and prediction intervals become significantly narrower (p 0.05). Calibration requirements are reduced by 50-90% at matched accuracy, enabling experiments otherwise infeasible within clinical session-time constraints. Conclusion: Our results demonstrate that cross-session structure in stimulation responses is consistent enough to support pretraining, providing the first empirical evidence that meta-learning approaches are viable for neural stimulation. Significance: The robustness and sample efficiency gains directly address known obstacles to deploying model-based stimulation controllers. Our results motivate community efforts to assemble standardized multi-site stimulation datasets and to further explore meta-learning for robust closed-loop stimulation.
[LG-40] chnical Comparative Benchmarking Study: Advanced AI Hybrid Methods for Renewable Energy Farm Optimization and Forecasting
链接: https://arxiv.org/abs/2608.26613
作者: Majid Masoumi,Asghar Dashtiy,Mohammad Dehghan,Mina Rajabi
类目: Machine Learning (cs.LG)
*备注:
Abstract:This study provides a comprehensive benchmarking of conventional machine learning (ML), ensemble learning, deep neural networks, recurrent architectures, Transformers, graph based models, and hybrid ensemble deep learning approaches under complementary renewable energy scenarios. Three datasets are considered: a large scale WEC dataset, a 16 WEC dataset, and operational 10 min SCADA measurements at the Penmanshiel wind farm. For structured WEC layout data, tree ensembles exhibited a clear advantage over conventional ML and neural predictors because randomized partitioning and boosting efficiently captured nonlinear layout power interactions without requiring explicit feature representation learning. The Extra Trees was the strongest model, achieving considerable results. Relative to the MLP baseline, this corresponds to an approximately 63.7% reduction in MAE, demonstrating the suitability of randomized tree ensembles for high dimensional structured WEC data. Also, STGCN reduced the MAE to approximately 167.0 kW and achieved R = 0.93 by explicitly learning spatial and temporal turbine interactions. The best overall forecasting accuracy was obtained by the RF BiLSTM hybrid, with an MAE=150.5 kW. Compared with standalone LSTM, this represents an approximately 75% reduction in MAE, while improving on STGCN by approximately 10.0%. Finally, the experiments reveal that no single AI architecture is universally optimal: randomized and boosted ensembles are particularly effective for structured WEC surrogate modeling, graph networks become advantageous when explicit spatial interactions dominate, and ensemble recurrent hybrids provide the strongest balance when nonlinear tabular relationships and temporal dynamics coexist.
[LG-41] SimCast-S2S: An Efficient Generative Model for Subseasonal Precipitation Forecasting via Transfer Learning from Climate Simulations
链接: https://arxiv.org/abs/2608.26594
作者: Hiep V. Dang,Antonios Mamalakis
类目: Machine Learning (cs.LG)
*备注: Manuscript submitted to Nature Communications
Abstract:Subseasonal-to-seasonal (S2S) precipitation forecasting has substantial financial and societal impact, yet remains challenging because of weak predictive signals, high associated uncertainty, and the computational cost of operational systems, which constrains simulation fidelity. We introduce SimCast-S2S, a generative latent-diffusion framework for probabilistic S2S precipitation forecasting that addresses three major bottlenecks in data-driven prediction. First, because S2S prediction requires uncertainty quantification rather than only deterministic point forecasts, SimCast-S2S is the first data-driven system that uses a diffusion-based generative pipeline for S2S prediction, enabling effective sampling from the underlying conditional distribution. Second, since generating large probabilistic ensembles is computationally costly in physical space, SimCast-S2S instead operates in a compact latent space learned by variational autoencoders, enabling efficient large-ensemble generation. Third, diffusion models typically require large training datasets; SimCast-S2S overcomes this via transfer learning with low-rank adaptation (LoRA), pretraining on large ensembles of climate simulations before fine-tuning on limited reanalysis data. On reanalysis data, SimCast-S2S outperforms deep learning baselines, including convolutional neural networks and U-Net architectures. Notably, despite using only a subset of atmospheric input variables and no post-processing, bias correction, or calibration, SimCast-S2S remains competitive with, and in many cases outperforms, state-of-the-art operational systems such as the ECMWF-S2S baseline. These results indicate that latent generative modeling combined with simulation-to-reanalysis transfer learning offers an efficient and scalable path toward data-driven probabilistic S2S precipitation forecasting.
[LG-42] GRAS: Guided Reduced-Variance Proposals and Adaptive Selection for Training-Free Reward Alignment in Discrete Diffusion
链接: https://arxiv.org/abs/2608.26585
作者: Kwanyoung Kim
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE); Quantitative Methods (q-bio.QM); Machine Learning (stat.ML)
*备注:
Abstract:Discrete diffusion models have become a strong, widely adopted class of generators for sequence data, and steering them toward a downstream reward at inference time, without any retraining, is increasingly important. Such training-free steering is done by gradient guidance, by search, or by combining the two. We study the combined regime and identify two weaknesses in how it is usually run: the guided proposal estimates its gradient from a single noisy sample, and the search then resamples particles at a fixed temperature that ignores how rewards spread across each denoising step. We address both with a small set of changes that add no denoiser cost. For the proposal, we lower the estimator variance with a Rao-Blackwellized reveal for differentiable rewards and a leave-one-out baseline for non-differentiable ones; for the search, we standardize the per-step values into a group-relative advantage and prove it collapses to a single active ingredient, an adaptive resampling temperature. We call the resulting method Guided Reduced-variance proposals and Adaptive Selection (GRAS). GRAS is simple yet effective: across regulatory DNA and protein design it attains the best training-free reward, outperforming prior training-free methods and matching or surpassing a reward-fine-tuned model, and it remains effective even for non-differentiable rewards.
[LG-43] Activation Outliers Matter: Robust Recovery for Quantized Multimodal LLM s
链接: https://arxiv.org/abs/2608.26581
作者: Tanzila Rahman,Mehran Taghian Jazi,Yunke Peng,Zhuang Ma,Anandharaju Durai Raju,Yao Wang,Xing Huang,Hei Yi Mak,Shadan Golestan,Hoang Le,Yonghan Dong,Wei Guo,Yaoyuan Wang
类目: Machine Learning (cs.LG)
*备注: 14 Pages, 5 figures, 5 tables
Abstract:Low-bit quantization offers a promising avenue for reducing the computational and memory demands of Multimodal Large Language Models (MLLMs). Recent hardware support for low-precision formats, ranging from MXFP8 to ultra-low-bit formats such as MXFP4 and HiF4, has accelerated research into efficient MLLM training and deployment. In this work, we present a systematic study of these quantization schemes in representative MLLMs that span both video generation and reasoning tasks. Our analysis shows that MXFP8 achieves near-lossless performance, whereas aggressive 4-bit quantization leads to significant degradation. Through extensive ablations, we identify activation quantization as the primary source of this performance loss, contributing substantially more than weight quantization. Motivated by this observation, we propose Residual Fallback Quantization (RFQ), a lightweight activation reconstruction framework that supplements the primary ulta-low-bit activation representation with an auxiliary quantized residual pathway. By explicitly modeling and compensating for quantization errors, RFQ improves activation fidelity while preserving the efficiency advantages of ultra-low-bit computation. RFQ requires no architectural modifications and incurs negligible computational overhead. Extensive experiments on Wan2.2 and Qwen3-VL demonstrate that RFQ consistently recovers a substantial portion of the performance lost under the quantization of MXFP4 and HiF4, significantly narrowing the gap to BF16 baselines across both generation and 4 reasoning benchmarks. Our findings establish activation quantization as the dominant bottleneck in ultra-low-bit MLLMs and highlight residual-based activation reconstruction as an effective and practical strategy for robust 4-bit deployment.
[LG-44] Arrive and Survive: Scaling Safe Goal-Conditioned Policy Learning from One-Bit Failure Signals
链接: https://arxiv.org/abs/2608.26571
作者: Guopeng Li,Yiyang Duan,Yiru Jiao,Chengcheng Xu
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注: 21 pages, 14 figures, 5 tables, Code: this https URL
Abstract:Contrastive reinforcement learning (CRL) scales effectively in goal-conditioned tasks by casting policy learning into a self-supervised contrastive objective. However, in a failure-terminated Markov decision process, established CRL considers pre-failure future goals only when constructing positive samples, without accounting for the probability mass removed by failure termination. Our theoretical analysis shows that this omission induces a systematic overestimation bias in goal-reaching values. Consequently, near-failure trajectories provide disproportionately strong supervision of success despite retaining little future occupancy. Unsafe actions can thereby be reinforced through catastrophic failure bootstrapping, leading to failed policy learning and unsustainable goal-reaching behaviours. To address this problem, we introduce two minimal yet strong corrections: mass-weighted InfoNCE corrects the overweighting of short surviving futures in critic learning, and a log-survival-mass score restores the missing survival mass in policy optimization. The resulting method, Safe Contrastive Reinforcement Learning (Safe-CRL), requires only the one-bit signal provided by failure termination to scale safe goal-conditioned policy learning. Across twelve failure-prone robot navigation and locomotion tasks, Safe-CRL consistently improves survival and substantially outperforms the Scaling-CRL baseline in goal-reaching performance. Additionally, deep Safe-CRL policies exhibit complex failure-avoidance behaviours. This study completes the CRL theory under failure termination and provides a scalable safe RL framework. The code is available via this https URL.
[LG-45] Hadamard Flattening and Gaussian Pooling Sketch for Least Squares with Coordinate-wise Guarantee
链接: https://arxiv.org/abs/2608.26552
作者: Zhao Song,Lichen Zhang
类目: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Abstract shortened to meet arxiv requirement
Abstract:Randomized sketch-and-solve algorithms accelerate overconstrained \ell_2 regression by replacing the input with a smaller problem. Standard subspace embeddings guarantee that the cost of the regression is nearly preserved, but coordinate-wise accuracy of the solution is more delicate: we want the solution vector itself to be close to the optimal solution in \ell_\infty norm. In particular, we want to find a vector x’\in \mathbbR^d such that |x’-x^*|\infty\leq \frac\epsilon\sqrt d\cdot |Ax^\star-b|2\cdot |A^\dagger|\rm op . Price, Song and Woodruff initiated the study of this problem and showed that the subsampled randomized Hadamard transform (SRHT) with O(\epsilon^-2 d^1+\Theta(\sqrt\log\log n/\log d)) rows achieves this guarantee. A subsequent work of Song, Ye, Yin and Zhang claimed to improve the row count to O(\epsilon^-2d\log^3 n) . Unfortunately, their proof relies on an independence assumption that does not hold in general, and we exhibit an explicit instance on which it fails. To achieve a truly nearly-linear-in- d row count, we introduce a new fast, dense randomized transform, which combines a randomized Hadamard flattening, a random permutation, and balanced, disjoint Gaussian pooling. Conditioned on the Hadamard-and-permutation stage, the sketched problem becomes an exact Gaussian regression in which the noise is independent of the entire sketched design; this conditional independence is exactly what the earlier argument was missing. Our sketch yields the \ell\infty guarantee with m=O(\epsilon^-2d\log d) rows, uses one Hadamard pass with a padded internal dimension N=\widetildeO(n+\epsilon^-2d^3) , and is efficient to apply: the sketched pair (SA, Sb) can be computed in O(Nd\log N)=\widetildeO(nd+\epsilon^-2d^4) time. Comments: Abstract shortened to meet arxiv requirement Subjects: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Machine Learning (stat.ML) Cite as: arXiv:2608.26552 [cs.DS] (or arXiv:2608.26552v1 [cs.DS] for this version) https://doi.org/10.48550/arXiv.2608.26552 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-46] Chart2SVG: Editable SVG Generation from Raster Chart Images
链接: https://arxiv.org/abs/2608.26544
作者: Jinning Cui,Lu Chen,Haoyan Shi,Yue He,Chenglong Wang,Mengyu Zhou,Weidong Huang,Yunhai Wang
类目: Machine Learning (cs.LG)
*备注:
Abstract:We present Chart2SVG, a multimodal large language model that converts static raster charts into structurally organized, semantically enriched SVGs that support programmatic editing. By incorporating chart-specific semantic tokens into a vision-language model, Chart2SVG captures both geometric primitives and their functional roles. To support robust structural recovery, we introduce Beagle+, a dataset of 33K canonicalized and structurally distilled chart samples. Our approach combines specialized training objectives with a rendering-aware post-training phase, producing SVGs that are both visually accurate and structurally consistent. To facilitate higher-level manipulations, we construct a Chart Structure Graph (CSG) that exposes visual dependencies, enabling tasks such as interactive exploration, chart repurposing, and layout reuse. Experiments show that Chart2SVG substantially outperforms baselines in reconstruction fidelity and downstream editing utility, advancing the development of intelligent and interactive visualization tools.
[LG-47] Predicting Quantifiability from Primary Screens to Prioritize Dose-Response Profiling
链接: https://arxiv.org/abs/2608.26538
作者: Sean Lim
类目: Machine Learning (cs.LG); Biomolecules (q-bio.BM)
*备注:
Abstract:High-throughput drug screening relies on low-cost primary assays to prioritize compounds for more expensive dose-response profiling, where potency is ultimately quantified. Current screening strategies largely focus on identifying compounds that will confirm biological activity on follow-up, implicitly assuming that confirmed activity will also yield a usable potency estimate. However, confirmed biological activity in screening does not necessarily translate into a quantifiable potency, because active compounds can still fail to produce a reportable dose-response estimate. We therefore present a framework for modeling quantifiability, whether follow-up testing will yield a usable potency estimate, as a distinct triage objective from biological activity. Quantifiability was strongly predictable from the preceding low-cost screen, with most predictive information arising from the observed screening features rather than molecular structure. Response-based predictors remained robust on previously unseen chemical scaffolds and generalized across held-out assay-mechanism families, while the probability of successful quantification varied strongly with response amplitude and assay context. These findings establish experimental measurability, distinct from biological activity, as a predictable property of screening outcomes and show that quantifiability-aware triage can improve the allocation of costly dose-response profiling capacity.
[LG-48] High Probability Derivative Bounds for Random tanh Neural Networks on a Hypercube
链接: https://arxiv.org/abs/2608.26526
作者: Josef Dick,Michael Feischl,Fabian Zehetgruber
类目: Machine Learning (cs.LG); Numerical Analysis (math.NA)
*备注: 31 pages, 0 figures
Abstract:We establish high-probability bounds for mixed input derivatives of wide random neural networks whose activation derivatives satisfy a factorial growth bound. Our main result specializes these estimates to \tanh networks with Xavier initialization. A direct deterministic analysis based on Euclidean operator norms of the weight matrices yields derivative bounds that generally grow exponentially with the depth. We show that this growth can be substantially improved for sufficiently wide Gaussian networks by isolating the term that is linear in the highest-order derivative and controlling the corresponding tangent directions by measurable finite nets. For scalar-output \tanh networks with Gaussian weights and Xavier initialization, we prove that there exist constants C,C_0,C_10 such that, whenever the common hidden width satisfies n \geq C\left(L^3n_0^2(1+\log n_0)+L^2\left(1+\log(L/\eta)\right)\right) , then, with probability at least 1-\eta , the estimate \left|D^u\mathcalR_\Phi^(L)(x)\right| \leq C_0 |u|! (C_1L)^|u|-1\prod_j\in u\beta_j(\eta,n_0) holds simultaneously for every non-empty u\subseteq[n_0] and every x\in[0,1]^n_0 . Thus, the first-order derivative bound is independent of the depth, while a square-free mixed derivative of order |u| grows at most polynomially as L^|u|-1 , apart from the coordinate factors. As consequences, we obtain high-probability bounds for the Euclidean Lipschitz constant and for weighted Sobolev norms of the network realization. The latter connect the derivative estimates to quasi-Monte Carlo integration and indicate how such regularity can enter the analysis of QMC-based training. Comments: 31 pages, 0 figures Subjects: Machine Learning (cs.LG); Numerical Analysis (math.NA) MSC classes: 65D30 Cite as: arXiv:2608.26526 [cs.LG] (or arXiv:2608.26526v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.26526 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-49] Algorithmic Principles For Multiclass Learning Are Hard To Come By: Limits of Regularization and Proper Learning
链接: https://arxiv.org/abs/2608.26516
作者: Julian Asilis,Shaddin Dughmi,Vatsal Sharan,Alec Sun,Shang-Hua Teng,Chang Wang
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 45 pages
Abstract:Two of the most fundamental questions in statistical learning theory are the following: which prediction problems are learnable, and how should they be learned? For the former, elegant answers often take the form of combinatorial dimensions. The latter question, however, has proved considerably more elusive: all known general-purpose multiclass learners rely on intricate orientations of exponentially large one-inclusion structures, and familiar algorithmic principles such as proper learning and regularization remain poorly understood. Motivated by prior work, we ask whether learning reduces to proper learning—possibly over a larger hypothesis class—and whether proper or improper multiclass learning can ultimately be captured by suitable regularizers. Our primary results answer both questions negatively, resolving three open problems from prior work. First, we exhibit a learnable multiclass problem that cannot be embedded in any properly learnable class, meaning learning cannot be reduced to proper learning by enlarging the hypothesis class. Second, we demonstrate that proper learning can require training error and characterize this phenomenon precisely: every properly learnable class admits a proper learner making o(m) errors on samples of size m , but every prescribed sublinear scale a_m=o(m) is necessary for some properly learnable problem. Third, regularization is not a general learner: we exhibit a properly learnable class that cannot be learned by any Structural Risk Minimization (SRM) learner, and a learnable class that cannot be learned by any local regularizer. We complement these impossibility results with a positive theory that gives two sufficient conditions for SRM learnability and characterizes SRM representability through integrability of revealed preferences. Comments: 45 pages Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML) Cite as: arXiv:2608.26516 [cs.LG] (or arXiv:2608.26516v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.26516 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-50] Sharp Minimax Regret for Infinite-Memory Logistic Prediction
链接: https://arxiv.org/abs/2608.26515
作者: Vaneet Aggarwal
类目: Information Theory (cs.IT); Machine Learning (cs.LG)
*备注:
Abstract:We study online prediction for a specific finite-alphabet, exogenously driven source with infinite input memory. Independent Rademacher inputs (U_t) are observed sequentially, and the next binary mark has logit \sum_j=1^t\theta_jU_t+1-j , where \abs\theta_j\leq r_j and \sum_jr_j\leq B . Regret is expected cumulative excess log loss. Lag j can affect prediction by scale r_j and enters only n_T,j=T-j+1 prediction rounds, leading to the lag-resolved spectrum \Gamma_T®=\sum_j=1^T\log!\left(1+n_T,jr_j^2\right) . For every summable envelope, a localized Bayesian mixture proves \cR_T®\leq C\Gamma_T® . For exponential and polynomial envelopes, under the stated finite-sample dimension condition, a Toeplitz-design converse proves \cR_T®\geq c\Gamma_T® , with constants allowed to depend on the fixed decay parameters and the logit bound. Thus \Gamma_T® is the minimax cumulative-regret scale for this source class in these canonical regimes, giving \Theta(\alpha^-1\log^2T) for r_j=Ae^-\alpha j and \Theta(T^1/(2s)) for r_j=Aj^-s , s1 . The converse is specific to the exogenous lagged model and is not a profile-only theorem for arbitrary stationary infinite-memory sources. Retaining only the most recent h inputs costs order \sum_jhn_T,j\theta_j^2 , yet the same worst-case truncation profile can correspond to polynomially different regret. A scaled online Newton predictor attains the spectrum upper bound.
[LG-51] A Unified Framework for Fair and Personalized Decentralized Learning under Communication Constraints
链接: https://arxiv.org/abs/2608.26493
作者: Krishnendu S. Tharakan,Carlo Fischione
类目: Machine Learning (cs.LG)
*备注:
Abstract:Decentralized learning systems aim to collaboratively train models across multiple clients without relying on a central coordinator. While decentralization improves scalability, privacy, and robustness, it also exacerbates three fundamental challenges: statistical heterogeneity across clients, fairness in client-level performance, and stringent communication constraints. This raises a natural question: \emphhow fair can decentralized learning be under limited communication? We address this question by presenting a unified framework for decentralized learning under communication constraints, bringing together graph-based personalization, agnostic fairness, and compressed event-triggered communication. Specifically, we propose a new algorithm DMFL-SQ, a decentralized multi-task learning algorithm that couples personalized model training over a communication graph with an agnostic mixture fairness objective, while reducing communication through sparsification, quantization, and event-triggered synchronization. We establish convergence guarantees for general non-convex objectives and show that DMFL-SQ achieves an \mathcalO(T^-1/2) rate in expected squared Moreau-envelope stationarity despite sparse, quantized, and event-triggered communication. We further derive PAC-Bayes generalization guarantees for the fairness-aware mixture objective. Experiments on CIFAR-10 and the real heterogeneous MUSMET EEG dataset demonstrate that DMFL-SQ substantially reduces communication while maintaining predictive performance and improving fairness across clients. Together, our theoretical and empirical results show that personalization, fairness, and communication efficiency can be jointly achieved in decentralized learning while preserving the dominant convergence rate.
[LG-52] Bayesian methods and Markov chain Monte Carlo algorithms for curve reconstruction and point cloud data analysis
链接: https://arxiv.org/abs/2608.26490
作者: Asir Intesar Tushar,Ioannis Sgouralis
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 32 pages and 12 figures
Abstract:Point-cloud data routinely captured by modern imaging and sensor technologies provide detailed geometric descriptions of objects and environments, but their analysis is hindered by large data volumes, localization noise, and missing information. In addition, existing point-cloud reconstruction pipelines typically return a single best-fit structure without uncertainty quantification. We introduce a fully Bayesian framework for representing point-cloud data and reconstructing closed curves, in which observed points are modeled as noisy perturbations of latent locations constrained to lie on the underlying curve that is regularized by a non-parametric prior. Posterior inference in our framework is carried out using a series of Markov chain Monte Carlo samplers tailored to point-cloud characteristics. Numerical experiments, including synthetic examples and real-world LiDAR datasets, show accurate reconstructions and quantified uncertainty over the recovered curves.
[LG-53] Shared Actors Need Not Share Critics: Effects of Value Mismatch in Parallel Reinforcement Learning
链接: https://arxiv.org/abs/2608.26481
作者: Zhenya Liu,Yang Meng,Zhuokai Zhao,Xuefeng Liu,Yuxin Chen
类目: Machine Learning (cs.LG)
*备注:
Abstract:When a single policy is trained in parallel across multiple environments of the same task, such as procedurally generated levels, randomized dynamics, or curricula, implementations commonly use one critic across all sampled environments. Yet different environments can assign different expected returns to the same input visible to the critic. A critic without environment information must then reconcile distinct value targets, systematically shifting the sampled advantages within individual environments. Using illustrative bandit models with multiple environments and a common optimal arm, we characterize how this value mismatch redistributes sampled policy updates, reinforcing unhelpful actions while attenuating or even reversing useful ones. The oracle processes using no baseline, the shared value, or the value specific to the sampled environment have the same mean logit update at a fixed policy and converge to the same optimal policy, yet their realized learning paths can differ sharply. The analysis motivates a minimal intervention: give only a logged environment index to the critic so that it can separate the value targets. Controlled CartPole and MuJoCo experiments expose the predicted shifted values, advantages, and performance gaps. In the more complex BipedalWalker and Procgen settings, the same intervention yields more stable learning and higher returns. Across all 16 Procgen games, the multihead conditional critic improves aggregate normalized return on 600 unseen levels per game by 40.8% . In conclusion, the theory identifies value mismatch as a direct mechanism through which critic sharing can degrade stochastic learning dynamics, not captured by scalar estimator variance alone, and the experiments show that conditioning on an index is broadly effective in parallel reinforcement learning.
[LG-54] Active Curriculum Refinement for Reinforcement Learning
链接: https://arxiv.org/abs/2608.26469
作者: Zhenya Liu,Yuxin Chen
类目: Machine Learning (cs.LG)
*备注:
Abstract:In many reinforcement learning (RL) domains, environments are connected by prerequisite relations, such as difficulty-increasing edits or parameter increments, which induce a directed acyclic curriculum graph (DAG). Although this structure is often exploited only implicitly, explicitly modeling it can improve training. We introduce PATH, a curriculum-learning framework that performs active learning over the curriculum graph. PATH first expands coverage by sampling diverse curriculum paths and then reallocates training toward regions that remain unmastered. Experiments across diverse environments show that PATH explicitly leverages the graph structure to achieve strong robustness and generalization.
[LG-55] Distributed Training using an Intelligent Network
链接: https://arxiv.org/abs/2608.26453
作者: Nihar Shah,Ben Blier
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC); Networking and Internet Architecture (cs.NI)
*备注:
Abstract:Distributed training across a wide area network (WAN) is challenging, as continuous parameter exchange by islands of compute is constrained by limited bandwidth, high latency, and uneven topology. We propose making the network an active participant in training. On the systems side, such networks should leverage (i) multicast technology to replicate outbound traffic and (ii) in-line FPGAs to aggregate inbound traffic, to ease egress and ingress bottlenecks. These technologies are used for training across workers within a data center, but this paper extends them to the WAN. On the algorithms side, we develop an optimization framework that produces rich synchronization schedules (namely, rotating cliques of islands) around the underlying network topology and these technologies, to maximize information exchange. Finally, we illustrate this on a nine-city topology modeled on the DoubleZero network, a live programmable WAN equipped with both technologies, and show how the optimal schedules shift with the network’s capabilities. Together, these can narrow the gap to the gold standard of colocated training.
[LG-56] oward Equitable Low-Carbon Mobility: Fairness-Aware Demand Prediction for Expanding Bike-Sharing Systems
链接: https://arxiv.org/abs/2608.26451
作者: Man Luo,Yixuan Zhao
类目: Machine Learning (cs.LG)
*备注:
Abstract:Bike-sharing systems are an important component of low-carbon urban mobility, but continued expansion creates challenges in both cold-start prediction and equitable resource allocation. Newly deployed stations lack historical ridership records, causing a mismatch between training and inference for graph-based models on evolving networks. Historical demand may also encode structural inequalities, as lower ridership in low-income neighborhoods can reflect limited infrastructure access rather than weak latent demand. Models trained directly on such data may therefore reinforce existing mobility disparities. We propose FairGIN, a fairness-aware graph neural network for demand prediction in expanding bike-sharing systems. FairGIN integrates three components. Expansion-Simulated Increment Training stochastically simulates network expansion during training to reduce the cold-start distribution gap. Attention-Based Knowledge Transfer combines station-adaptive temperature scaling with orthogonal embedding alignment to transfer representations from data-rich existing stations to data-sparse new stations. Fairness-Aware Optimization introduces income-stratified regularization and an equity-calibrated deployment score to support more inclusive station placement. Experiments on NYC and Seattle demonstrate that FairGIN achieves state-of-the-art predictive accuracy across diverse expansion scenarios while substantially reducing income-based disparities without compromising overall system efficiency.
[LG-57] Subgraph Filtering for Fair Graph Neural Networks ICONIP2026
链接: https://arxiv.org/abs/2608.26437
作者: Haohui Lu,jiyuan Tian,Fangyu Zhou,Shahadat Uddin
类目: Machine Learning (cs.LG)
*备注: ICONIP 2026
Abstract:Graph neural networks (GNNs) can exhibit unfair behavior even when sensitive attributes are excluded from node features, because graph topology and message passing propagate group-correlated signals under sensitive homophily. Existing fairness-aware GNN methods mainly constrain representations or prediction distributions at a global level, without explicitly controlling the local structural pathways through which biased information propagates during aggregation. We propose Subgraph Filtering for Fair Graph Neural Networks (SF-GNN), a lightweight and architecture-agnostic framework that mitigates structural bias at its source. SF-GNN identifies bias-prone edges by combining sensitive homophily with structural propagation amplifiers, including hub participation and triadic closure. It then incorporates stochastic edge filtering into each message-passing step to selectively downweight or remove these edges while preserving the remaining graph structure. Training further incorporates a statistical-parity regularizer with a warm-up schedule to stabilize optimization. Experiments on five benchmark datasets show that SF-GNN achieves consistent fairness improvements while maintaining competitive predictive performance, leading to a better fairness–accuracy trade-off than recent fairness-aware GNN baselines.
[LG-58] NeoTriFuse: Reliability-Aware Multimodal Fusion under Missingness Heterogeneity for Neonatal Mortality Risk Prediction ICONIP2026
链接: https://arxiv.org/abs/2608.26436
作者: Jiyuan Tian,Qincheng Shen,Ye Lin,Yu Gao,Haohui Lu
类目: Machine Learning (cs.LG)
*备注: ICONIP 2026
Abstract:Neonatal mortality risk prediction from bedside monitoring data remains challenging due to extreme class imbalance, heterogeneous clinical risk factors, multi-scale temporal dynamics, and substantial missingness. We propose NeoTriFuse, a reliability-aware multimodal fusion framework for missingness-heterogeneous neonatal monitoring data. Unlike conventional multimodal approaches that treat missingness primarily as a preprocessing issue, NeoTriFuse models missingness as an explicit reliability signal that dynamically modulates modality contributions during fusion. The framework integrates static perinatal variables, local-global temporal encoders, and patient-level statistical summaries through reliability-guided gating mechanisms, while jointly optimizing mortality prediction and an auxiliary length-of-stay objective. NeoTriFuse achieves competitive performance, with an F1 score of 0.6736 +/- 0.0216 and an AUROC of 0.9454 +/- 0.0056. Ablation studies indicate that the local-global temporal architecture and patient-level summary branch contribute most substantially to predictive performance, while reliability-aware gating provides additional improvements on threshold-dependent metrics under heterogeneous observation completeness. Sensitivity analyses further suggest stable performance across nearby hyperparameter settings. Overall, the findings support reliability-aware multimodal fusion as a practical approach for neonatal mortality prediction under realistic clinical missingness conditions.
[LG-59] FedCMAPSS: A Benchmark for Federated Learning in Remaining Useful Life Estimation
链接: https://arxiv.org/abs/2608.26433
作者: Amelia Sorrenti,Matteo Pennisi,Concetto Spampinato,Simone Palazzo
类目: Machine Learning (cs.LG)
*备注: Accepted at the 21st IEEE Conference on Industrial Electronics and Applications (ICIEA 2026)
Abstract:Data-driven prognostics and health management has emerged as a key enabler for Industry 4.0, yet the development of robust remaining useful life (RUL) estimation models is often limited by the scarcity of run-to-failure data. While federated learning offers a promising paradigm to collaboratively train predictive models without sharing sensor data, research efforts have operated so far in the absence of a common evaluation framework. To address this gap, this paper introduces FedCMAPSS, a benchmark for federated RUL estimation based on the commonly-used NASA C-MAPSS dataset. We define a set of five standardized tasks designed to simulate real-world industrial challenges, ranging from ideal IID settings to extreme statistical heterogeneity, and conduct a systematic evaluation of state-of-the-art federated optimization algorithms across multiple neural architectures. By establishing reproducible baselines and making the source code and data splits publicly available, this work aims to provide a standard foundation for developing and comparing federated predictive maintenance solutions.
[LG-60] Beyond Capability Benchmarks: Learning Operational Fingerprints of LLM Cloud Services from Production Incident Metadata
链接: https://arxiv.org/abs/2608.26332
作者: Meiwei Zhang,Eduardo Miranda,Bruce Baynes,Suvigya Jain,Wanlong Chen,Tao He,Sergey Borodavkin
类目: Machine Learning (cs.LG)
*备注:
Abstract:Managed LLM services are now part of real production systems, but model selection and service planning still rely heavily on capability benchmarks that reveal little about operational behavior after deployment. We present Operational Embedding (OpEmbed), a framework for learning compact operational fingerprints of LLM cloud services from structured, privacy-preserving support-case metadata, without using case text. OpEmbed aggregates model–time windows into an eight-channel operational signature and learns a low-dimensional representation via temporal contrastive learning, cross-view reconstruction, and generational-ordinality regularization. Evaluated on more than 33,000 production support cases spanning seven LLM families over 26 months at Google Cloud, OpEmbed recovers interpretable family- and version-level structure, improves leave-one-model-out operational forecasting over non-learned baselines, remains useful under limited early-window data, and supports cross-model fault-type transfer. We report the practical lessons learned from building and evaluating this tool for model onboarding, support readiness assessment, and operational monitoring.
[LG-61] Privacy Without Regret: Differentially Private Inference-Time Alignment
链接: https://arxiv.org/abs/2608.26324
作者: Ishi Jain,Nandini Bhattad,Sayak Ray Chowdhury
类目: Machine Learning (cs.LG)
*备注:
Abstract:Best-of-N (BoN) sampling is the simplest and most widely deployed inference-time alignment strategy, but it suffers from two distinct problems: reward hacking, in which the selected response exploits errors in the proxy reward model, and the absence of any privacy protection for the sensitive human preference data used to train that reward model. We show that a single intervention-adding calibrated noise to reward scores before selection-resolves both. Our first result, Private Best-of-N (PrivBoN), establishes that Gumbel noise at an appropriate scale simultaneously provides \epsilon -differential privacy and implements KL-regularized alignment. Whenever the privacy budget exceeds a critical threshold \epsilon^* , the privacy-mandated noise is the regret-optimal regularization, and privacy imposes zero additional alignment cost-matching the information-theoretic skyline of Huang et al. (2025). Because \epsilon^* depends on an unknown coverage coefficient, we introduce Private Inference-Time Pessimism (PrivITP), which combines \chi^2 -regularized rejection sampling with a two-phase Gaussian mechanism. PrivITP achieves ex-post (\epsilon,\delta) -DP with a privacy cost independent of the number of responses n , cleanly decouples the regularization parameter from the privacy parameter, and attains the skyline up to a noise-inflation term. Experiments across several language models, datasets, and reward models confirm our results: PrivBoN and PrivITP are scaling-monotonic (unlike BoN, which degrades past a critical n ), and PrivITP matches or outperforms PrivBoN at equivalent privacy levels, with the largest gains in the strong-privacy regime.
[LG-62] Algebraic Multigrid Acceleration for Efficient Label Spreading
链接: https://arxiv.org/abs/2608.26309
作者: Antonia van Betteray,Jonathan Klees,Miriam Schäfers,Matthias Rottmann
类目: Machine Learning (cs.LG)
*备注:
Abstract:Modern machine learning models rely on large amounts of labeled data. However, manual annotation of large-scale datasets is expensive and time-consuming. Label spreading is a semi-supervised learning technique that addresses this challenge by propagating information from a few labeled examples to a larger pool of unlabeled data. Despite its effectiveness, its application to large-scale, high-dimensional datasets is limited by computational costs and memory constraints. To address these limitations, we propose Algebraic Multigrid Acceleration for Efficient Label Spreading (AMELS), an efficient label spreading framework that improves scalability by fast construction of neighborhood graphs and the incorporation of algebraic multigrid solvers. The latter is an iterative solver that replaces the ordinary random walk iteration typically performed in label spreading. Due to the multilevel nature of algebraic multigrid solvers, AMELS spreads given label information across a graph of any size in a single multigrid cycle. We demonstrate that AMELS achieves significant runtime reductions compared to existing implementations while also being more robust to hyperparameter choices in terms of both runtime and classification accuracy. Our framework therefore enables efficient label spreading on large-scale image datasets and produces accurate labels even when only a few labeled samples are available.
[LG-63] Muon with Finite Newton-Schulz: The Smoothing Benefit in Nonsmooth Nonconvex Optimization
链接: https://arxiv.org/abs/2608.26288
作者: Mingyi Li,Taira Tsuchiya
类目: Machine Learning (cs.LG); Optimization and Control (math.OC); Machine Learning (stat.ML)
*备注: 37 pages, 3 figures
Abstract:Muon has emerged as a strong optimizer for the matrix-valued parameters in large language model pretraining, approximately orthogonalizing its momentum with a few Newton-Schulz iterations. Existing theory either replaces this iteration with the exact polar factor it approximates, or treats its finite depth as an approximation error, and thus the iteration Muon actually runs can only hurt the guarantees. We show that finite Newton-Schulz can instead be beneficial for nonsmooth nonconvex optimization. To this end, we analyze Muon through the online-to-nonconvex conversion, which views the update rule as an online learner and converts its regret bound into a stationarity guarantee. The finite Newton-Schulz iteration smooths the discontinuous polar map into a Lipschitz map of the singular values, and Muon with finite Newton-Schulz can be regarded as an online learner with a smoothed spectral potential. This smoothing is exactly what the conversion needs: we prove that a Newton-Schulz depth growing only logarithmically in the target accuracy suffices for convergence to stationary points in nonsmooth nonconvex optimization, whereas Muon with the exact-polar update may fail to converge. The resulting sample complexity bounds match the best-known guarantees for nonsmooth nonconvex optimization and are optimal for smooth nonconvex optimization up to problem-dependent factors. The argument extends beyond Newton-Schulz to general spectral maps with the same smoothing property.
[LG-64] Multi-Dataset Inverse Problem Solving with Distributed Generative AI
链接: https://arxiv.org/abs/2608.26283
作者: Daniel Lersch,Steven Goldenberg,Johann Rudi,Markus Diefenthaler,Kevin Brager,Xingfu Wu,Yaohang Li,Nobuo Sato
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注: 23 pages, 14 figures
Abstract:Extracting a shared set of unknown, not directly measurable quantities from multiple, heterogeneous datasets is a common challenge across scientific domains. A prominent example is the combination of datasets obtained from different measurements with different settings (e.g. varying detector resolutions). Analyzing such datasets jointly, rather than independently or after naive merging, is essential for obtaining precise and unbiased estimates of the unknowns, but requires careful treatment of dataset heterogeneity and is computationally demanding. We present a generalized framework for simultaneously analyzing multiple heterogeneous datasets in the context of generative AI-based inverse problem solvers. Building on our recent Scalable Asynchronous Generative Inverse Problem Solver (SAGIPS) framework, we extend the well-established distributed data-parallel training paradigm to non-identically distributed datasets, where each dataset is controlled by the same set of unknown inference parameters but covers a different region of the available feature space. Each dataset is processed through its own forward operator and discriminator, providing complementary constraints that collectively guide a shared generator toward global parameter consistency. We validate the approach using a controlled setup inspired by a multi-detector scattering experiment. We provide numerical evidence that our framework is robust to different data fidelities, which arise from unknown detector systematics in the Rutherford experiment, and we show the scaling behavior on multi-GPU leadership computing systems. The results show that our approach is well suited for real-world multi-dataset analyses in which experimental conditions vary across measurements.
[LG-65] Constraint-Aware Physics-Informed Neural Networks for Static Shape Estimation of Co-Manipulative Continuum Robots
链接: https://arxiv.org/abs/2608.26273
作者: Rana Danesh,Pari Qarehdaghi,Farrokh Janabi-Sharifi
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Static shape estimation of co-manipulative continuum robots (CCRs) is challenging because the continuum arms and manipulated flexible object form a closed chain that must satisfy both static equilibrium and geometric loop-closure constraints. This paper presents a constraint-aware physics-informed neural network (PINN) for static shape estimation of a tendon-driven CCR modeled using the geometric variable strain formulation. The proposed method incorporates a projected static equilibrium residual and a configuration-level geometric residual to enforce the governing mechanics and closed-chain geometry. In simulation, the PINN is compared with a purely data-driven artificial neural network (ANN) under limited and noisy training data. With 140 samples and 50% label noise, the PINN reduces the relative configuration error, equilibrium residual, and closed-chain residual by 67.88%, 67.35%, and 88.06%, respectively. Using the full dataset, the PINN achieves 0.1597% relative configuration error with an inference time of 0.1773 ms, compared with 17.97 s for an iterative nonlinear solver. Experimental fine-tuning reduces the marker RMSE from 2.657 mm to 0.497 mm and increases R2 from -0.788 to 0.937. These results demonstrate accurate, physically consistent, and computationally efficient static shape estimation of closed-chain CCRs.
[LG-66] Pruning Binarized Neural Networks: A Dedicated Framework and Globally Weighted Algorithms
链接: https://arxiv.org/abs/2608.26233
作者: Roan Rubiales,Jean Pierre David
类目: Machine Learning (cs.LG)
*备注: 9 pages, 3 figures, 5 tables, 3 algorithms
Abstract:Extreme compression of deep neural networks, up to full binarization, dramatically reduces memory footprint and arithmetic complexity, facilitating deployment on constrained edge hardware with field-programmable gate arrays (FPGAs) and microcontrollers. Although combining binarization with pruning promises additional efficiency gains, existing pruning strategies are ill-suited to binarized representations and rarely translate into meaningful hardware savings. We introduce a PyTorch-based, research-oriented framework that incorporates freezing and pruning mechanisms for designing and optimizing binarized neural networks. The framework enables rapid and reproducible evaluation of state-of-the-art approaches and the fast prototyping of new ones. Leveraging this framework, we propose a novel pruning method that accounts for the relative importance of learned parameters across abstraction levels. Such a global weighting mechanism consistently achieves a superior trade-off between model accuracy and pruning rate, achieving a 70% pruning rate on VGG11 with constant accuracy, while state-of-the-art results reach only 41% in the binarized setting.
[LG-67] Universality and sharp thresholds for ellipsoid fitting
链接: https://arxiv.org/abs/2608.27372
作者: Frederic Koehler,Youngtak Sohn
类目: Probability (math.PR); Disordered Systems and Neural Networks (cond-mat.dis-nn); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注: 79 pages, 4 figures
Abstract:We establish a sharp phase transition for fitting random vectors by an ellipsoid. The random vectors have independent subgaussian coordinates with mean zero, variance one, and a common fourth moment, and the number of vectors is proportional to the square of the dimension. We identify an explicit satisfiability threshold such that, with high probability, a positive definite ellipsoid passes through every data point below the threshold, whereas no positive semidefinite fit exists above it. We also determine the optimal squared fitting error throughout the unsatisfiable regime. In particular, the threshold depends on the coordinate distributions only through their common fourth moment, revealing a fourth moment universality phenomenon. For standard Gaussian data the threshold is 1/4 , resolving the ellipsoid fitting conjecture.
[LG-68] A Finite Sample Analysis for Quantile Temporal Difference Learning in Distributional Reinforcement Learning
链接: https://arxiv.org/abs/2608.27313
作者: Zijie Cheng,Xiang Li,Yang Peng,Zhihua Zhang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:We establish a global finite-sample guarantee for synchronous quantile temporal-difference learning (QTD) in tabular distributional reinforcement learning. The proof separates two stability mechanisms. A global comparison argument, based on the order monotonicity of reward cumulative distribution functions and the W_\infty contraction of the distributional Bellman operator, brings an arbitrarily initialized iterate into a local neighborhood. Inside that neighborhood, we linearize the QTD mean field. Its Jacobian is a nonsingular M -matrix, and the associated positive semigroup permits a variance-sensitive martingale analysis. For stepsizes \alpha_t=c(t+1)^-a with a\in(1/2,1) , the leading last-iterate fluctuation is of order \widetilde O\bigl(T^-a/2/\sqrt1-\gamma\bigr) and has no polynomial dependence on the number of quantiles. The deterministic transient and the required burn-in can still depend on the smallest Bellman-target density, which is of order m^-1 in the worst case. The result therefore distinguishes sharply between the local stochastic fluctuation and the global sample complexity.
[LG-69] Recovering Expert Critic-Sourced Network Adjacency between Musical Artists from Acoustic Distributions: A Construct-Validity Approach RECSYS
链接: https://arxiv.org/abs/2608.27291
作者: Elena Badillo-Goicoechea,Fengfeng He
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: Accepted workshop paper at USR Workshop, RecSys, 2026, Minneapolis, MN, USA
Abstract:Music recommendation relies primarily on two signals: user-item interactions, which fail in the cold-start regime, and intrinsic musical content, available for any recording. We argue that a third, largely untapped signal is both richer and more principled: critical adjacency, the pairwise relation established when an expert critic explicitly links two artists in long-form prose. It encodes deliberate judgments about which artists belong together. Prior work established its internal validity, showing it recovers coherent, interpretable communities and can match collaborative filtering in user-satisfaction simulations, with no user data. What has been missing is external validation: whether this critic-sourced relation is grounded in the music itself versus sociological context. We test it against acoustic content, reframing the question as one of construct validity. Representing artists as empirical distributions over 80 low-level Essentia acoustic descriptors and modeling pairwise proximity via marginal optimal-transport (Wasserstein) distances, we evaluate how far critical adjacency is sonically recoverable under a cold-start, artist-disjoint split. Our ensemble recovers these edges at out-of-sample AUC of 0.767 (95% CI 0.761-0.775). Recoverability rises monotonically with critical consensus, reaching 0.865 on multi-source attested edges. Stratified evaluations align with sociological models of genre: tightly bounded, scene-based genres show higher recoverability than broad industry umbrella terms. Critical discourse is thus a rich source of information for recommendation, decomposing into a reproducible “sonic core” and a “sociological remainder” driven by narrative positioning, subcultural context, and canonical placement. The work offers both a scalable cold-start discovery mechanism and a sociologically grounded approach to MIR and MRS research. Comments: Accepted workshop paper at USR Workshop, RecSys, 2026, Minneapolis, MN, USA Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG) Cite as: arXiv:2608.27291 [stat.ML] (or arXiv:2608.27291v1 [stat.ML] for this version) https://doi.org/10.48550/arXiv.2608.27291 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-70] Over-The-Air Extreme Learning Machines with Nonlinear Stacked Intelligent Metasurfaces
链接: https://arxiv.org/abs/2608.27137
作者: Kyriakos Stylianopoulos,Mattia Fabiani,Giulia Torcolacci,Davide Dardari,George C. Alexandropoulos
类目: ignal Processing (eess.SP); Machine Learning (cs.LG)
*备注: To be presented at Int. Workshop Signal Process. Adv. Wireless Commun. (SPAWC), Athens, Greece, Sept. 2026
Abstract:The recently envisioned goal-oriented communications paradigm requires machine learning inference to be performed directly on wirelessly transferred data. This paper presents an eXtremely Large (XL) Multiple-Input Multiple-Output (MIMO) system that operates as an Extreme Learning Machine (ELM) to execute Over-The-Air (OTA) binary classification. To reduce hardware complexity, the receiver is equipped with cascaded metasurfaces terminating in a single radio-frequency chain. A front metasurface layer applies a fixed nonlinear response to the incoming signal, acting as the ELM’s activation function. Subsequent tunable linear metasurface layers physically approximate the trained network weights directly in the wave domain. Numerical evaluations across diverse datasets showcase that our XL MIMO architecture achieves classification accuracy comparable to idealized digital models, thereby proving the viability of low-complexity, wave-domain OTA learning.
[LG-71] Linear Independence of Polynomial Compositions and Identifiability of Deep Neural Networks
链接: https://arxiv.org/abs/2608.27113
作者: Kathlén Kohn,Giovanni Luca Marchetti,Alex Massarenti,Massimiliano Mella
类目: Commutative Algebra (math.AC); Machine Learning (cs.LG); Algebraic Geometry (math.AG)
*备注:
Abstract:Motivated by theoretical problems in deep learning, we conjecture that post-composing a fixed number of pairwise distinct nonconstant polynomials with a generic polynomial of sufficiently large degree yields linearly independent polynomials. This generalizes Newman–Slater’s theorem on powers of polynomials. We establish several cases of this conjecture and its origin-passing variant: We prove the result for two polynomials, and for an arbitrary number of polynomials when their degrees are bounded. Furthermore, we show how the conjecture implies a complete understanding of the identifiability (i.e., parameter symmetries) of deep fully connected neural network architectures with generic polynomial activation functions. In particular, for network architectures with layer-specific activations of increasing degree, our established versions of the conjecture fully characterize the set of parameters yielding the same end-to-end network function. As a special case, we fully resolve the identifiability of shallow polynomial networks.
[LG-72] Representation Measurements Under Function-Preserving Reparameterizations
链接: https://arxiv.org/abs/2608.27020
作者: Abdullah Karasan
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Hidden coordinates are not uniquely determined by a language model’s input–output function, so representation-derived measurements should be invariant to function-preserving changes of basis. This study shows that column-permutation parallel analysis violates function-preserving reparameterization invariance because its reference distribution and selected component count can change while the model function and observed covariance spectrum remain fixed. More generally, a data-internal reference procedure cannot simultaneously preserve every coordinate marginal, remain orthogonally equivariant, and remove cross-coordinate covariance. Empirically, across five models, three retrieval domains, and 75 transformations, median component-count disagreement is 0.79 and median fixed-threshold decision disagreement is 0.26. A centering-only control isolates the reference-driven effect, with 1,141 of 1,200 component counts changing despite an unchanged observed spectrum, whereas independent parallel analysis seeds change none of the corresponding decisions. By contrast, orthogonally invariant comparator scores remain numerically stable with similar held-out discrimination. Together, these results show that parallel analysis-derived component counts and decisions can reflect hidden-coordinate choice rather than a well-defined property of the model.
[LG-73] Why not to use the Gaussian kernel
链接: https://arxiv.org/abs/2608.26974
作者: Toni Karvonen,Chris J. Oates
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Numerical Analysis (math.NA); Statistics Theory (math.ST)
*备注:
Abstract:Kernels measure similarity or correlation in tasks such as regression and classification. The Gaussian kernel, other names of which include squared exponential and radial basis function kernel, is one of the most popular in Gaussian process regression. We argue that the Gaussian kernel is best avoided and should never be used as a default. The argument rests on two results demonstrating that the Gaussian kernel is extremely brittle. First, the Gaussian kernel gives rise to a conditional variance that is unrealistically small. If the variance is used to quantify predictive uncertainty, catastrophic overconfidence is almost inevitable. Second, a small variance goes hand in hand with numerical ill-conditioning, so that to use the Gaussian kernel in practice requires tricks such as nugget terms that effectively modify the underlying regression or classification model. These problems are caused by the unnatural smoothness of the Gaussian kernel, a fact we are far from the first to take notice of. The problem is not the Gaussian form itself but the analyticity of the kernel: Our argument is more broadly that analytic kernels are best avoided. For stationary kernels analyticity is essentially equivalent to an exponential decay of the spectral density.
[LG-74] Bridging short- and medium-range weather forecasting with machine learning
链接: https://arxiv.org/abs/2608.26822
作者: Timothy A. Smith,Mariah Pope,Sergey Frolov,Brett Basarab,Daniel Abdi,Paul Madden,Isidora Jankov
类目: Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning (cs.LG)
*备注:
Abstract:The National Oceanic and Atmospheric Administration (NOAA) employs independent prediction systems for distinct forecast products. While some separation is practical, we argue that combining short- and medium-range weather into a single prediction system would provide the public with a useful distillation of global weather and its impacts. To this end, we present Nested-EAGLE (Experimental Artificial intelligence Global and Limited-area Ensemble): a 0.25° global weather model with a 6 km refinement over the Contiguous United States (CONUS). The model achieves significantly lower mean-squared error in near-surface and low-level quantities over CONUS compared to NOAA’s Global Forecast System and High-Resolution Rapid Refresh (HRRR), while remaining competitive throughout the rest of the global atmosphere. We show that the skill gains for near-surface fields stem from incorporating high-resolution regional analysis data into training through the nesting process. Forecasts of precipitation amounts are less skillful than those from HRRR, owing to deterministic training. However, we show that Nested-EAGLE provides the most accurate forecasts of storm locations at longer leads, despite blurred extrema. Our results motivate future work to extend the skill gains beyond CONUS and improve precipitation representation.
[LG-75] Incremental Recommendation via Causal Models RECSYS’26
链接: https://arxiv.org/abs/2608.26804
作者: Athanasios Vlontzos,David Gustafsson,Michael O’Riordan,Ciarán M. Gilligan-Lee
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
*备注: Accepted at the CONSEQUENCES Workshop @ RecSys’26
Abstract:Recommendation impressions are a finite resource, hence delivering a recommendation to a user who would discover the content organically yields no incremental value and displaces other recommendations that could. We address this by extending an existing production recommendation model to a causal architecture using holdback data that is already collected as part of routine experimentation infrastructure, requiring no new data collection. A central challenge is that attribution windows differ between treated and holdback observations: treated users are attributed a stream within a short direct-response window, while holdback users are attributed organic streams over a multi-day window. This mismatch makes naive treatment-effect subtraction invalid. We resolve this with a dual-threshold targeting policy that delivers a recommendation only when the probability of a treated stream is high and the probability of organic stream is low. In a production-scale A/B test on millions of Spotify users, this policy reduces recommendation impressions by 7% with no statistically significant reduction in overall recommended content consumption. We further show that joint training with holdback data improves calibration of the treated head relative to the production baseline, and argue this can be taken as evidence that causal models learn more generalisable representations than models trained on observational data alone.
[LG-76] Neural Renormalization Group Flow for Percolation
链接: https://arxiv.org/abs/2608.26764
作者: Anaclara Alvez,Luca Camagna,Sergio Chibbaro,Cyril Furtlehner,François Landes,Gianluca Manzan,Lorenzo Mensi
类目: Disordered Systems and Neural Networks (cond-mat.dis-nn); Machine Learning (cs.LG)
*备注: 7 pages, 5 figures
Abstract:Machine learning offers a possible route to data-driven real-space renormalization when the relevant observables are nonlocal and difficult to prescribe explicitly. We explore this idea for two-dimensional site percolation developping a supervised, scale-shared neural architecture. The model recursively applies the same learned coarse-graining rule across scales, producing a latent field from which the crossing probability is predicted, while a corresponding fine-graining decoder reconstructs the largest-cluster mask. Trained only on small lattices, the model extrapolates to substantially larger systems, recovers the spanning cluster with high fidelity, and produces observables obeying the expected finite-size scaling near the critical point. We observe that to get such performance it is key that the learned latent representation exhibits critical fluctuations and scale-dependent flows consistent with the renormalization-group structure of percolation.
[LG-77] A Unified Descriptive-Complexity Framework for Model Selection under Correlated Designs
链接: https://arxiv.org/abs/2608.26618
作者: Yanhang Zhang,Wei Liu,Yuhong Yang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 73 pages, 14 figures, 4 tables
Abstract:Model selection becomes particularly challenging under strong predictor dependence and model-class uncertainty, especially when there are exponentially many models. We propose a Descriptive-Complexity Information Criterion (DCIC) that regularizes large candidate model collections through Kraft-admissible code lengths. Under sub-Weibull noise, we establish selection consistency through approximation-error separation without relying on RIP-type conditions, together with nonasymptotic oracle risk bounds that remain valid under model misspecification. The same coding principle places heterogeneous classes on a common complexity scale at a small additional class-identification cost. This extension yields class–model recovery under suitable identifiability conditions and risk adaptation across classes. We further develop a complexity-guided search path that makes the computation–statistics trade-off explicit. Large penalties yield polynomial-size retained search regions with high probability, whereas smaller penalties sharpen the oracle risk benchmark. Numerical experiments illustrate stable support recovery and favorable estimation performance under strong dependence and model-class uncertainty.
[LG-78] Dynamical phase selection controls compute scaling in looped transformers
链接: https://arxiv.org/abs/2608.26556
作者: Gunn Kim
类目: Disordered Systems and Neural Networks (cond-mat.dis-nn); Statistical Mechanics (cond-mat.stat-mech); Machine Learning (cs.LG)
*备注: 5 pages and 3 figures
Abstract:A looped transformer performs inference by iterating a weight-tied map, making its computation a dynamical process whose cost is set by the resulting inference dynamics. Here we show that networks with identical architecture and objective, trained to identical accuracy, nevertheless realize distinct dynamical phases depending strongly on initialization, and that the bifurcation defining each phase determines how test-time compute scales. The phases are distinguished by their bifurcation mechanisms, including a saddle-node fold and a Neimark-Sacker-type transition to bounded nonstationary motion. In the fold phase, a one-dimensional normal-form reduction predicts both the relaxation-time and spectral-gap amplitudes from local derivatives of the trained map, yielding the parameter-free relation \tau(\varepsilon)[1-\lambda_\max(-\varepsilon)]\to\pi . Composed with a regular distribution of problem difficulty, the same critical slowing down produces the workload-level tail P(\tauN)\sim N^-2 . In the Neimark–Sacker phase, the fold scaling law disappears rather than merely changing its prefactor. Thus, test-time compute is not determined by architecture alone. It is governed by the dynamical phase of the solution found by training.
[LG-79] Interpreting Latent Protein Language Model Features with Geometric Annotations
链接: https://arxiv.org/abs/2608.26419
作者: Siddharth Setlur,Djordje Mihajlovic,Darrick Lee
类目: Quantitative Methods (q-bio.QM); Machine Learning (cs.LG)
*备注: 25 pages, 6 figures
Abstract:Protein language models (pLMs) encode information about protein sequences which enable downstream tasks such as structure prediction, but their internal representations are not well understood. Sparse autoencoders (SAEs) provide a promising tool to disentangle latent pLM representations into interpretable features, but existing annotation pipelines largely rely on protein-level annotations derived from database labels and LLM annotations of top activating sequences. Such annotations can overlook the localized residue-level and geometric patterns encoded by sparse features. We introduce an automated and scalable method for interpreting SAE features in ESM-2 by using geometrically inspired features of the protein \textC_\alpha backbone. Across ESM-2 8M layers, an FDR-controlled discovery analysis shows that local geometry is significantly associated with many SAE features, with varying levels of predictive strength, expanding coverage beyond database and sequence-based methods. In particular, geometry can distinguish SAE features sharing the same database annotation, revealing substructure within known biological labels. A significant portion of SAE features activate on unannotated metagenomic protein sequences enabling us to use our SAE annotations to better understand these sequences. In addition, ablation experiments at the level of contact prediction show that removing found geometric features shifts ESM-2’s predicted contact maps in the direction of the descriptor. This provides a robust method of annotating proteins activated within SAE neurons at a residue level, providing a bridge between mechanistic interpretability and structural biology.
[LG-80] owards a universal meta-optics solver via large language models
链接: https://arxiv.org/abs/2608.26417
作者: Huanshu Zhang,Lei Kang,Yuyan Chen,Luxiang Wang,Zhaolong Cao,Douglas H. Werner
类目: Optics (physics.optics); Machine Learning (cs.LG)
*备注: Accepted for publication in Nano Letters
Abstract:Metasurface design increasingly requires fast models that can operate across structurally distinct device families, rather than retraining a separate surrogate for every geometry class. Conventional neural network surrogates often depend on fixed-dimensional descriptors, family-specific output formats, and repeated architecture tuning, which limits their scalability across heterogeneous meta-atoms. Here, we present a unified large language model (LLM) workflow for multi-family metasurface modeling and inverse-design. Geometries, design parameters, and optical response channels were converted into a shared instruction-following text format and used to fine-tune Gemma-2-9B across 8 metasurface families. Compared with single-family baselines, the joint model simultaneously predicted the optical responses of all metasurface families while reducing the MSE for each family by an average of 56.5%. The same representation was also used for inverse design. These results show that a shared sequence-based LLM interface can provide a practical route to cross-family metasurface design while reducing the need for task-specific surrogate architectures.
[LG-81] Cross-simulator transfer with foundation model summaries: Towards robust SKA-era reionization inference
链接: https://arxiv.org/abs/2608.26354
作者: Yannic Pietschke,Caroline Heneka,Ayodele Ore,Romain Meriot
类目: Cosmology and Nongalactic Astrophysics (astro-ph.CO); Instrumentation and Methods for Astrophysics (astro-ph.IM); Machine Learning (cs.LG); Data Analysis, Statistics and Probability (physics.data-an)
*备注: 12 pages, 5 figures, prepared for submission to AA
Abstract:Simulation-based inference (SBI) for parameter estimation is vulnerable to model misspecification: neural summaries and density estimators trained on a specific forward model typically fail when applied to data drawn from another model, or from real observations, and no training simulator can capture the full observational pipeline of a real measurement exactly. We show that a self-supervised Vision Transformer (ViT), pretrained label-free on a fast approximate simulator, produces transferable data summaries that generalize across simulators. Without retraining, it can be reused as a frozen encoder to infer astrophysical parameters from a completely different simulator that resolves the radiative transfer explicitly, on which it has never seen either data or parameters. As a concrete use case in 21cm cosmology, SKATR, a ViT pretrained with a Joint Embedding Predictive Architecture (JEPA), serves as a foundation model for reionization inference from upcoming SKA measurements: SKATR is pretrained once on 67k low-cost, noiseless semi-numerical 21cmFAST lightcones, then frozen and applied to hydrodynamical Loreli II lightcones, where a lightweight conditional flow matching head infers five astrophysical parameters; the encoder is never shown Loreli data, its parameters, or any noise. In our comparison, SKATR yields the most precise and best-calibrated posteriors across all five parameters, matching the accuracy of the fully-supervised in-domain baseline while requiring 2.6x fewer radiative-transfer simulations. Under realistic SKA AA* noise, only SKATR remains simultaneously accurate, informative, and calibrated, outperforming even a supervised baseline retrained from scratch on noisy data. Self-supervised pretraining on computationally efficient semi-numerical simulations is therefore a viable route to calibrated, simulator- and noise-agnostic reionization inference for the SKA-era.
[LG-82] District-Level Food Environment Indicators and Social Vulnerability in São Paulo
链接: https://arxiv.org/abs/2608.26299
作者: Pedro Lemes Sixel Lobo,Eric Tokuda,Kuruvilla Joseph Abraham,Roberto Fray,Dirce Maria Marchioni,Alexandre Cláudio Botazzo Delbem,Rogerio Salvini
类目: Physics and Society (physics.soc-ph); Machine Learning (cs.LG); Applications (stat.AP)
*备注:
Abstract:Urban food environments may reflect broader socioeconomic inequalities, but district-level evidence remains limited in Brazilian cities. This study examined whether indicators of food retail and street-market availability discriminate between levels of social vulnerability across the 96 districts of São Paulo. We conducted an exploratory cross-sectional ecological analysis integrating the São Paulo Social Vulnerability Index (IPVS), establishment records from the Relação Anual de Informações Sociais (RAIS), and street-market data from CAISAN. Census-sector information was aggregated at the district level. Twenty districts without an IPVS classification were excluded, resulting in 76 observations. The outcome distinguished districts classified as IPVS level 1 from those classified as levels 2–7. Predictors described the densities of healthy and unhealthy food establishments, the number of street markets, and the availability of establishments selling fresh or in natura food. Eight conventional machine-learning classifiers were evaluated using leave-one-out cross-validation. Reported mean F-scores ranged from 0.62 to 0.75, with XGBoost obtaining the highest value. In the Random Forest model, the densities of healthy and unhealthy food establishments jointly accounted for approximately 60% of the total impurity-based feature importance. These findings indicate that publicly available food-environment indicators contain information associated with the district-level distribution of social vulnerability. However, the small ecological sample, class imbalance, outcome binarization, and cross-sectional design limit predictive generalization and preclude causal or household-level interpretations.
[LG-83] A causal graph-informed temporal convolution architecture for interpretable retail electricity price forecasting
链接: https://arxiv.org/abs/2608.26234
作者: Yufan Ji,Abdollah Shafieezadeh,Noah Dormady
类目: Applications (stat.AP); Machine Learning (cs.LG)
*备注:
Abstract:Retail electricity markets in deregulated systems face significant price volatility and complex interactions with forward and futures products, posing challenges for effective operational decision-making. This study introduces a Causal Graph-Informed Temporal Convolutional Network (CG-TCN), a forecasting architecture that integrates a learned causal graph into a temporal convolutional network via a graph-neural embedding to enhance both forecasting accuracy and interpretability of retail electricity price dynamics. It first applies a multi-resolution decomposition to isolate semiannual, quarterly, and monthly trends from high-frequency fluctuations. A causal graph is then discovered over these components and key covariates, including wholesale forward prices and retail contract attributes such as early termination fees, with domain constraints that preserve causal directionality and exogeneity. The learned causal structure is encoded as an adjacency embedding that conditions the TCN’s convolutions and attention, aligning representation learning with causal pathways. Using ten years of daily 12-month fixed-price residential contracts from Ohio’s deregulated market, we find that wholesale forward prices primarily determine long-term retail price trends, whereas contract attributes influence short-term fluctuations. CG-TCN consistently outperforms benchmark models, achieving mean absolute percentage errors of 3.08%, 3.82%, and 5.43% for one-, ten-, and fifteen-step-ahead forecasts of daily retail electricity median prices, respectively. By combining predictive performance with interpretability, CG-TCN provides transparent, policy-relevant insight to support market analytics, consumer protection, regulatory oversight, risk assessment and procurement planning in competitive electricity markets.
[LG-84] Classical and Hybrid Quantum Machine Learning for Trigger-Like Event Selection on CMS Open Data: An Eight-Qubit PCA-Constrained Benchmark
链接: https://arxiv.org/abs/2608.26224
作者: Tariq Mahmood,Muhammad Awais Rafique,Talab Hussain,Juan Pablo Perez Aguilar,Alfredo Raya,Muhammad Ahsan
类目: High Energy Physics - Phenomenology (hep-ph); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex)
*备注:
Abstract:Event triggering sits at the heart of high-energy physics, where the rare events of interest must be retained while an overwhelming background is discarded under tight latency and bandwidth budgets. This work compares four classical machine learning models, namely a support vector machine, an artificial neural network, a convolutional network and a long short-term memory network, with four hybrid quantum counterparts, on a trigger-like binary classification task built from CMS open data. The label is defined by an invariant-mass window, and the inputs combine reconstructed kinematics with physics-motivated derived variables: the pseudorapidity difference, the wrapped azimuthal difference, the angular separation and the total transverse momentum. The quantum models run under a fixed resource budget of eight qubits, a principal-component compression to sixteen features and state-vector simulation. Every model shares the same stratified split, the same preprocessing and a common decision threshold, and performance is reported through accuracy, ROC-AUC, F1-score, precision and recall. The strongest classical model is the artificial neural network, at 93.53 percent accuracy and 0.9819 ROC-AUC, while the strongest quantum model is the quantum convolutional network, at 90.89 percent accuracy and 0.9731 ROC-AUC, with the quantum neural network close behind. The quantum-kernel and recurrent quantum approaches trail both, which places the trainable hybrid embeddings ahead within this budget. The study is meant as a controlled reference point rather than a claim of quantum advantage.
[LG-85] RACE: Retrospective Streaming Generation of Physical Fields under Sparse Structured Sensing AAAI2027
链接: https://arxiv.org/abs/2608.26219
作者: Xinyu Zhang,Lihao Chen,Panqi Chen,Lei Cheng,Ting Zhang,Jianlong Li,Shikai Fang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 8 pages, 4 figures, 1 table (main text); 11 figures, 15 tables in the 20-page appendix. Under review at AAAI 2027
Abstract:Reconstructing continuous physical fields from sparse measurements is central to scientific monitoring, inverse modeling, and digital-twin construction. Generative reconstruction has recently emerged as a promising paradigm for this task by learning data-driven physical priors that complete plausible full fields from limited observations. However, existing methods largely assume fixed, batch conditioning, whereas real sensing systems often produce structured streams: probes scan local regions, instruments observe moving fields of view, and communication constraints may leave entire frames missing. We propose TRACE, a retrospective streaming generative reconstruction framework for physical fields under structured sensing. TRACE performs approximate Bayesian inference in a learned continuous-coordinate latent space, converting sparse off-grid measurements into generative latent evidence, fusing it with a state-space temporal prior through Kalman-style filtering, and refining under-observed past frames via retrospective smoothing. Experiments on active matter, ocean sound-speed fields, and supernova simulations show that TRACE matches or surpasses frame-wise generative reconstructors, offline spatiotemporal methods, and streaming data-assimilation baselines in reconstruction quality under temporally sparse and spatially localized sensing protocols.
[LG-86] Real-time virtual circuits for plasma shape control via neural network emulators: integration and testing in the MAST-U PCS ICIP
链接: https://arxiv.org/abs/2608.26216
作者: Matthew J. Marshall,Edward Jones,Graham J. McArdle,Alasdair Ross,Kamran Pentland,Nicola C. Amorisco,Charles Vincent,Martin Kochan,Colin Hogben,Graham Jones,Adam Stephen,George K. Holt,Adriano Agnello
类目: Plasma Physics (physics.plasm-ph); Machine Learning (cs.LG)
*备注: 8 pages, 3 figures, presented at 15th Technical Meeting on Control Systems, Data Acquisition, Data Management and Remote Participation in Fusion Research and submitted for review at Fusion Engineering and Design on July 31 2026
Abstract:The deployment of advanced, AI-enabled control algorithms in tokamak experiments requires robust integration with existing plasma control system (PCS) architectures and extensive pre-experimental validation. In this contribution, we describe the integration and testing of neural-network-emulated virtual circuits for plasma shape control within the MAST Upgrade (MAST-U) PCS environment. The neural network models predict the plasma shape using the plasma current, poloidal field coil currents, and plasma profile parameters. In this paper, we explain how they are deployed via a real-time C++ inference server that interfaces with the PCS, returning the shape prediction and its Jacobian, and how, from the latter, virtual circuit matrices and updated coil current requests are computed for real-time actuation. Emphasis is placed on the validation workflow and best practices adopted to ensure confidence in the proposed control framework prior to experimental deployment. This work demonstrates practical AI-based shape control components for fusion control systems, with direct relevance for upcoming MAST-U experiments and future devices.
[LG-87] Generative Monte Carlo Sampling for Constant-Cost Particle Transport
链接: https://arxiv.org/abs/2512.13965
作者: Joseph A. Farmer,Aidan Murray,Johannes Krotz,Ryan G. McClarren
类目: Computational Physics (physics.comp-ph); Machine Learning (cs.LG)
*备注: 10 pages, 4 figures
Abstract:We present Generative Monte Carlo (GMC), a novel paradigm for particle transport simulation that integrates generative artificial intelligence directly into the stochastic solution of the linear Boltzmann equation. By reformulating the cell-transmission problem as a conditional generation task, we train neural networks using conditional flow matching to sample particle exit states, including position, direction, and path length, without simulating scattering histories. The method employs optical coordinate scaling, enabling a single trained model to generalize across any material. We validate GMC on two canonical benchmarks, namely a heterogeneous lattice problem characteristic of nuclear reactor cores and a linearized hohlraum geometry representative of high-energy density radiative transfer. Results demonstrate that GMC preserves the statistical fidelity of standard Monte Carlo, exhibiting the expected 1/\sqrtN convergence rate while maintaining accurate scalar flux profiles. While standard Monte Carlo computational cost scales linearly with optical thickness in the diffusive limit, GMC achieves constant O(1) cost per cell transmission, yielding order-of-magnitude speedups in optically thick regimes. This framework strategically aligns particle transport with modern computing architectures optimized for neural network inference, positioning transport codes to leverage ongoing advances in AI hardware and algorithms.
附件下载
