本篇博文主要内容为 2026-09-10 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 (2026-09-10)
今日共更新653篇论文,其中:
- 自然语言处理共97篇(Computation and Language (cs.CL))
- 人工智能共153篇(Artificial Intelligence (cs.AI))
- 计算机视觉共107篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共165篇(Machine Learning (cs.LG))
- 多智能体系统共12篇(Multiagent Systems (cs.MA))
- 信息检索共11篇(Information Retrieval (cs.IR))
- 人机交互共21篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] Avatar: Toward Autonomous End-to-End Orchestration of Scientific Workflows using LLM s
【速读】：该论文旨在解决科学工作流管理系统（Scientific Workflow Management Systems, WMSs）在自动化执行过程中依赖固定、人工调优规则所带来的灵活性与适应性不足的问题。尽管大语言模型（LLM）代理有望实现更自主的工作流编排，但其应用仍面临关键挑战：如何确定代理推理的引入位置、如何控制其潜在风险，以及在何种场景下能真正提升系统性能。为此，论文提出Avatar架构，采用基于角色（actor-based）的设计，包含编排器（orchestrator）、执行器（executor）和溯源监控器（provenance monitor）三个核心组件。其核心创新在于通过单一适配器验证的动作目录（adapter-validated action catalog），使每个角色的决策策略可插拔——既支持传统规则驱动模式，也支持基于LLM的智能推理模式，从而在不同WMS中统一运行核心逻辑。实验基于Academy框架在三个工作负载上进行评估，结果表明：在规则模式下，Avatar能够完全复现原生执行行为，且仅需一个不变的核心即可支撑所有任务；而在LLM驱动模式下，系统显著降低55%的计算资源浪费，并将GPU繁忙时间减少40%。综上，Avatar的关键突破在于实现了规则与智能代理控制的无缝集成，为构建能够自我推理编排而非遵循预设规则的工作流系统提供了可行路径。
链接: https://arxiv.org/abs/2609.10509
作者: Suman Raj,Hai Duc Nguyen,Haochen Pan,Ryan Chard,Kyle Chard,Ian Foster
机构: University of Chicago (芝加哥大学); Argonne National Laboratory (阿贡国家实验室)
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Multiagent Systems (cs.MA)
备注: 8 pages, 6 figures, 1 table
Abstract:Scientific workflow management (WMSs) systems automate execution, yet orchestrate using fixed, hand-tuned rules. LLM agents promise more autonomous orchestration, but it remains unclear where to introduce agentic reasoning, how to bound its risk, and when it actually helps. We present Avatar, an actor-based architecture comprising an orchestrator, an executor, and a provenance monitor. Each actor’s decision policy is pluggable (rule-based or LLM-backed) via a single adapter-validated action catalog, so conventional and agentic control run on the same core across different WMSs. We present an implementation using the Academy framework and evaluate Avatar across three workloads. We observe that Avatar’s rule mode reproduces native execution, with a single unchanged core running all three. Moreover, LLM-backed Avatar reports a reduction of compute wastage by 55% and cuts GPU-busy time by 40% . Overall, we envision Avatar as a step toward workflow systems that reason about their own orchestration rather than follow pre-fixed rules.
[MA-1] Glyph: A Multi-Strategy Agent ic System for Column Description and Sensitivity-Ontology Tagging of Enterprise Data Catalogs
【速读】：该论文旨在解决企业数据湖中因数据表增长速度远超人工治理能力，导致列级描述缺失与治理标签未分配的问题，即“文档债务”（documentation debt），这一问题严重制约了数据发现、访问控制及合规性。其解决方案的核心在于提出Glyph系统，通过将列描述生成与列类型标注两个耦合任务建模为由状态化图（stateful graphs）协调的多智能体（multi-agent）架构。关键创新点包括：1）描述生成器（Descriptor）基于列的上游源代码（从企业GitHub按需检索），通过“推理-行动”循环实现主动增强型生成（active Retrieval-Augmented Generation），确保生成内容可追溯；2）标签器（Tagger）采用并行运行的三种互补策略——描述标签器、业务线正则表达式标签器与基于微调对比编码器（fine-tuned contrastive encoder）的元数据标签器，结合递归排名融合（Reciprocal Rank Fusion, RRF）对各策略输出进行加权融合；3）针对元数据编码器引入基于批内对比学习的目标函数，使同标签检索性能在内分布保留测试集上，NDCG@10从0.55提升至0.92，MAP@100从0.19跃升至0.90。此外，系统具备无值依赖设计、每标签溯源机制及渐进降级能力，显著区别于以往工作及商用敏感度扫描工具，实现了生成式AI驱动的元数据治理在生产环境中的可审计性与可操作性。
链接: https://arxiv.org/abs/2609.10430
作者: Kostia Kudriavtsev,Parvez Rafi,Sha Sundaram
机构: 未知
类目: Multiagent Systems (cs.MA); Information Retrieval (cs.IR)
备注:
Abstract:Enterprise data lakes accumulate tables faster than human stewards can document or classify them, leaving columns with missing descriptions and unassigned governance labels. This documentation debt undermines data discovery, access control, and regulatory compliance. We present Glyph, a production system that frames two coupled problems, column description generation and column type annotation for data classification, as cooperating LLM agents orchestrated as stateful graphs. The Descriptor grounds generation in the pipeline source code that produces each column, retrieved on demand from an enterprise GitHub via a reasoning–acting tool loop (active Retrieval-Augmented Generation). The Tagger assigns labels from a governed 275-leaf Data Classification Ontology by running three complementary strategies in parallel (a description tagger, a line-of-business regex tagger, and a metadata tagger backed by a fine-tuned contrastive encoder over a vector database), then fuses their ranked outputs with Reciprocal Rank Fusion (RRF). We fine-tune a 6-layer MiniLM metadata encoder with an in-batch contrastive objective, lifting same-tag retrieval on an in-distribution held-out split from NDCG@10 0.55 to 0.92 (MAP@100 0.19 \rightarrow 0.90 ) relative to the stock base encoder. We report end-to-end multi-label tagging quality under a recall-weighted F2 objective across three evaluation groups, an ablation isolating each strategy and the RRF fusion, and the engineering decisions that distinguish Glyph from prior column-type-annotation work and from commercial value/regex sensitivity scanners: value-free and code-grounded design, per-tag provenance, and graceful degradation. Together these make multi-agent LLM cataloging auditable and operable as a production service.
[MA-2] A traffic management system for large and heterogeneous vehicles in narrow industrial environments
【速读】：该论文旨在解决高密度工业环境中自动化导引车（AGV）协同调度的难题，传统基于协商的优先级分配方法在复杂非标准化场景下易引发效率低下与交通冲突。其核心解决方案是提出一种基于终身多智能体路径规划（L-MAPF）算法的新型AGV交通管理系统，该系统依托非均匀有理B样条（NURBS）生成的道路图实现对复杂工业布局的精确建模。关键创新在于将改进的有界时域冲突消解搜索（CBS）方法嵌入滚动时域冲突消解策略中，并引入动态扩展的时间窗机制以增强对狭窄双向通道内高密度交通流的冲突处理能力。相比现有先进方法，该方案专为非网格化、异构且高密度的现实工业环境设计，支持不同尺寸与性能的AGV并行运行。主要贡献包括：具备自适应时间窗调节能力的任意时间冲突消解策略、支持安全合规的实时执行层接口，以及先进的死锁检测与恢复机制。实验结果表明，在真实工业场景中，系统相较传统规则驱动管理与前沿优先级式L-MAPF变体，可实现最高达11%的吞吐量提升，同时保障连续稳定运行与更高运营效率。
链接: https://arxiv.org/abs/2609.10400
作者: Alessandro Bonetti,Silvia Proia,Simone Guidetti,Lorenzo Sabattini
机构: University of Modena and Reggio Emilia (摩德纳与雷焦艾米利亚大学); University of Bologna (博洛尼亚大学)
类目: Robotics (cs.RO); Multiagent Systems (cs.MA)
备注:
Abstract:The coordination of Automated Guided Vehicles (AGVs) in high-density industrial environments represents a critical challenge within Logistics 4.0, as traditional traffic management methods often lead to inefficiencies caused by negotiation-based priority assignment. To overcome the resulting limitations, this paper presents an innovative AGV traffic management system based on a Lifelong Multi-Agent Path Finding (L-MAPF) algorithm operating on roadmaps generated with Non-Uniform Rational B-Splines (NURBS) curves. The approach guarantees locally optimal coordination and ensures safe operation of large and heterogeneous AGVs. Building on this concept, the proposed framework integrates a modified version of the Bounded Horizon Conflict Based Search (CBS) technique within a Rolling Horizon Conflict Resolution strategy, utilizing an extended time horizon for each agent to enable effective conflict resolution in corridors identified by a topological map. In contrast to state-of-the-art methods for AGV fleet traffic management, the proposed solution is designed for real-world, non-standardized (i.e., non-grid-like) industrial settings characterized by narrow bidirectional corridors and high-traffic density, where AGVs of various sizes and capabilities operate simultaneously. Key contributions include an anytime conflict resolution strategy with adaptive time horizon regulation, an execution layer for safe and standard-compliant interaction with real AGVs, and an advanced mechanism for deadlock detection and resolution. Experimental results obtained in realistic industrial environments demonstrate higher throughput, with improvements of up to 11% over a conventional rule-based traffic management system, a state-of-the-art industrial method, and a priority-based L-MAPF variant, while maintaining continuous operation and improved efficiency.
[MA-3] MOONWALK: Mediating Operations with Intent-Evidence-Action Alignment Across Junior-Supervisor Review Workflows in Animation/VFX Pre-Production
【速读】：该论文旨在解决动画与视觉特效（VFX）前期制作评审过程中，创意意图在跨团队协作中因信息模糊、迭代过程中的标准漂移以及决策依据缺失而导致的执行偏差问题。其核心挑战在于如何将松散表述的创意需求（如简报、动态规格、异构参考素材及口头决策）转化为可被初级艺术家清晰理解并执行的具体修改任务，同时确保评审判断具备可追溯的证据支撑。解决方案的关键是提出一种“意图-证据-行动”对齐的设计框架：通过建立共享的项目意图记录，将评审判断锚定在具体的参考文档或规格说明上，并由主管授权将决策转化为明确的修订任务，直接关联原始注释。该框架在名为MOONWALK的专业预生产评审系统中实现，集成共享意图记录、参考/规格锚定、结构化在制品对比及主管授权的任务规划功能。在此流程中，生成式AI（Generative AI）承担行政协调职责（如识别缺失上下文、组织笔记），而创作者仍保有完整的审美主导权与最终优先级决策权。基于工作室环境的实证研究显示，相较于仅使用聊天机器人（chatbot）的非结构化对话界面，MOONWALK显著提升了意图一致性、决策可追溯性与检查清单可执行性，验证了结构化工作流相较于非结构化对话式AI在专业创作场景中的优越性。
链接: https://arxiv.org/abs/2609.10385
作者: Shih-Yu Lai,Wen-Fan Wang,Sai Ling,Shaune Jan,Bing-Yu Chen,Xiang Anthony Chen
机构: National Taiwan University(国立台湾大学); MoonShine Animation Studio(月光动画工作室); Cornell Tech(康奈尔科技); HCI Research, UCLA(人机交互研究，加州大学洛杉矶分校)
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Graphics (cs.GR); Multiagent Systems (cs.MA)
备注:
Abstract:Animation and VFX pre-production review requires teams to translate loosely specified creative intent–briefs, evolving specifications, heterogeneous references, and verbal decisions–into revisions that junior artists can execute without repeated clarification. In practice, criteria drift across iterations, review judgments lose their evidential basis, and the reasoning behind a request rarely survives the senior-junior handoff. We contribute a design framework for intent-evidence-action alignment: intent is articulated into a shared project record, judgments are anchored to grounded evidence, and authorized decisions are converted into clear revision tasks tied directly to reference notes. We instantiate this framework in MOONWALK, a professional pre-production review system comprising a shared intent record, reference/specification anchoring, structured work-in-progress comparison, and supervisor-authorized action planning. In this workflow, AI handles administrative coordination–flagging missing context and organizing notes–while artists retain full creative direction. An in-studio study with professional practitioners compares MOONWALK with a chat-only (chatbot) interface using matched production materials, while participants’ existing workflows provide a retrospective ecological baseline. Results indicate stronger intent alignment, decision traceability, and checklist executability, while also showing that aesthetic authority and final prioritization must remain with practitioners. The evaluation establishes the value of the integrated structured workflow over unstructured conversational AI chatbot. Code: this https URL
[MA-4] How neighbourhood ideology shapes misinformation belief in densely tied social networks
【速读】：该论文旨在解决现有基于主体的模型在模拟虚假信息传播时，未能有效整合个体意识形态偏见对信念形成影响的问题。传统模型虽关注社会影响作为虚假信息扩散的核心驱动力，但忽略了个体意识形态强度及其与虚假信息契合度之间的动态交互作用。本文提出的关键解决方案在于构建一个能够同时表征个体意识形态偏见强度和虚假信息与其意识形态契合程度的动态模型，并通过网络结构中的社会互动机制，刻画社会影响如何调节意识形态强度与信息契合度的共同作用。研究发现，邻近个体的意识形态强度对信念的影响显著依赖于其在网络中的连接紧密程度，揭示了网络拓扑结构与个人意识形态偏见在虚假信息传播中的协同效应。因此，该研究强调，在建模虚假信息传播时，必须综合考虑网络结构特征与个体意识形态因素。
链接: https://arxiv.org/abs/2609.10277
作者: Soroush Karimi,Marcos Oliveira,Diogo Pacheco
机构: University of Exeter (埃克塞特大学); Vrije Universiteit Amsterdam (阿姆斯特丹自由大学)
类目: ocial and Information Networks (cs.SI); Multiagent Systems (cs.MA); Physics and Society (physics.soc-ph)
备注: 18 pages, 9 figures, 1 table
Abstract:With the rapid spread of news on social media, understanding the propagation of misinformation is becoming increasingly important. One factor that affects individuals’ vulnerability to false information is their ideological predisposition. Despite the large number of agent-based models that focus on social influence as a driver of the spread of false claims, they often fail to explicitly integrate personal ideological biases into belief formation. In this work, we explore how misinformation spreads through the interaction between individuals’ ideological biases and social influence. Our model accounts for both the strength of individuals’ ideological biases and the extent to which a false claim aligns with their ideology. Social influence modifies the effects of ideological intensity and false claim alignment through network interactions. Notably, the influence of neighbours’ ideological intensity on belief is strongly affected by how well those neighbours are connected to one another. These results highlight the importance of considering both network structure and personal ideological biases when modelling misinformation propagation.
[MA-5] UnitBoost: Managing Compound LLM Systems with a Merge Operator Not a Model
【速读】：该论文旨在解决复合大语言模型（Compound LLM）系统中由高层管理代理（meta-agent）带来的协调难题，尤其是其在生成式决策过程中集中了三项控制决策（输出整合、任务分配与终止判断），导致行为不透明且对调用顺序敏感的问题。其核心解决方案是提出一种名为UnitBoost的非生成式元级操作算子，取代传统的生成式管理器。该算子的关键在于：通过任务给定的单元映射（unit map）将工作者输出转化为槽值提案，采用受限的argmax机制组装最终输出，并将未填充或无支持的槽作为显式的残差（residual）传递至下一轮；该设计实现了顺序无关性（order-free）、单元溯源记录（unit provenance）以及可验证的优化保证——在无耦合约束条件下，基于相同准入得分的单元级最大化优于任意完整候选解的选择。实验表明，在三个独立基准测试上，该方法在任务得分上超越使用黄金标签选择的最佳单一候选者0.060–0.195绝对点，并优于输入匹配的生成式管理器0.048–0.076；仅替换管理步骤即可使六种复合系统配置提升0.013–0.182。在FanOutQA任务中，残差引导的迭代使细胞F1从0.4778提升至0.5524，且对照实验验证了真实残差优于随机目标和常规重读策略；同时，无需标签的供给信号可有效识别单轮无效后进入耗尽状态。研究进一步量化了三种无增益场景（不可分割单元、单元身份不可得、按输出单元计费的终点）下的性能瓶颈，并将跨单元耦合定义为修复成本。总体而言，该方案以牺牲语义自由度为代价，获得了顺序不变性、单元溯源能力及可测试的失败判定条件，显著提升了系统的可解释性与鲁棒性。
链接: https://arxiv.org/abs/2609.09815
作者: Xing Zhang,Guanghui Wang,Yanwei Cui,Mengdie Flora Wang,Peiyang He
机构: AWS Generative AI Innovation Center
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Multiagent Systems (cs.MA)
备注:
Abstract:Compound LLM systems often solve a coordination problem by adding a higher-level LLM. The resulting meta-agent reads workers’ outputs, writes the final answer, allocates later calls, and decides when to stop. It is expressive, but it also concentrates three control decisions in an opaque, order-sensitive model call. We ask whether the manager needs to be generative at all. UnitBoost replaces that model with a defined meta-level operator: a task-given unit map turns worker outputs into slot-value proposals, a constrained argmax assembles the output, and the slots left unfilled or unsupported become an explicit residual for the next round. The operator is order-free, records unit provenance, and gives a simple guarantee: without coupling constraints, unit-wise maximization under the same admission score dominates selection of any complete candidate. On three held-out benchmarks, it exceeds the best single candidate chosen with gold labels by 0.060-0.195 absolute task-score points and input-matched generative managers by 0.048-0.076. Replacing only the management step improves six compound-system configurations by 0.013-0.182. Residual-directed rounds raise FanOutQA cell F1 from 0.4778 to 0.5524; matched controls show that the true residual outperforms random targets and ordinary rereading, while a label-free supply signal flags exhaustion after one unproductive round. The same analysis measures three conditions in which no such gain is available (one indivisible unit, unavailable unit identity, and an endpoint that charges for every emitted unit) and quantifies cross-unit coupling as a repair cost. The manager gives up semantic freedom and gains order invariance, unit provenance, and testable failure conditions.
[MA-6] Who Are They to Each Other? Multi-Agent Reasoning for Speaker Relationship Inference
【速读】：该论文旨在解决从口语对话中推断说话人关系这一社会感知语音理解中的关键问题，尤其针对现有方法在监督训练成本高、难以扩展，以及大语言模型（LLM）在推理时对细微、分布式且多模态的关系线索缺乏结构化处理能力的局限。其核心解决方案是提出一种无需训练的多智能体推理框架，通过结构化智能体间的交互实现关系判断的生成、质疑与仲裁。该框架包含两种互补设计：一是多角色多智能体辩论（Multi-Role Multi-Agent Debate），赋予各智能体基于社会理论的不同角色或视角，以支持更丰富的推理；二是多智能体竞争（Multi-Agent Compete），通过成对仲裁机制淘汰较弱判断，保留最具说服力的结果。实验在Seamless Interaction数据集上验证了两种方法在不同模态设置下的有效性，结果表明其在多数场景下优于零样本及现有多智能体基线。人工评估进一步揭示该任务对人类亦具挑战性，尽管在含文本的设定下，大模型表现可超越人类标注者，但在纯音频场景下仍逊于人类，说明当前模型尚未充分捕捉声学线索。综上，研究证实结构化的推理时智能体交互能显著提升关系推断性能，但声学特征的建模仍是当前瓶颈。
链接: https://arxiv.org/abs/2609.09628
作者: Yaohan Guan,Yen-Ju Lu,Yuzhe Wang,Junhyeok Lee,Jesus Villalba,Laureano Moro Velazquez,Thomas Thebaud,Najim Dehak
机构: Johns Hopkins University (约翰霍普金斯大学)
类目: Multiagent Systems (cs.MA); Computation and Language (cs.CL); Sound (cs.SD)
备注:
Abstract:Inferring speaker relationships from spoken conversations is an important step towards socially aware speech understanding. However, this task remains underexplored, and supervised modeling is costly to train and scale. At the same time, existing inference-time LLM approaches provide limited structure for handling subtle, distributed, and multimodal relational cues that may support multiple plausible interpretations. To address these limitations, we introduce a training-free multi-agent reasoning framework that organizes inference through structured interaction among LLM agents, allowing relationship judgments to be proposed, challenged, and adjudicated without task-specific training. We instantiate this framework with two complementary designs. We propose Multi-Role Multi-Agent Debate as a task-specific adaptation of standard multi-agent debate for speaker relationship inference, assigning agents complementary roles or social-theory-grounded perspectives rather than a single undifferentiated viewpoint. In contrast, we introduce Multi-Agent Compete, a competition-based protocol that compares agent judgments through pairwise adjudication, eliminates weaker candidates, and retains the most defensible one. We evaluate these methods on the Seamless Interaction dataset across different modality settings, covering both binary classification and fine-grained relationship-detail prediction. Results suggest that they improve over zero-shot and existing multi-agent baselines in most cases. Human evaluation further suggests that this task is challenging even for people. LLM methods can sometimes outperform human annotators in text-included settings but are less competitive in the audio setting. Together, these findings suggest that relationship inference benefits from structured inference-time interaction among agents, while acoustic cues are not yet fully captured by current models.
[MA-7] Agent ic AI-enabled Semantic Commissioning of a Cognitive Digital Twin for Reconfigurable Manufacturing
【速读】：该论文旨在解决可重构制造系统中认知数字孪生（Cognitive Digital Twin, CDT）快速定制化部署的关键挑战。传统数字孪生（Digital Twin, DT）构建方法主要聚焦于几何重建，往往忽视了实现自主推理所必需的深层语义融合与功能互操作性。其解决方案的核心在于提出一种基于智能体（agent）的、由人工智能驱动的端到端自动化调试工作流，采用LangGraph作为多智能体编排引擎，实现双路径协同合成：语义路径通过检索增强生成（Retrieval Augmented Generation, RAG）从非结构化文档中提取技术规范，功能路径则利用模型上下文协议（Model Context Protocol, MCP）自主发现并绑定实时工业遥测数据。实验验证表明，在机器人加工单元中的应用实现了97.2%的平均精度（mAP），并将部署周期从数周缩短至平均2小时，标志着从人工脚本编写向自主编排范式的根本性转变。
链接: https://arxiv.org/abs/2609.09503
作者: Yangyang Liu,Xun Xu,Jan Polzer
机构: 未知
类目: Robotics (cs.RO); Multiagent Systems (cs.MA)
备注:
Abstract:Rapid bespoke commissioning of the Cognitive Digital Twin (CDT) is a major challenge in reconfigurable manufacturing. Traditional digital twin (DT) construction methods primarily focus on geometric reconstruction, often neglecting the deep semantic integration and functional interoperability necessary for autonomous reasoning. This paper proposes an agent-based, AI-driven workflow to automate end-to-end CDT debugging. The system utilises LangGraph as a multi-agent orchestration engine to achieve dual-path synthesis: the semantic path extracts technical specifications from unstructured documents using Retrieval Augmented Generation (RAG), while the functional path autonomously discovers and binds to real-time industrial telemetry data using Model Context Protocol (MCP). Experimental validation in a robotic machining cell demonstrates that the system achieves a mean average accuracy (mAP) of 97.2% in perception and reduces the deployment cycle from several weeks to an average of 2 hours, marking a paradigm shift from manual scripting to autonomous orchestration.
[MA-8] Networked Admissibility-Preserving Control for Directed Safe Coordination
【速读】：该论文旨在解决具有异构非对称输入约束的标量代理系统在固定加权有向图通信拓扑下，实现安全关键协调控制的问题。核心挑战在于：代理输出需始终位于一个共同移动的安全走廊（safety corridor）内，同时实际执行的物理输入必须满足各自不同的非对称边界限制，且系统需在分布式架构下保证稳定性与一致性。其解决方案的关键在于提出一种网络化可容许性保持控制（Admissibility-Preserving Control, APC）架构，其中包含可容许性保持输入实现（Admissibility-Preserving Input Realization, APIR） 与对数障碍坐标（logarithmic barrier coordinate） 构成的协同机制。该设计通过精确级联结构，使实现误差指数衰减，驱动非对称一致性的动态演化；同时，基于左佩龙向量（left Perron vector）和初始APIR失配，导出闭式障碍坐标极限，确保安全走廊与执行器区间在正向不变的前提下保持均匀裕度，并实现命令的有界性与指数一致性。进一步地，研究给出了方向特异性充分条件以保证正负控制需求不越界，并证明在强连通性及增益与兼容性条件下，部分钉定（partial pinning）策略可将恒定的障碍参考从非空知情子集传播至整个网络，从而确定整体安全走廊轨迹。非平衡权重示例验证了方向性判据的有效性及预测的集体运动行为。
链接: https://arxiv.org/abs/2609.09384
作者: Abhinav Sinha,Lohitvel Gopikannan,Shashi Ranjan Kumar
机构: University of Cincinnati (辛辛那提大学); Indian Institute of Technology Bombay (印度理工学院孟买分校)
类目: ystems and Control (eess.SY); Multiagent Systems (cs.MA); Robotics (cs.RO); Dynamical Systems (math.DS)
备注:
Abstract:This paper addresses safety-critical coordination for scalar agents whose distributed commands are implemented through constrained physical-input dynamics. Agents communicate over a fixed weighted digraph with a directed spanning tree, while their outputs must remain inside a common moving safety corridor and their realized inputs must satisfy heterogeneous asymmetric bounds. We propose a networked Admissibility-Preserving Control (APC) architecture in which an Admissibility-Preserving Input Realization (APIR) governs physical inputs and a logarithmic barrier coordinate represents the safety corridor. The synthesis yields an exact cascade in which exponentially decaying realization errors drive nonsymmetric consensus dynamics. For every compatible compact initial set, the closed-loop system admits a unique complete solution, renders the moving corridor and actuator intervals forward invariant with uniform margins, keeps commands bounded, and achieves exponential consensus. We derive direction-specific sufficient conditions under which positive and negative control demands remain within their corresponding actuator limits. The analysis yields a closed-form barrier-coordinate limit determined by the left Perron vector and initial APIR mismatch. Under strong connectivity and the stated gain and compatibility conditions, partial pinning propagates a constant barrier reference from a nonempty informed subset and assigns the induced safety corridor trajectory. A non-weight-balanced example illustrates the directional certificate and predicted collective motion.
[MA-9] Smart Adaptive Computing Across the Continuum: LLM s in IoT-Edge-Cloud Resource Management WWW
【速读】：该论文旨在解决在物联网（IoT）、边缘与云（Cloud Continuum）多层资源管理中，如何实现持续、上下文感知的决策问题，尤其针对动态变化的约束条件带来的挑战。现有深度强化学习（DRL）方法虽能有效应对此类复杂决策问题，而大语言模型（LLM）正被越来越多地用于增强DRL系统，但二者之间的架构关系尚未明确。为此，作者基于Wang等人提出的连续体编排系统（Continuum Orchestration Systems）分类框架，引入两个新维度：AI增强范式（AI Augmentation Paradigm）用于衡量LLM的使用方式，反馈通道（Feedback channel）则刻画执行反馈是否以及通过何种系统路径回传至LLM，以在LLM编排层闭合MAPE控制环。通过对六种近期系统架构的应用分析，研究发现当前方案普遍缺乏在云连续体环境下同时实现完整LLM编排与完整代理层反馈的整合设计。这一缺口可归因于跨层级反馈抽象机制的缺失，该机制对于弥合各层级间信号异构性与LLM编排器之间的语义鸿沟至关重要。
链接: https://arxiv.org/abs/2609.09348
作者: Antonino Vaccarella,Lanpei Li,Vincenzo Lomonaco,Massimo Coppola
机构: Institute of Information Science and Technologies “Alessandro Faedo” (ISTI), National Research Council of Italy (CNR), 56124 Pisa, Italy; Department of Computer Science, University of Pisa, 56127 Pisa, Italy; Department of AI, Data and Decision Sciences, LUISS University, 00197 Rome, Italy
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: This is the Accepted Manuscript, after peer review, not the Version of Record; no post-acceptance changes. Presented at FRAME 2026 (Euro-Par 2026). Subject to Springer’s AM terms of use: this https URL
Abstract:Managing resources across IoT, edge, and cloud layers calls for continuous, context-aware decisions under constraints that rarely stay fixed. Deep reinforcement learning (DRL) handles this class of problems well, and large language models (LLMs) are increasingly used to augment DRL pipelines, yet the architectural relationship between the two is seldom made explicit. We build on Wang et al.'s taxonomy of Continuum Orchestration Systems employing DRL techniques and extend it with two further dimensions. The AI Augmentation Paradigm measures how LLMs are exploited, while the Feedback channel captures whether and through which system path the execution feedback returns to the LLM in order to close the MAPE control loop at the LLM Orchestration layer. We apply this taxonomy to six recent system architectures and find a common gap, as none combines full LLM orchestration with full agent-layer feedback in a Cloud Continuum setting. We relate this gap to a missing cross-tier feedback abstraction, bridging the incommensurable per-tier signals and the LLM Orchestrator.
[MA-10] Scores Alone Do Not Prove Discovery: The Discovery Certification Protocol for Auditing AI Research Agents
【速读】：该论文旨在解决生成式AI研究中成果可复现性与可信度评估的难题，核心问题在于如何在缺乏透明机制的情况下验证AI研究代理（AI research agents）所宣称的有效性改进是否真实且可重复。其解决方案的关键在于提出“发现认证协议”（Discovery Certification Protocol, DCP），通过一系列结构化、可执行的验证门限实现对研究成果的严格审计：Gate 1 验证在封闭评估环境下是否存在实质性改进；Gate 2 在控制起始信息与观测网络内容的前提下，检验不同代理是否能独立达到目标并提供可恢复的证据，同时触发“核心否决机制”（Core veto）；DCP核心要求包括充分的对照实验、零观察到的可复现结果，以及在单次注册实验中可复现性的有限样本上界。此外，可选的Gate 3 通过独立的零假设校准，量化真实反馈相对于基准策略的平均效应，并结合注册效应裕度形成最终证据。实证部分在SQLite优化和虚拟催化剂控制等任务中成功实现了96轮零复现记录，上界为0.0468，且每对研究均显示30次真实可复现而0次中性可复现，通过了60对零假设检验。整个协议由一个无需大语言模型（LLM-free）的确定性验证器复现决策，确保结果透明、可审计。DCP构建了一个通用的证据语言体系，支持对有用成果、替代路径及反馈效应的标准化评估，为跨研究的可比性与可信性提供了技术基础。
链接: https://arxiv.org/abs/2609.09219
作者: Jingjie Ning,Shanshan Zhong,Xiaochuan Li,Ji Zeng
机构: 未知
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:AI research agents combine prior knowledge, public sources, and experimental feedback to produce useful results. The Discovery Certification Protocol (DCP) turns claims about these results into executable recovery and feedback tests. Gate 1 validates useful improvement on sealed evaluation. Gate 2 gives matched agents the registered starting information and observed Web content while withholding the target research history. Every valid method reaching the numerical target supplies a recovery witness and triggers the Core veto. DCP Core requires adequate controls, zero observed recoveries, and a finite-sample bound on recovery in one fresh registered episode. Optional Gate 3 measures the average effect of truthful feedback relative to a specified neutral policy from a shared checkpoint. DCP Evidence adds this effect after independent null calibration and a registered effect margin. Two controlled audits exercise the complete protocol in SQLite optimization and virtual catalyst control under different models. Each produced zero recoveries in 96 episodes, with an upper bound of 0.0468. Each paired study yielded 30 truthful recoveries and zero neutral recoveries, with passing 60-pair null studies. Additional cases exercise Core, recovered, and audit-incomplete decisions. A deterministic, LLM-free verifier reproduces the decisions from frozen evidence. DCP provides a common evidence language for useful outcomes, alternative routes, and feedback effects across AI research.
[MA-11] CompEvo: Competition-Induced Evolution for Multi-Agent in News-Driven Time Series Forecasting
【速读】：该论文旨在解决多智能体新闻驱动时间序列预测中的两个核心问题：一是“思维退化”（degeneration of thought）问题，即多个智能体在迭代过程中趋向于相似的证据搜寻行为，导致多样性丧失；二是缺乏理论基础的问题，现有策略更新机制多依赖启发式方法，缺乏可证明的优化框架。其解决方案的关键在于提出一种竞争诱导演化框架CompEvo，通过引入演化博弈论（evolutionary game）形式化建模，确保均衡存在性与优化收敛性，并构建一个可训练的多智能体演化系统，集成策略执行、基于适应度的可微选择以及竞争驱动的策略演化机制。该框架使异构智能体能够探索多样化的新闻证据，将预测反馈转化为可微的影响权重，并在竞争压力下持续演化策略，在保持有效逻辑的同时维持行为多样性。实验结果表明，CompEvo在四个真实数据集上平均降低RMSE 27.3%、MAPE 26.2%，且分析显示其成功维持了多样化且专业化的智能体行为模式。
链接: https://arxiv.org/abs/2609.09195
作者: Yuxuan Zhang,Yangyang Feng,Yong Guan,Daifeng Li,Kexin Zhang,Junlan Chen,Bowen Deng,Jun Liu,Zehua Zeng
机构: Sun Yat-sen University (中山大学); North China Electric Power University (华北电力大学); Unilumin Group Co., Ltd. (联创光电)
类目: Neural and Evolutionary Computing (cs.NE); Multiagent Systems (cs.MA)
备注:
Abstract:News-driven time series forecasting uses evolving textual events together with historical observations to predict future values, supporting applications such as market risk monitoring and resource scheduling. In multi-agent settings, two challenges still remain. The first is degeneration of thought, where agents converge to similar evidence-seeking behaviors. The second is insufficient theoretical grounding, where strategy updates are often heuristic and lack a principled formulation. To address the above challenges, we propose CompEvo, a competition-induced evolution framework for multi-agent news-driven time series forecasting. For theoretical grounding, we introduce an evolutionary game formulation to guarantee equilibrium existence and optimization convergence. Building on this formulation, we construct a trainable multi-agent evolution framework that integrates strategy execution, fitness-based differentiable selection, and competition-induced strategy evolution. CompEvo enables heterogeneous agents to explore diverse news evidence, converts forecasting feedback into differentiable influence weights, and evolves agent strategies under competitive pressure to preserve effective logic while maintaining diversity. Experiments on four real-world datasets show that CompEvo reduces RMSE by 27.3% and MAPE by 26.2% on average over strong baselines. Further analysis indicates that CompEvo successfully maintains diverse and specialized agent behaviors.
自然语言处理
[NLP-0] IdeaAMBIG: Benchmarking Implementation-Critical Gaps in Research-Idea Specifications
【速读】： 该论文旨在解决科研方法描述在可实现性（codification readiness）方面的不足问题，即尽管研究思想具有新颖性、逻辑一致性和科学合理性，但其提出的方法往往缺乏足够详细的说明，导致专业执行者或代码代理无法忠实复现。核心挑战在于：如何评估并提升科研方法规范的可编码化程度——即是否具备足够的方法学信息，使具备能力的实施者无需依赖未经证实的假设即可构建出预期方法。解决方案的关键在于构建基于证据的规范（evidence-grounded specifications）及其支持性解答，并引入IdeaAMBIG基准，涵盖660个实例（包括163个来自可重现性报告和GitHub问题的真实缺陷，以及497个注入到可编码参考文献中的受控合成缺陷），用于评估三个关键能力：可编码化准备度评估、缺陷定位与澄清动作生成。实验表明，当前大语言模型（LLM）在真实场景下的缺陷恢复率仅为9.6%，但在已知缺陷的前提下，澄清动作的成功率可达80.6%；而在“理想”条件下（提供黄金级解答），下游可编码化准备率从14%跃升至98%。结果揭示，缺陷定位是当前系统的主要瓶颈，而一旦获得缺陷标注，模型的澄清能力显著增强。
链接: https://arxiv.org/abs/2609.10539
作者: Yiling Ma,Yilun Zhao,Sihong Wu,Manasi Patwardhan,Arman Cohan
机构: Yale University (耶鲁大学); TCS Research
类目: Computation and Language (cs.CL)
备注: Preprint. 74 pages, 18 figures
Abstract:A research idea may be novel, coherent, and scientifically plausible, yet its proposed method may remain insufficiently specified for faithful implementation. We study the codification readiness of implementation-facing research-method specifications, defined by whether they provide sufficient methodological information for a competent implementer or coding agent to construct the intended method without unsupported assumptions. We construct evidence-grounded specifications and their supported resolutions from papers, codebases, issue threads, and reproduction artifacts. We introduce IdeaAMBIG, a benchmark of 660 evidence-grounded instances: 163 real-world gaps from reproducibility reports and GitHub issues, and 497 controlled synthetic gaps injected into codification-ready references. IdeaAMBIG evaluates three capabilities: codification-readiness assessment, defect localization, and clarification action generation. Defect localization receives only the specification, whereas clarification additionally receives the annotated defect. Across 13 LLMs, the best model achieves 9.6% Macro Defect Recovery Rate on real-world instances but 80.6% Macro Clarification Action Success Rate when given the defect. In an oracle study, supplying the gold resolution raises the downstream codification-ready rate from 14% to 98%. Across all evaluated models, defect localization is the main bottleneck, with stronger clarification given the defect.
[NLP-1] IBIB: A Protocol for Measuring Enterprise AI Systems by Serving Route Not Model Identifier
【速读】： 该论文旨在解决当前大模型评估体系中存在的核心问题：现有18项审计基准测试仅依赖广告模型标识符进行评分，而忽视了实际可用能力的多维度构成——包括权重、服务路径（serving route）、精度（precision）、输出契约（output contract）及工具调用框架（harness）等关键因素。这种“以标识符代能力”的做法导致测量误差，无法真实反映系统在复杂任务中的实际表现。其解决方案的关键在于提出一种名为IB2的协议，通过三个核心机制实现可报告、可验证的能力测量：第一，金盲能力绑定预检（gold-blind capability-binding preflight），在任务执行前验证服务路径是否满足评估契约；第二，包含可靠性的初筛评分规则（reliability-inclusive first-pass scoring rule），将失败响应保留在评分中但排除不支持的能力项；第三，结构化无偏评分仲裁机制（adjudication is structurally score-blind），确保评估过程不受评分结果影响。该协议将评估过程本身作为可复现的“范式”而非数据集，其参考实现包含128个锁定任务与987条关于文档、表格、图表、工具及数据库操作的断言。实验表明，同一权重在不同运行中表现出显著差异，且模型标识符无法揭示性能极限；此外，不同评估套件的性能饱和度分布不均，尤其受受控数据库操作和多标签关联的影响，因此采用区间支撑的分辨率分组替代传统排名，并对输出标签切割进行多重性调整。更重要的是，服务端选择（serving-arm choice）显著影响性能表现，例如某版本精度从77.38提升至82.54，区间跨度达[0.11, 10.60]，但该差异源于访问模式、工具生成方式及工具调用解析器的差异，而这些并非终端接口属性。最终，若剔除失败响应作为分母，点序会发生变化，说明可靠性纳入直接影响结论，而非仅改变表述。
链接: https://arxiv.org/abs/2609.10494
作者: Blake Stenstrom,Charangan Vasantharajan,Brian Sathianathan
机构: Iterate.ai(迭代人工智能)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 42 pages, 4 figures
Abstract:Enterprises deploy systems, not checkpoints. Usable capability depends jointly on weights, serving route, precision, output contract, and harness, yet all 18 audited benchmarks score advertised model identifiers. We treat this as measurement error and give a protocol that makes it reportable. It has three parts. A gold-blind capability-binding preflight verifies that a route can execute the evaluation contract before any task reaches it; a reliability-inclusive first-pass scoring rule keeps failure in the score while keeping unsupported capability out; and adjudication is structurally score-blind. We call the protocol IB2 and release its algorithms, classification tables, request contract, and manifest schemas. Its reference instantiation, 128 locked tasks and 987 assertions over document, spreadsheet, chart, tool and database work, stays sealed: the procedure is the artifact, not the corpus. Across eleven systems, four results. Capability availability is measurable: two complete single-route runs on identical weights later failed distinct predicates of the finalized binding gate, while a third passed that gate before a fresh run. The advertised identifier exposed neither limit. Discrimination is not uniform: four of seven suites saturate under a six-system band, with the spread almost entirely from governed database work and multi-tab joins, so we report interval-backed resolution groups, not ranks; two of the nominal five-label output’s four cuts fail multiplicity adjustment. Serving-arm choice moved one declared revision and precision from 77.38 to 82.54, paired interval [0.11,10.60], though the arms differ in access mode, harness generation, and the serving tool-call parser, and harness generation is a property of our evaluator, not any endpoint. Excluding failed responses from denominators changes the point ordering, so reliability inclusion changes a conclusion, not its wording.
[NLP-2] Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoning
【速读】： 该论文旨在解决当前生成式AI在复杂任务中普遍存在“英语中心化”（English-centric）的问题，即模型尽管可接收多语言输入，但其推理过程仍主要依赖英语，导致非英语用户难以获得准确、自然且语义一致的响应，同时可能丢失原始问题的文化或语言特异性信息。为应对这一挑战，论文提出并推进“第二语言推理”（L2 reasoning），即让模型能够根据用户提示的语言进行一致性推理，从而构建从提示到答案的“在地化桥梁”。其解决方案的关键在于采用数据驱动的方法，通过优化监督微调（SFT）中的数据构成与调度策略，实现跨语言推理能力的泛化。研究构建了3.35B参数规模的Tiny Aya L2-Thinker模型，在60种语言、涵盖数学、常识推理、指令遵循、开放式生成和文化推理等6个基准上的实验表明，该模型在绝大多数语言中实现了超过93%的L2推理准确率，同时保持了强性能。研究进一步揭示，将L2推理能力推广至未见语言的关键路径包括：扩大语言覆盖范围、利用广泛可得的多语言非推理数据，以及保留充分的英语推理基础。这些发现表明，推理是一种语言无关的行为，可通过精心设计的数据混合策略在类型差异显著的语言间迁移，而无需在每种目标语言中都提供专门的推理标注数据。论文开源了模型权重与多语言推理数据集，以推动可访问、在地化推理的研究发展。
链接: https://arxiv.org/abs/2609.10445
作者: Mehrnaz Mofakhami,Ananya Sahu,Alejandro R. Salamanca,Daniel D’souza,Alexandre Berard,Thomas Euyang,Marzieh Fadaee,Julia Kreutzer
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Reasoning language models have made substantial advances on a variety of complex tasks, yet their capabilities remain overwhelmingly English-centric: models primarily reason in English regardless of the language they are prompted in. This is inaccessible for non-English-speaking users, risks losing the intent of the original question, and forgoes knowledge more readily expressed in the target language. In this work, we advance L2 reasoning, the ability of a model to reason consistently in the language of the user’s prompt, thus building an in-language bridge between the prompt and the answer. We approach this problem from a data-centric angle, investigating how to optimize data composition and scheduling in SFT for reasoning generalization. Building Tiny Aya L2-Thinker at 3.35B scale, we achieve an L2 reasoning rate above 93% across 60 languages on 6 benchmarks spanning math, commonsense reasoning, instruction following, open-ended generation, and cultural reasoning while keeping performance strong. We show the path to generalizing L2 reasoning to held-out languages goes through broader language coverage, readily available multilingual non-reasoning data, and a sufficient English reasoning backbone. These findings indicate that reasoning is a language-agnostic behavior that can be transferred across typologically diverse languages through careful data mixing and without requiring reasoning supervision in every target language. We release our model weights and multilingual reasoning data to support further research on accessible, in-language reasoning.
[NLP-3] ConvMem: Convolutional Memory for Long-Context Reasoning
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在处理超长上下文时因固定上下文长度限制而导致的性能瓶颈问题。现有方法如MemAgent采用序列化处理方式，虽通过分段读取并迭代更新固定大小记忆来扩展有效上下文，但其固有的串行计算模式导致高延迟，且依赖代价高昂的强化学习（Reinforcement Learning, RL）训练，易在特定数据集上过拟合。为此，本文提出ConvMem，一种无需训练、高度可并行化的框架，将长上下文推理重构为分层卷积结构。其核心创新在于借鉴卷积神经网络（Convolutional Neural Networks, CNNs）思想，将提示特定查询的大语言模型视为卷积核，通过可配置步幅（Configurable Strides）与跳跃连接（Skip Connections）实现证据的稳健捕获与跨层级传播，并利用多核卷积（Multi-Kernel Convolution）将复杂查询分解为解耦的语义通道，从而显著降低误差累积风险。该设计不仅将推理路径从线性链式结构压缩为对数级树状结构，还实现了文本段落与推理线程层面的大规模并行化。实验在RULER-HotpotQA和RULER-2WikiMultiHopQA数据集上验证了ConvMem在保持训练自由的同时，优于现有无训练基线方法，并有效规避了基于参数先验的强化学习模型在分布外任务中常见的过拟合问题。
链接: https://arxiv.org/abs/2609.10441
作者: Hongming Zhang,Zhaozhen Gu,Fengshuo Bai,Ming Hao,Qingyang Zhang,Yuanyuan Wang,Shiyang Tang,Yanna Wang,Bo Xu
机构: Institute of Automation, Chinese Academy of Sciences(自动化研究所，中国科学院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:While Large Language Models (LLMs) have demonstrated impressive capabilities, they often struggle with extremely long contexts due to fixed context limits. To address this, sequential approaches like MemAgent extend the effective context by reading text in segments and iteratively updating a fixed-size memory. However, this sequential paradigm suffers from high latency and requires costly reinforcement learning (RL) training, which can lead to overfitting on specific datasets. To overcome these limitations, we propose ConvMem, a training-free, highly parallelizable framework that reformulates long-context reasoning as a hierarchical convolution. Inspired by CNNs, ConvMem treats an LLM prompted with a specific query as a convolutional kernel. This kernel summarizes text segments hierarchically, shortening the reasoning path from a linear chain into a logarithmic tree. Specifically, ConvMem integrates \textitConfigurable Strides and \textitSkip Connections to ensure robust evidence capture and propagation, while employing \textitMulti-Kernel Convolution to decompose complex queries into disentangled semantic channels. This design not only mitigates error accumulation but also enables massive parallelization across both text segments and reasoning threads. Experiments on RULER-HotpotQA and RULER-2WikiMultiHopQA demonstrate that ConvMem outperforms training-free baselines and avoids the risk of overfitting to parametric priors often observed in RL-trained models on out-of-distribution tasks.
[NLP-4] Do speech foundation models really learn words? INTERSPEECH2026
【速读】： 该论文旨在解决自监督语音基础模型（self-supervised speech foundation models）在下游任务中表现优异但其内在表征机制尚不清晰的问题，尤其关注这些模型是否真正学习到了独立于语音形式的词汇语义或句法信息。传统研究多聚焦于模型对音素和词的判别能力，但这种判别能力可能仅源于对语音形式（如音素）的良好编码，而非对词汇本身（如身份、语义或句法属性）的独立表征。为此，论文提出一种基于残差化（residualization）的解耦方法，通过部分消除语音内容中的音素信息，验证了在深层网络中，HuBERT与wav2vec 2.0确实能够学习到与局部语音内容解耦的词汇表示，具备合理的词汇独立表征能力。该解决方案的关键在于利用残差化实现音素信息的剥离，从而揭示模型中潜在的更高阶语言信息，并显著提升词汇发现任务中的表现，为理解自监督语音模型如何编码语言结构提供了新的实证依据。
链接: https://arxiv.org/abs/2609.10434
作者: Robin Huo,Ewan Dunbar
机构: 未知
类目: Computation and Language (cs.CL); Sound (cs.SD)
备注: Proceedings of Interspeech 2026
Abstract:Self-supervised speech foundation models are now used in a wide array of downstream applications, including traditional speech recognition and as the basis for tokens in speech-aware language models. Attempts to understand their usefulness have largely focused on probing their representations’ ability to discriminate phonemes and words. However, discriminative ability for words need not imply specialized representation of words per se. Good discrimination of words may be explained by good encoding of word form (phonemes) rather than form-independent word representations encoding identity or syntactic/semantic properties. By partialling out phoneme information using residualization, we show that, in later layers, HuBERT and wav2vec 2.0 do in general learn representations which encode words with reasonable fidelity independently of local phonetic content. We show that this simple approach to disentanglement can enhance higher-order linguistic information in word discovery tasks.
[NLP-5] Can Foundation Models Moderate Online Content? Evaluating Instruction- vs. Example-Driven Policy Operationalization
【速读】： 该论文旨在解决内容审核政策日益复杂化背景下，如何实现其一致且可靠执行的难题。当前，尽管基础模型具备处理此类任务的基本能力，但其在实际内容审核中的可靠性仍缺乏充分验证。为此，研究系统比较了视觉-语言模型（VLM）引导的两种竞争范式：基于指令的推理范式（模型从政策原则出发进行推理）与基于示例的泛化范式（模型从历史案例中学习并推广）。研究依托于ModerationBench这一新基准，包含来自Bluesky平台的4,000条人工标注的真实网络帖子。实验结果表明，基础模型在随机测试集上的F₁分数（0.60）显著优于Bluesky现有部署系统的0.22，近乎提升两倍；且两种范式均能达到相近的最优性能。因此，该研究的关键在于证明了通过合理设计的VLM引导策略，可实现大规模、可扩展且可靠的政策执行路径，为内容审核的自动化与标准化提供了可行方案。
链接: https://arxiv.org/abs/2609.10410
作者: Ayan Majumdar,Shounak Paul,Pushpdeep Singh,Ines Abdelaziz,Sayeh Jarollahi,Seungeon Lee,Krishna P. Gummadi,Ingmar Weber,Abhisek Dash
机构: MPI-SWS(德国马普所计算机科学研究所); Saarland University(萨尔大学); INRIA(法国国家信息与自动化研究院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 33 pages, 28 figures, 8 tables
Abstract:The growing complexity of content moderation policies presents a critical challenge for their consistent operationalization. While foundation models possess the basic capabilities needed to confront this challenge, whether they can reliably moderate online content remains an unanswered question. In this paper, we systematically compare two competing paradigms for Vision-Language Model (VLM) guidance: an instruction-driven approach where models reason from policy precepts, and an example-driven approach where they generalize from prior precedents. We ground this investigation in ModerationBench, a new benchmark of 4,000 manually annotated, in-the-wild posts from the Bluesky platform. Our experiments reveal that foundation models can substantially outperform Bluesky’s deployed moderation system, nearly tripling its F_1 score (0.60 vs. 0.22) on Random Posts in the benchmark, with both instruction- and example-driven paradigms achieving comparable peak effectiveness. Our findings thus chart a path toward reliable and adaptable policy operationalization at scale.
[NLP-6] Retrofitting Code Using LLM s to Support Exceptional Behavior
【速读】： 该论文旨在解决在大型代码库中手动编写异常处理代码（Exception Related Code, ERC）效率低下且易出错的问题。现有方法难以自动化补全缺失的ERC，尤其是在缺乏显式异常处理逻辑的情况下，如何根据异常行为测试（Exceptional Behavior Tests, EBTs）自动生成符合预期的抛出语句、保护条件及try/catch块。其解决方案的关键在于提出一种名为EXCODER的框架，通过融合静态与动态程序分析技术，提取上下文信息并进行“上下文工程”（context engineering），以增强大语言模型（LLM）对复杂异常处理逻辑的理解与生成能力。具体而言，EXCODER将程序结构、数据流、控制流以及运行时行为等多源上下文信息有效整合并注入到LLM推理过程，从而显著提升生成代码的正确性与可测性。实验结果表明，在基于GitHub Java项目构建的基准上，结合Qwen 2.5 Coder 32b模型，EXCODER在开发者编写的测试套件中实现了85.92%（pass@1）、86.18%（pass@5）和86.51%（pass@10）的通过率，相较基线分别提升了12.56、12.82和13.15个百分点，验证了该方法在测试驱动的异常处理代码生成任务中的有效性。
链接: https://arxiv.org/abs/2609.10397
作者: Linghan Zhong,Jiyang Zhang,Jayanth Srinivasa,Junyi Jessy Li,Milos Gligoric
机构: The University of Texas at Austin, USA; Cisco Systems, USA
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL)
备注: ISSRE 2026
Abstract:Exception Related Code (ERC), which includes throw statements, conditions (if statements) that guard those throw statements, and try/catch blocks, is an essential component of software systems, allowing developers to detect and handle exceptional states that deviate from the expected program behavior. However, manually writing ERC across large codebases is tedious. We propose a novel task: retrofitting existing code with ERC. Namely, given code (without ERC) and Exceptional Behavior Tests (EBTs) (e.g., check if method throws InvalidArgumentException if null is given as the value to the argument) we aim to automatically generate missing ERC, such that the given tests pass. We design and implement Exception Coder (EXCODER) that performs context engineering to help Large Language Models (LLMs) tackle this task. EXCODER integrates static and dynamic program analysis with LLMs by providing the extracted contextual information to the LLMs. To evaluate EXCODER, we build a benchmark constructed from GitHub Java repositories, where we systematically remove ERC in 304 methods from 75 projects. Our results demonstrate that EXCODER provides an effective, though imperfect, solution to this problem in automated code generation, offering developers the first way to implement ERC following test-driven development. When combined with Qwen 2.5 Coder 32b, EXCODER achieves pass@1, 5, and 10 rates of 85.92% (12.56 percentage points over baseline), 86.18% (12.82 p.p. over baseline), and 86.51% (13.15 p.p. over baseline), respectively, on developer-written test suites. Our manual inspection of the generated code further reveals limitations of EXCODER, pointing to directions for future work.
[NLP-7] Rosetta at AlexandriaX-2026: LoRA-Adapted NileChat for Context-Aware Dialectal Arabic Dialogue Translation EMNLP
【速读】： 该论文旨在解决上下文感知的英语到方言阿拉伯语对话翻译问题，即在保留对话上下文信息的前提下，将英语对话准确转换为对应的阿拉伯语方言。其核心挑战在于如何有效建模方言特异性与对话连贯性之间的关系。解决方案的关键在于基于NileChat-3B模型，采用结构化系统/用户提示（structured system/user prompts）对低秩适配器（LoRA adapter）进行微调，通过提示工程显式地引导生成过程以适应目标方言和上下文语境。对于无约束赛道，进一步在MADAR和PADIC数据集上进行预训练以增强泛化能力。实验结果表明，外部预训练仅对其中13种方言中的2种有提升作用，且整体性能略有下降，暗示存在负迁移现象，提示需谨慎设计多源数据融合策略。
链接: https://arxiv.org/abs/2609.10395
作者: Nada Esmaeil,Fathima Rena,Sibi Subhash,Osama Elgendy,Mina Naguib,Salma Omar,Muhammad Arif
机构: 未知
类目: Computation and Language (cs.CL)
备注: 5 pages, 3 tables, accepted to the AlexandriaX 2026 Shared Task at ArabicNLP 2026 (co-located with EMNLP)
Abstract:This paper describes the Rosetta system for Subtask 1 (Context-Aware English-to-Dialectal Arabic Dialogue Translation) of the AlexandriaX shared task, participating in both constrained and unconstrained tracks. The approach fine-tunes a LoRA adapter on NileChat-3B using structured system/user prompts that condition generation on dialect and dialogue context. For the unconstrained track, the adapter is additionally pretrained on MADAR and PADIC. Rosetta ranked 4th in the constrained track (spBLEU 26.10) and 5th in the unconstrained track (spBLEU 25.09). The experimental results demonstrate that external pretraining helps only two of thirteen dialects while slightly hurting overall performance, suggesting negative transfer.
[NLP-8] Why Is Video Still So Expensive? A Survey of Inference-Efficiency Mechanisms in Video and Audiovisual LLM s WWW
【速读】： 该论文旨在解决视频大语言模型（VideoLLM）在实际部署中面临的计算与内存开销过大的问题，尤其是在实时、移动端及资源受限场景下的应用瓶颈。其核心挑战在于，随着视频帧数和上下文长度的增加，模型的推理成本呈显著上升趋势，严重制约了高效部署。解决方案的关键在于提升推理效率，具体通过一系列优化机制实现：包括帧采样策略、多模态编码压缩、连接层级别的令牌剪枝，以及大语言模型（LLM）预填充与解码阶段的优化。这些方法分别作用于视频理解流水线的不同阶段，有效降低了参数量、每输入的浮点运算次数（FLOPs）、延迟、内存占用以及视觉与音频令牌数量。论文系统梳理了自2022年末以来主流VideoLLM中的高效推理技术，并整合了在统一主机模型与输入协议下报告的准确率-成本对比数据，区分了可比性结果与跨论文异构证据，揭示了当前音频-视觉联合效率研究的不足与标准化评估体系的缺失，同时维护了一个持续更新的开源资源库以支持后续研究。
链接: https://arxiv.org/abs/2609.10355
作者: Killian Steunou,Yannis Tevissen,Mounîm A. El Yacoubi
机构: SAMOVAR Laboratory, Télécom SudParis, Institut Polytechnique de Paris, Palaiseau, France (SAMOVAR 实验室，巴黎高等电信学院，巴黎综合理工学院，帕莱索，法国); Moments Lab, Paris, France (时刻实验室，巴黎，法国)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Multimedia (cs.MM)
备注: Supplementary material at this https URL
Abstract:Video understanding has rapidly evolved toward video large language models (VideoLLMs): systems that couple video representations with pretrained large language models and condition generation on a textual prompt. Their strong performance on captioning, question answering, retrieval and temporal grounding comes at a computation and memory cost that grows with frame count and context length, limiting deployment in real-time, mobile and resource-constrained settings. This survey covers inference-efficiency mechanisms for visual and audiovisual VideoLLMs that report concrete reductions in parameter count, FLOPs per input, latency, memory, or visual and audio token count. We analyze bottlenecks across frame sampling, modality encoding, connector-level token reduction, and LLM prefilling and decoding. We organize methods by the pipeline stage at which they act, covering VideoLLMs developed since late 2022 together with earlier frame-sampling and vision-encoder mechanisms that remain components of current pipelines. We assemble literature-reported accuracy–cost comparisons under shared host models and input protocols wherever available, distinguish them from heterogeneous cross-paper evidence, and identify gaps in audiovisual efficiency and standardized evaluation. We maintain a repository at this https URL.
[NLP-9] From Symbolic Perception to Logical Deduction: A Framework for Guiding Language Models in Geometric Reasoning
【速读】： 该论文旨在解决生成式人工智能在平面几何推理任务中面临的挑战，即如何有效融合视觉感知与数学逻辑推理，同时克服大型多模态模型（LMMs）计算开销大、决策过程不透明的问题。其核心解决方案在于构建一个基于纯大型语言模型（LLM）的可解释推理框架，关键创新点是引入几何视觉解析器（Geometric Vision Parser），将图形信息转化为符号表达形式，并结合符号求解器（Symbolic Solver）进行形式化演绎推理，从而显著降低幻觉现象并提升推理过程的可解释性。为实现严格评估，研究团队从2025年中国中考（Zhongkao）试题中精心构建了一个具有高难度和数据新颖性的基准测试集，以检验模型深层逻辑推导能力。实验结果表明，该方法在性能上可媲美Gemini 2.5 Pro，同时输出更清晰、符合人类思维习惯的解答过程。
链接: https://arxiv.org/abs/2609.10335
作者: Weichen Dai,Rafael Medeiros Cabral,Ziyi Shou,Yan Cao,Xin Shen,Dongcai Lu,Yi Zhou
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Plane geometry remains a significant challenge in AI, requiring the integration of visual perception and mathematical reasoning. While Large Multimodal Models (LMMs) naturally handle visuo-linguistic inputs, they are often computationally intensive and opaque. We demonstrate that a pure Large Language Model (LLM), when equipped with specialized modules, can rival state-of-the-art LMMs on complex geometry problems. Our framework integrates a Geometric Vision Parser, which translates diagrams into symbolic form, with a Symbolic Solver that performs formal deductions, thereby mitigating hallucinations and promoting interpretable reasoning. To enable rigorous evaluation, we curate a benchmark of challenging problems from the 2025 Chinese Zhongkao examinations, ensuring data novelty and testing deeper deductive skills. Experiments demonstrate that our approach achieves performance comparable to Gemini 2.5 Pro while delivering clearer, human-like solutions.
[NLP-10] On-Policy Distillation for Vision-Language Model Adaptation an Effective Paradigm on Low-Quality Multimodal Data
【速读】： 该论文旨在解决现有视觉-语言知识蒸馏方法在类别和领域分布发生偏移时，依赖固定教师模型预测作为统一训练目标所导致的目标不可靠问题。其核心挑战在于：传统方法将教师输出作为静态、全局一致的监督信号，无法根据样本自身可靠性及训练阶段动态调整，从而限制了模型在泛化场景下的性能表现。本文提出OnPoKD（on-policy knowledge distillation），首次将基于策略的动态知识蒸馏引入视觉-语言模型适应任务中，关键创新在于将蒸馏目标的构建视为一个可学习的策略决策过程。具体而言，OnPoKD通过一个轻量级控制器，融合教师模型、学生模型与零样本先验中的可靠性与分歧性线索，动态生成样本级自适应的蒸馏目标；该控制器通过有界策略动作，实时平衡教师监督、零样本引导与硬标签锚定，实现对不同样本可靠性和训练阶段的灵活响应。控制器利用验证集反馈进行优化，使目标构建更注重迁移能力而非仅拟合训练数据分布。由于控制器仅在训练阶段使用，不影响推理架构与测试开销，可无缝集成至现有蒸馏流程。大量实验表明，OnPoKD在基础到新类泛化与跨数据集迁移任务中均显著优于现有强基线，验证了其动态目标构建机制的有效性。
链接: https://arxiv.org/abs/2609.10321
作者: Hongyuan Zhang,Xianda Guo,Yanlun Peng,Qianlong Yang,Yubin Guo,Pinhan Fu,Mulin Chen,Xiaozhen Qiao,Ping Luo
机构: The University of Hong Kong (香港大学); Great Wall Motor (长城汽车); Wuhan University (武汉大学); China University of Petroleum (East China) (中国石油大学（华东）); University of Science and Technology of China (中国科学技术大学); Northwestern Polytechnical University (西北工业大学); University of Science and Technology of China (中国科学技术大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Knowledge distillation offers an efficient route to transfer a task-adapted vision-language teacher to a compact student. The training target in current vision-language distillation methods is typically constructed from the teacher prediction and applied uniformly to all training samples, making it unreliable under class and domain shifts. In this paper, we argue that distillation target construction should be treated as a dynamic training decision rather than a fixed recipe. To this end, we propose OnPoKD, an on-policy distillation framework for vision-language model adaptation. To the best of our knowledge, OnPoKD is the first framework that applies on-policy distillation to vision-language model adaptation by learning target construction as a policy decision. OnPoKD learns a lightweight controller that constructs sample-wise adaptive targets using reliability and disagreement cues from the teacher model, student model, and zero-shot prior. Instead of relying on a fixed teacher prediction, the controller dynamically balances teacher supervision, zero-shot prior guidance, and hard-label anchoring through bounded policy actions, allowing the distillation target to adapt to varying sample reliability and training stages. The policy controller is updated with validation feedback, encouraging target construction to optimize transferability rather than merely fitting the training distribution. Since the controller is only used during training, OnPoKD can be seamlessly integrated into existing vision-language distillation pipelines while preserving the original inference architecture and test-time cost. Extensive experiments on Base-to-novel generalization and Cross-dataset transfer benchmarks show that OnPoKD consistently improves over strong vision-language distillation baselines.
[NLP-11] RiLM: Parameter-Efficient Language Modeling via Geodesic Decoding
【速读】： 该论文旨在解决小型语言模型（参数量低于百万级）在边缘部署、领域适应及可复现研究中的效率瓶颈问题，核心挑战在于传统模型中输出矩阵 $ W_{\text{out}} \in \mathbb{R}^{d \times |V|} $ 占据了约三分之一的模型容量，造成资源浪费。其解决方案的关键是提出黎曼语言模型（Riemannian Language Models, RiLM），彻底摒弃显式的输出层：通过将上下文状态建模为黎曼流形上的轨迹，利用当前状态与词表嵌入之间的测地线距离平方来生成下一个词的概率分布，实现输入与输出共享同一嵌入映射，解码过程完全由几何结构决定。作者在平坦空间 $ \mathbb{R}^d $（Flat RiLM）和庞加莱球 $ \mathbb{H}^d $（HypRiLM）上进行了实例化，采用一个共享的多层感知机组合映射 $ \phi （约29万参数， d=128 $, $ |V|=2000 $）。在WikiText-2数据集上，HypRiLM达到54.2 ± 0.2的验证困惑度，显著优于Flat RiLM（87.6 ± 0.6）及所有基线模型（如LSTM、Transformer、SSM，困惑度113–147），领先最强基准约2倍；在Penn Treebank及10,000词表规模的应力测试中，测地线解码展现出跨语料库和大词汇量的泛化能力，且双曲曲率有助于选择性表示。此外，论文还揭示了朴素双曲递归导致的边界坍缩现象，并提出莫比乌斯稳定化机制以恢复训练稳定性。研究聚焦于小规模模型的可控对比实验，不追求全词汇量下的最先进性能。
链接: https://arxiv.org/abs/2609.10305
作者: Fang Li
机构: Oklahoma Christian University (奥克拉荷马基督教大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Language models under one million parameters matter for edge deployment, domain adaptation, and reproducible research, yet a two-layer LSTM or Transformer at embedding width d = 128 still spends roughly one third of its capacity on the output matrix W_out in R^(d x |V|). We propose Riemannian Language Models (RiLM), which remove that layer entirely: context unfolds as a trajectory on a Riemannian manifold, and next-token probabilities arise from squared geodesic distance between the current state and vocabulary embeddings. The same embedding map serves input and output – decoding is geometry. We instantiate the framework on flat R^d (Flat RiLM) and the Poincare ball H^d (HypRiLM) with a shared MLP composition map phi (~290k parameters, d = 128, |V| = 2000). Across five seeds on WikiText-2, HypRiLM reaches 54.2 +/- 0.2 validation perplexity versus 87.6 +/- 0.6 for Flat RiLM; tied and matched LSTM, Transformer, and SSM controls remain at 113-147 PPL on WT-2 – HypRiLM leads by roughly 2x over the strongest tied recurrent baseline (SSM, 113.0 +/- 3.8). Penn Treebank and a 10k-vocabulary stress test confirm that geodesic decoding transfers across corpora and larger |V|, while hyperbolic curvature helps selectively. We also characterize boundary collapse in naive hyperbolic recurrence and show how Mobius stabilization restores trainability. Claims are scoped to controlled small-model comparisons, not full-vocabulary state of the art.
[NLP-12] he Semantic Bottleneck: Leverag ing Semantic Representations for Non-Invasive Speech Decoding
【速读】： 该论文旨在解决非侵入式脑电信号（如脑磁图，MEG）在语音解码中因信噪比低而导致的细粒度音素或单词重建困难的问题。传统方法通常试图直接从神经信号中解码低层次的声学或词汇特征，但由于信号噪声大、时空分辨率受限，难以实现精确还原。针对这一挑战，研究提出了一种基于语义中间表征的新范式——Brain2Semantics2Text，其核心创新在于将句子级的神经响应映射至语义空间（semantic manifold），通过语义嵌入（semantic embedding）作为桥梁，实现从神经活动到自然语言的逆向生成。该方法的关键在于利用高阶语义表示在皮层区域广泛分布且随时间缓慢演变的神经科学证据，以语义内容为解码目标，规避了对词级别对齐的依赖，从而有效缓解了低信噪比带来的建模难题。通过构建可靠的神经-语义映射并采用相应的优化策略，该方法显著提升了非侵入式脑-文本解码在句子层面的表现，优于以往方法。
链接: https://arxiv.org/abs/2609.10296
作者: Gilad D. Landau,Dulhan Jayalath,Oiwi Parker Jones
机构: University of Oxford (牛津大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 12 pages, 8 figures
Abstract:Non-invasive speech decoding remains constrained by the low signal-to-noise ratio of neural recordings, which makes fine-grained reconstruction of phonemes or individual words difficult. Motivated by neuroscientific evidence that high-level semantic representations are distributed across cortical regions and evolve over slower temporal scales, we hypothesize that semantic content may provide a more suitable target for non-invasive decoding than low-level acoustic or lexical features. We introduce Brain2Semantics2Text, a method that reconstructs text through an intermediate semantic embedding space. Our model maps sentence-level MEG responses into a semantic manifold and then inverts the predicted embeddings into natural language. This semantic bottleneck enables recovery of high-level meaning without word-level alignment. We describe the core principles of the approach, its implementation, and the strategies used to mitigate the challenges of learning a reliable neural-to-semantic mapping. Finally, we compare against prior non-invasive Brain2Text methods and show improved sentence-level results.
[NLP-13] KVShareArena: KV-Cache Reuse Across Contexts and Model Checkpoints
【速读】： 该论文旨在解决大语言模型（LLM）服务系统中键值缓存（KV cache）在非前缀场景下的失效问题，即当需复用的文本不再位于提示（prompt）起始位置时，传统缓存机制因位置信息错误或模型检查点不一致导致无法正确使用。这一问题在检索增强生成（RAG）服务器和多智能体协调系统中尤为突出：前者为每个查询动态拼接不同的检索片段，后者则需读取其他智能体生成的报告。由于缓存中的注意力位置信息与当前上下文不匹配，且可能由不同模型检查点生成，直接复用会导致推理偏差甚至性能劣化。其核心解决方案在于对缓存进行修复，关键突破是提出一种基于位置校正（position correction）的方法——该方法无需重新编码即可恢复缓存的有效性，在仅需少量来源时已足够；而当需要同时利用多个来源时，则需通过部分重编码或训练微调等代价更高的方法才能恢复50%至70%的性能增益。研究构建了跨提示上下文与模型检查点的基准测试平台KVShareArena，全面评估各类修复方法在真实工作负载下的表现，并引入计算开销、内存占用及请求延迟的成本核算机制。实验表明，针对新生成的智能体报告，仅依赖压缩的缓存策略远逊于位置校正，且在不同检查点写入缓存的情况下，训练自由方法具有更强鲁棒性，而基于特定检查点训练的适配器则显著退化。相关工具链（Harness、冻结查询集、成本会计）已开源并支持自动化提交与公开排行榜。
链接: https://arxiv.org/abs/2609.10266
作者: Xi Shi,Qian Lou
机构: University of Central Florida (中佛罗里达大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:LLM serving systems already reuse KV caches, but only when the reused text sits at the very start of the prompt. Two growing workloads break this condition: a retrieval-augmented generation server assembles a different set of retrieved chunks for every query, and a multi-agent coordinator reads reports written by other agents. Reused inside a new prompt, a cache carries the wrong positions and never attended to the other sources. The cache may also have been written by a different checkpoint of the same model family, which changes the stored values. Repair methods for such caches have appeared in three separate communities, each measured on its own terms, and existing benchmarks test only exact-prefix reuse, where nothing is lost. KVShareArena benchmarks KV-cache reuse across prompt contexts and model checkpoints on retrieved chunks and agent reports. It scores every method by the fraction of the gap it recovers between no cache and full recomputation, and charges compute, memory, and per-request latency with the cache in hand, reporting the one-time cost of building a cache separately. We find that correcting positions, which needs no recomputation, is enough until a question needs several sources at once. There, only methods that pay, by re-encoding part of the cache or by training, recover half to two thirds of the gap; unrepaired caches can be worse than no cache. Cache-compression methods that are harmless on a single prompt fall significantly behind position correction on freshly written agent reports. These patterns hold across three model boards. When a different checkpoint wrote the cache, training-free methods are barely affected, while an adapter trained on one checkpoint’s caches loses quality. Harness, frozen querysets, and cost accounting ship as a pip package with an automated submission workflow and a public leaderboard.
[NLP-14] DiSCo: A Distribution-First Steering and Cultural Prior Evaluation Framework for Measuring Cultural Preference Bias in LLM s
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在跨文化情境下存在的系统性文化偏好问题，即模型在处理具有文化根基的日常任务时，倾向于优先选择特定文化（如英美文化）的响应，从而导致本地化适配不佳、用户信任度下降以及行为不公。现有文化评估基准通常依赖单一“正确”答案来衡量准确性，难以捕捉多文化合理响应下的模型默认偏好，并且混淆了模型的固有文化倾向与上下文驱动的适应能力。为此，论文提出DiSCo（Distribution-first, forced-choice）评估框架，通过四层级上下文梯度（C0–C3）分离模型的默认文化先验并测试其可调控性。基于涵盖12个文化的DiSCo-Bench（304个样本），对六种指令微调的LLM进行评估发现，模型默认偏好高度集中，英美文化合计占据约35%的选择比例，远超其文化代表性；更重要的是，基于提示词的引导策略反而加剧了高资源文化与低资源文化之间的选择差距，且显式注入文化事实对分布的扰动微乎其微，表明仅靠提示工程无法有效缓解文化偏好偏差。因此，解决方案的关键在于建立以分布为核心、可分离默认偏见与上下文适应性的评估范式，并揭示当前提示调优方法在消除文化偏见方面的根本局限。
链接: https://arxiv.org/abs/2609.10253
作者: Bhuvan Arora,Devesh Saraogi,Sravya Varada,Dhruv Kumar
机构: BITS Pilani(毕尔拉理工学院), India
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) are increasingly deployed in globally used assistants, yet their default choices in culturally grounded everyday situations can systematically favour some cultures over others, affecting localisation, user trust, and equitable behaviour. Existing cultural benchmarks evaluate accuracy against a single “correct” answer, making it difficult to characterise an LLM’s cultural preference prior when multiple culturally grounded responses are all valid; they also conflate default preferences with context-driven adaptation. We propose DiSCo, a distribution-first forced-choice evaluation framework that isolates default cultural priors and tests steerability via a four-level context gradient (C0–C3). Using DiSCo-Bench (304 items) derived from BLEnD spanning 12 cultures, we evaluate six diverse instruction-tuned LLMs. Default priors are heavily concentrated, with UK and US together absorbing approximately 35% of all selections despite representing only 2 of 12 cultures. Most critically, prompt-based steering consistently widens the selection gap between high- and low-resource cultures, and injecting explicit cultural facts produces negligible distributional disruption, confirming that cultural preference bias cannot be resolved through prompt-based personalisation alone.
[NLP-15] wo-Token Features and Small-Large Ensembles for VLM Hallucination Detection
【速读】： 该论文旨在解决字符级视觉语言模型（VLM）幻觉检测中的难题，即识别图像描述中由模型生成的不真实或错误文本内容。其核心挑战在于如何在缺乏大量标注数据的情况下，有效区分真实可见文本与模型幻觉生成的虚假文本。解决方案的关键在于构建一个双组件集成框架：首先微调一个小型（约40亿参数）的视觉语言模型作为逐标记分类器，通过分析自身隐藏状态中两个连续标记的特征表示来判断是否存在幻觉；其次，在推理阶段将该小模型与一个约4000亿参数的零样本（zero-shot）大型视觉语言模型判别器进行集成，以增强泛化能力。为提升集成多样性，研究利用大模型生成的合成幻觉数据作为训练来源，并通过验证集选择最优的特征层、训练数据及光学字符识别（OCR）对齐策略。该方法在多语言测试集上取得了显著性能，尤其在主要评估指标Cor-lbl上表现优异，证明了小模型与大模型协同集成在低资源幻觉检测场景下的有效性。
链接: https://arxiv.org/abs/2609.10244
作者: Eli Schwartz
机构: IBM Research(IBM研究院)
类目: Computation and Language (cs.CL)
备注:
Abstract:We present our system for the SHROOM-Visions 2026 shared task on character-level VLM hallucination detection. A small ( 4 B-parameter) VLM is fine-tuned as a per-token classifier reading a two-token feature from its own hidden states, and is ensembled with a \sim 400B zero-shot VLM judge at prediction time. Both components see off-the-shelf OCR of any visible in-image text. We use synthetic hallucination data generated by the large model as a source of ensemble diversity, and use validation to select feature layer, training data and OCR grounding. Our official entry reaches mean Cor 0.487 / Cor-lbl 0.387 on the hidden test set, placing 6 th/ 28 (EN), 6 th/ 21 (FR), 8 th/ 21 (IT) and 7 th/ 22 (ZH) on the task’s primary Cor-lbl metric.
[NLP-16] Φ-Bench: Can Large Language Models Engineer the Infrastructure That Powers Them?
【速读】： 该论文旨在解决现有基准测试无法有效评估大语言模型（LLM）在开放性、长周期的大型语言模型基础设施工程任务中能力的问题。当前主流基准多聚焦于孤立的内核优化、预定义算子或固定优化目标，难以反映模型在真实复杂系统中进行全栈式、端到端工程设计与优化的能力。为此，论文提出Φ-Bench，一个系统性评估大语言模型在构建和优化大语言模型基础设施栈方面能力的基准。其关键在于：基于前沿研究中的实际优化问题，并结合真实世界代码仓库中的工程实践，覆盖从底层内核函数补全到长期规划的系统实现与整体性能调优等多层次、多粒度的任务，从而全面衡量大语言模型在复杂基础设施工程中的综合能力。该基准为揭示当前大语言模型在自主优化未来人工智能基础设施过程中所面临的挑战提供了重要依据。
链接: https://arxiv.org/abs/2609.10226
作者: Leilei Ding,Shumin Wang,Yuting Huang,Fanqi Wan,Yinmin Zhang,Qi Han,Yiming Xu,Feiyuan Zhang,Xiaomeng Chu,Guoliang You,Wuyang Zhang,Daxin Jiang,Yanyong Zhang
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) have demonstrated remarkable capabilities in reasoning and code generation, raising the prospect that they could assist in developing and optimizing the very infrastructure that powers them. However, existing benchmarks mainly focus on isolated kernels, predefined operators, or pre-specified optimization targets, and therefore fail to evaluate the ability of LLMs to perform open-ended, long-horizon LLM infrastructure engineering. To address this gap, we present \Phi -Bench, a benchmark for systematically evaluating LLMs on engineering the LLM infrastructure stack. Derived from optimization problems studied in frontier research and grounded in real-world code repositories, \Phi -Bench provides broad coverage of the LLM infrastructure stack and spans tasks of varying complexity, ranging from localized kernel-level function completion to long-horizon implementation and end-to-end system optimization. Extensive experiments on frontier LLMs reveal their current capabilities and limitations in engineering complex LLM infrastructure, offering insights into the challenges that remain on the path toward autonomous optimization of future AI infrastructure.
[NLP-17] hrough the Looking Glass: Directly Reading and Writing Transformers
【速读】： 该论文旨在解决生成式模型中“单个标记（token）的预测究竟由多少组件决定”这一核心问题，揭示了Transformer架构内部决策过程的可解释性与稀疏性本质。其关键在于通过计算每个单元（unit）和通道（channel）对最终logit输出的贡献绝对值，量化组件在预测中的实际作用；进一步发现尽管有数千至数十万组件参与，但其贡献呈显著符号差异——大量成分实际上在抑制目标标记，净贡献仅由少数关键组件驱动。研究指出，在基准模型中仅有53个组件承载90%的预测信息，其中13个为不可替代的核心组件，8个即可独立完成预测；跨十二个不同规模（124M至7B参数）的模型验证表明，支撑预测的“充分集”仅占模型总参数量的1%至3%，且不随模型规模增长而扩大。更关键的是，这些有效组件可被精确识别：其输入来自模型自身权重与激活状态，无需额外训练或拟合，且能明确标识出“写入”（从预测反推）与“读取”（基于本层权重）的双重角色，其中最强输入的预测性能比随机基线高出58.9%。此外，通过上游源排序可解析出语法类别等语义结构，甚至发现一个未被模型显式持有的关联可通过一个空闲单元实现，仅需极小代价（约0.25%的保留损失）。最终，研究揭示了一个具有顺序保持特性的激活机制，使组件输入达到“仪器上限”，并支持两阶段安装，从而实现了对模型内部动态的高精度追踪与干预。
链接: https://arxiv.org/abs/2609.10210
作者: Mark Oskin
机构: University of Washington (华盛顿大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:How many of a transformer’s components decide a token? Counted by the absolute value of each unit’s and channel’s contribution to the logit, one prediction rests on thousands to hundreds of thousands of them. But contributions are signed, and across eighteen models the mass pushing away from the predicted token is a median of seven times the mass carrying it. Divide by the net and the count is dozens: on the baseline, 53 components carry ninety percent of a prediction, 13 it cannot survive losing, and 8 suffice to produce it alone. Across twelve models trained elsewhere, 124M to 7B parameters, the sufficient set runs from two components to sixteen, and what a prediction draws on, followed all the way back, is one to three percent of the model, a share that does not grow with size. Three quarters of a layer’s update is a fixed linear map of the state it received. Everything is read from the model’s own parameters and activations, with nothing trained or fitted, and it names a component on both sides: what it writes, from the predictions it drives, reaching close to half of every model; what it reads, from its weights in the frame of its own layer, at 58.9 percent above chance over its eight strongest inputs. Sorting the remainder by upstream source yields grammatical categories the embedding cannot see. A name can be acted on. An association the model does not hold installs into one spare unit, key and value read from the weights, for a quarter of a percent of held-out loss, a fortieth of what a rank-one update costs. An installed attention head and a unit two layers above it make an edit fire only where a token occurred earlier in the context, and a unit the model trained for itself is driven from two layers upstream, 86 percent of the effect passing through it. An order-preserving activation puts a unit’s inputs at the instrument’s ceiling, at the price of a two-part install. Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG) Cite as: arXiv:2609.10210 [cs.CL] (or arXiv:2609.10210v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.10210 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-18] Politics of Feelings: Emotional Expression and Legislative Effectiveness in the U.S. Congress
【速读】： 该论文旨在解决现有政治传播研究中对情绪表达与实质性立法成果之间关联性探讨不足的问题。传统研究多聚焦于描述情绪表达的模式，却未深入分析其是否影响立法成效。为此，本文基于1973至2024年间美国国会超过170万次演讲，利用基于Transformer架构的情绪分类器，量化八种离散情绪（愤怒、恐惧、厌恶、悲伤、喜悦、热情、自豪与希望），系统考察情绪表达随时间、政策议题、议员特征的变化规律及其与立法有效性的关系。其解决方案的关键在于：通过计算方法精确识别并量化离散情绪，揭示情绪维度在立法沟通中的作用机制；研究发现，情绪表达整体呈上升趋势，且在不同政策领域和意识形态立场间呈现系统性差异；更重要的是，特定情绪类型对立法有效性具有差异化影响——热情与自豪正向促进立法成效，而愤怒则产生负向效应；此外，情绪效价（valence）与情绪多样性正相关于立法有效性，而情绪强度则呈负相关。该研究证明，基于计算模型的离散情绪测量可有效揭示立法话语中的情感维度，深化对立法者在民主制度中沟通、互动与绩效的理解。
链接: https://arxiv.org/abs/2609.10198
作者: Segun Aroyehun
机构: University of Konstanz(康斯坦茨大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Emotions are a pervasive feature of political communication, yet existing research has focused primarily on describing patterns of emotional expression rather than examining whether they are associated with consequential legislative outcomes. We address this gap by investigating the expression and correlates of discrete emotions in more than 1.7 million speeches delivered in the U.S. Congress between 1973 and 2024. Using a transformer-based emotion classifier, we measure eight discrete emotions: anger, fear, disgust, sadness, joy, enthusiasm, pride, and hope. We examine how these emotions vary over time, across policy topics, legislator characteristics, and their relationship with legislative effectiveness. We find that congressional speeches are becoming emotionally expressive over time. Emotional expression also varies systematically across policy domains and ideological positioning of legislators. Notably, the relationship between emotional expression and legislative effectiveness depends on the specific emotions expressed: enthusiasm and pride are positively associated with effectiveness, whereas anger exhibits a negative association. Emotional valence and emotional diversity are positively associated with legislative effectiveness, while emotional intensity is negatively associated with legislative effectiveness. These findings demonstrate that computationally derived measures of discrete emotions can provide insight into affective dimensions of legislative speeches and facilitate our understanding of how legislators communicate, interact, and perform within democratic institutions.
[NLP-19] Who Argues What? Joint Argument-Entity Detection and Classification in Political Debates EMNLP2026
【速读】： 该论文旨在解决政治辩论中论点可解释性不足的问题，即现有论点挖掘（Argument Mining, AM）方法仅关注论点片段及其角色标注，而忽略了支撑论点的关键实体（如人物、事件、地点、政党等）信息，导致论点难以被充分理解。其核心挑战在于缺乏对辩论中涉及的辩论命名实体（Debate Named Entities, DNE）的系统标注与联合建模机制。为此，本文提出两项关键创新：一是构建DNE-ElecDeb数据集，作为USElecDeb的实体增强版本，在论点性与非论点性文本跨度中均标注了DNE，从而定义了辩论命名实体识别（DNER）任务；二是提出联合论点与实体标记（Joint Argument and Entity Tagging, JAET）框架，基于解码器-only的生成式大语言模型（Generative AI），在不修改原始对话文本的前提下，通过插入内联的论点与实体标签，实现论点与实体信息的联合建模。实验表明，在BIO标注评估下，JAET在联合AM+DNER任务上相对于最强的串行管道分别提升了27.3%（无类型）和41.9%（有类型）的相对F1值，且这些性能增益无法通过独立模块组合恢复，证明了联合建模的有效性。此外，该方法在说服性作文任务上也展现出显著泛化能力（提升26.6%和52.7%），验证了其跨领域适用性。通过将论点与实体层面的表征统一于同一视图，本工作为更深入的政治辩论理解提供了新范式。
链接: https://arxiv.org/abs/2609.10192
作者: Lucio La Cava,Stefano Francesco Monea,Sergio Greco
机构: University of Calabria (卡莱布里亚大学); DIMES Dept. (数字信息与系统工程系)
类目: Computation and Language (cs.CL)
备注: Accepted at EMNLP 2026 - Findings
Abstract:Political debates are often analyzed through Argument Mining (AM) to investigate the key arguments that drive them. However, political arguments are rarely interpretable from argumentative spans alone, as claims and premises generally depend on the entities (e.g., people, events, locations, parties) they mention. Existing AM resources and methods typically annotate argumentative spans and roles, but do not provide a paired debate-entity layer for asking which Debate Named Entities (DNE), e.g., actors and events, are invoked within debates. In this work, we address these data and methodological gaps by (i) introducing DNE-ElecDeb, an entity-enriched version of the USElecDeb dataset that adds DNEs in both argumentative and non-argumentative spans and defines Debate Named Entity Recognition (DNER) as the task of detecting DNEs, and (ii) proposing Joint Argument and Entity Tagging (JAET), a generative framework that fine-tunes decoder-only LLMs to insert inline argument and entity tags into debate turns while preserving the original transcript. Under BIO-tagging evaluation, JAET improves relative F1 on the joint AM+DNER task by +27.3%, resp. +41.9%, under the untyped, resp. typed setting over the strongest sequential AM-DNER pipelines, demonstrating that such gains cannot be recovered by composing two independent modules. Notably, similar margins replicate on Persuasive Essays (+26.6%, resp. +52.7%), showing effective generalization to domains orthogonal to political debates. By unifying argumentative and entity-level representations within a single view, our contributions pave the way for richer political debates understanding.
[NLP-20] YallaMorph: A Benchmark for Evaluating Arabic Morphological Generation in Large Language Models EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在阿拉伯语形态生成中缺乏精确的形态句法控制问题，即模型虽能生成流畅文本，但难以根据显式的词干与特征输入实现精准的形态生成。现有评估多聚焦下游任务，未能直接检验从明确词汇及特征输入出发的受控形态生成能力。为此，研究提出YallaMorph，一个大规模阿拉伯语形态生成基准，涵盖动词、名词、形容词及其附着代词形式（cliticized forms）以及无效构型，全面覆盖阿拉伯语形态多样性。通过在60万条基准样本上对多语言及专为阿拉伯语优化的LLMs进行带符号（diacritized）与无符号（undiacritized）设置下的评估，结果表明：阿拉伯语形态生成仍具挑战性，尤其在附着形式、未见形态及罕见形态方面表现显著不足。其解决方案的关键在于构建系统化、精细化的基准测试框架，以直接量化模型在受控条件下的形态生成能力，从而推动对阿拉伯语形态学建模的深入研究。
链接: https://arxiv.org/abs/2609.10153
作者: Mahmoud Reda,Salam Khalifa,Reham Marzouk,Nizar Habash
机构: Computational Approaches to Modeling Language (CAMeL) Lab; New York University Abu Dhabi(纽约大学阿布扎比分校); Stony Brook University(石溪大学); Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026
Abstract:Arabic morphology remains challenging for large language models, since fluent generation does not guarantee accurate morphosyntactic control. Existing Arabic evaluations mainly target downstream tasks and do not directly test controlled morphological generation from explicit lexical and feature-based input. We introduce YallaMorph, a large-scale benchmark for Arabic morphological generation covering verbs, nouns, adjectives, their cliticized forms, and invalid configurations. We evaluate multilingual and Arabic-oriented LLMs under diacritized and undiacritized settings over 600K benchmark entries. Results show that Arabic morphological generation remains difficult, especially for cliticized, unseen, and morphologically rare forms.
[NLP-21] Active Adaptation Not Static Defense: Temporal Dynamics of Preventative Steering in Adversarial Fine-Tuning EMNLP2026
【速读】： 该论文旨在解决大语言模型在微调过程中遭受恶意攻击导致有害人格漂移（harmful persona drift）的问题，尤其关注训练阶段的防御机制。现有方法“预防性引导”（Preventative Steering）通过在微调期间注入不良特质的人格向量并在评估时移除以实现保护，但其长期有效性的内在机制尚不明确。研究通过分析优化过程中的时间动态，发现防御效果源于早期的补偿适应阶段与后续稳定状态阶段，其中校正信号逐渐衰减；在参数空间中，注意力输出投影被识别为防御更新的主要残差写入路径。通过干预增量保留（IDP）及延续实验进一步表明，仅保留或重新注入权重偏移无法维持防护效果，说明该方法依赖于持续的主动适应而非静态防御机制。基于此发现，本文提出渐进强度调度（Progressive Intensity Scheduling, PIS），即初始采用适中强度的注入，并在静态强度对齐开始衰减后逐步增强注入强度。在Qwen2.5与Gemma-3模型上的实验验证表明，PIS在提升安全鲁棒性的同时显著降低了有害特质的表达，优于传统静态强度引导策略。
链接: https://arxiv.org/abs/2609.10142
作者: Jing Guan,Yachao Yang,Zhaoliang Liu,Yuyao Zhang,Fanyu Meng,Junlan Feng
机构: JIUTIAN Research(九天研究); Beijing University of Posts and Telecommunications(北京邮电大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
备注: Accepted to Findings of EMNLP 2026
Abstract:Large language models remain fragile against malicious fine-tuning, motivating training-time defenses against harmful persona drift. Preventative Steering injects undesirable-trait persona vectors during fine-tuning and removes them at evaluation time, yet the mechanism behind its lasting protection remains unclear. Analyzing its temporal optimization dynamics, we find that the defense emerges from an early compensatory adaptation phase followed by a steady-state phase where the corrective signal decays; in parameter space, attention output projections emerge as the dominant residual-write route for defensive updates. Through Intervention Delta Preservation (IDP) and IDP Continuation experiments, we further show that preserving or reinjecting the weight offset fails to maintain protection, indicating that preventative steering relies on active adaptation rather than a static defense. Motivated by this finding, we propose Progressive Intensity Scheduling (PIS), which starts with a moderate injection strength and increases it after static-strength alignment begins to decay. Across the evaluated Qwen2.5 and Gemma-3 models, PIS improves safety robustness over static-strength steering while reducing harmful trait expression.
[NLP-22] If Its Not Buggy Dont Fix It: On the Dynamics of Iterative Bug-fixing with LLM s
【速读】： 该论文旨在解决生成式 AI 在软件开发中作为自动化程序修复工具时所面临的“盲迭代使用”问题，即在代码审查过程中反复依赖大语言模型（Large Language Models, LLMs）进行缺陷修复，却未充分考虑其误判与自我强化错误的潜在风险。研究发现，LLMs 在无缺陷程序中持续误报缺陷，且对存在缺陷程序的实际修复率低于对正确代码造成的破坏率；进一步分析揭示，这种迭代过程常陷入一种“伪修复循环”——相同修改被反复添加与删除，无法收敛。通过机制探针分析，研究识别出一个控制编辑倾向的“引导向量”（steering vector），表明大语言模型内部存在对“有缺陷代码”的表征，并可能因错误激活该表征而导致虚假修复行为。因此，该研究的关键解决方案在于揭示了生成式 AI 在自主修复场景下的内在动态机制，强调需建立基于明确目标的终止条件，以避免在模糊目标下产生不可控的伪修复循环。
链接: https://arxiv.org/abs/2609.10123
作者: Xietao Wang-Lin,Anton Isopoussu,Louis Mahon
机构: University of Warwick(华威大学); UnlikelyAI
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) have become ubiquitous in software development, with LLM-based automated program repair tools increasingly used during code review. In this report, we explore the iterative blind use of LLMs as bug-fixers. Across multiple models and repair environments, we find that LLMs consistently claim to detect bugs in entirely bug-free programs while the rate of repair of buggy programs is less than that of the damage to correct programs. We also explore the long-term dynamics of this iterative process, and find that this frequently reaches a pseudo-bug-fixing cycle where the same changes are added and removed again ad infinitum. Lastly, via mechanistic probing, we unveil the existence of a steering vector which controls the editing propensity, suggesting that LLMs have an internal representation of ``buggy code", and that this representation is what is falsely activated to induce pseudo-bug fixing. These results provide insight towards the dynamics of fully autonomous bug-fixing systems, as well as stopping conditions under ambiguous goals.
[NLP-23] ProbPlug: A Plugin Uncertainty Network for Reliable Confidence in LLM Binary Classification PRICAI2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在高风险场景下部署时面临的预测可靠性问题，尤其关注基于LLM的分类任务中置信度估计（confidence estimation）与置信度校准（confidence calibration）不足的挑战。现有研究虽已广泛探讨LLM的置信度估计，但针对其分类任务的系统性校准方法仍相对匮乏。为此，论文提出ProbPlug——一种轻量级的置信度估计框架，专用于基于LLM的二分类任务。其核心创新在于利用冻结的LLM内部的标记级特征（token-level features）作为输入，通过自注意力（self-attention）模块对隐藏表示进行聚合，进而预测输出结果的正确性。该方法无需修改基础模型，可无缝集成至原有推理流程中，具有极低的额外计算开销。实验表明，ProbPlug在多种文本及多模态大模型任务上均能提供更可靠的置信度估计，显著提升分类性能，并展现出优异的任务泛化能力。因此，ProbPlug为实现高效、可靠且通用的LLM分类置信度估计提供了切实可行的解决方案。
链接: https://arxiv.org/abs/2609.10122
作者: Jianzong Wang,Chuhang Liu,Botao Zhao,Zuheng Kang,Xulong Zhang,Xiaoyang Qu,Junqing Peng,Zhiewei Ye,Yayun He
机构: Ping An Technology (Shenzhen) Co., Ltd., Shenzhen, China; Tsinghua Shenzhen International Graduate School, Shenzhen, China; Hubei University of Technology, Hubei, China
类目: Computation and Language (cs.CL)
备注: Accepted by the 23rd Pacific Rim International Conference on Artificial Intelligence. (PRICAI 2026)
Abstract:Large language models (LLMs) have achieved strong performance across a broad range of classification settings, yet the reliability of their predictions remains a major obstacle to deployment in high-stakes scenarios. Although confidence estimation for LLMs has been widely studied, confidence calibration for LLM-based classification remains underexplored. We introduce ProbPlug, a lightweight confidence estimation framework for LLM-based binary classification, which predicts whether an output is correct using internal token features extracted from a frozen LLM. ProbPlug employs a self-attention module to aggregate hidden representations and can be integrated into the original inference pipeline without modifying the base model. Experiments across multiple tasks involving both text-based and multimodal large models show that ProbPlug provides more reliable confidence estimates, improves classification performance with negligible additional overhead, and exhibits strong generalization across tasks. These results indicate that ProbPlug serves as a practical solution for confidence estimation in LLM-based classification. Our code is publicly available at Github.
[NLP-24] Data-Centric Post-Training for Financial Reasoning : Mining Distillation and Verifiable Learning
【速读】： 该论文旨在解决金融领域大语言模型在推理能力后训练（post-training）过程中因可用高质量、可验证的推理数据稀缺而导致性能下降的问题。现有问答对普遍存在推理过程不明确、上下文信息不足或答案不可靠等问题，而金融教材虽内容丰富，但需经过复杂转换才能用于训练。其解决方案的关键在于构建一个以数据为中心的全流程管道：通过挖掘开源推理轨迹、提炼金融指令数据，并基于知识图谱引导从金融教育材料中生成结构化问答对；随后通过语义去重及三类轻量级序列分类器筛选出与金融相关的有效样本，排除表述模糊的问题，并识别适合采用强化学习与规则验证器结合的任务。在模型适配方面，采用自蒸馏微调（self-distilled SFT）和模型合并等策略，以保留初始模型已有的金融能力，避免普通监督微调导致的性能退化。实验结果表明，自蒸馏SFT相较于初始模型提升1.0–2.8个百分点，等权重合并恢复3.0个百分点，且在硬任务上结合强化学习（GRPO）进一步提升0.4至3.0个百分点，充分证明了“保留感知型”适应方法在提升金融推理能力的同时，能有效防止性能退步。
链接: https://arxiv.org/abs/2609.10113
作者: Zhirayr Hayrapetyan,Andrei Kalmykov,Denis Kokosinskii,Dmitry Stanishevskii,Dmitry Zmitrovich
机构: 未知
类目: Computation and Language (cs.CL)
备注: 12 pages, 1 figure, 3 tables. Technical report
Abstract:Financial text, textbooks, and question-answer pairs are abundant, but only a small fraction is directly usable for reasoning-focused post-training. Existing QA pairs often lack explicit reasoning, sufficient context, or reliably verifiable answers, while textbooks must first be transformed into synthetic training examples. We present a data-centric pipeline that constructs complementary corpora by mining open-source reasoning traces, distilling financial instruction data, and generating knowledge-graph-guided question-answer pairs from financial educational material. After semantic deduplication, three lightweight sequence classifiers select finance-relevant examples, reject under-specified questions, and identify tasks suitable for reinforcement learning with compact rule-based verifiers. For model adaptation, we study supervised fine-tuning and reinforcement learning, while self-distilled fine-tuning and post-training model merging are used to prevent the loss of financial capabilities already present in the starting model. We evaluate the adapted language models using FINESSE-Bench, reporting aggregate performance and changes relative to their starting checkpoints. Across the selected comparisons, ordinary SFT reduces FINESSE-Bench accuracy by 3.2-4.0 percentage points, whereas self-distilled SFT improves over the corresponding starting models by 1.0-2.8 points. Equal-weight merging recovers 3.0 points over its SFT parent and finishes 0.9 points above the original model; GRPO on hard tasks adds 0.4 points after self-distilled SFT or 3.0 points when applied directly to verifiable tasks. These results show that retention-aware adaptation can improve financial reasoning without the regressions observed after ordinary SFT.
[NLP-25] RAP: Research Attention Prediction Reveals Target-Conditioned Evidence Acquisition Biases
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在作为科研代理时，难以有效追踪研究关注点动态变化的问题。由于学术综述与研究思路本身缺乏可唯一验证的客观结果，传统评估方法难以衡量模型对研究热点演进的预测能力。为此，作者提出“研究关注点预测”（Research Attention Prediction, RAP）这一滚动基准，覆盖278个AI/ML领域及1,390个时间片段，要求LLM在限定时间范围内的arXiv语料库中检索并预测未来六个月在八个固定研究方向上的论文占比。研究发现，尽管搜索机制普遍有助于提升性能，但四种诊断模型在组合准确性上均逊于一个基于精确计数的指数加权移动平均（EWMA）基线。关键瓶颈在于：在累积历史信息条件下，状态传递（State carry-forward）优于直接预测（Direct Forecast），而这一性能反转可归因于以预测为导向的策略在获取近期证据方面存在系统性偏差；即使拥有完整历史数据，面向未来的更新仍受限，仅有GPT-5.5结合开放搜索的微调版本略超EWMA基线。此外，基于实际结果进行微调可显著提升Qwen3-4B在后期起始场域的预测相关性（Spearman相关系数提升0.105），尤其在变化剧烈的时段表现更优。因此，解决方案的关键在于通过引入基于真实观测结果的反馈机制和改进证据获取策略，以增强模型对未来研究趋势的动态建模能力。
链接: https://arxiv.org/abs/2609.10092
作者: Yingqian Wu,Jingcong Liang,Siyuan Wang,Zhenfei Yin,Philip Torr,Junchi Yu,Zhongyu Wei
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) increasingly act as research agents, yet their ability to track shifts in research attention is difficult to evaluate because reviews and research ideas lack uniquely verifiable outcomes. We introduce Research Attention Prediction (RAP), a rolling benchmark covering 278 AI/ML fields and 1,390 episodes. At each cut-off, an LLM agent searches a temporally restricted arXiv corpus and predicts the next six months’ paper shares across eight frozen research directions. Search generally helps, but all four diagnostic models perform worse than an exact-count exponentially weighted moving average (EWMA) baseline in compositional accuracy. We identify two linked bottlenecks. Under cumulative-history access, State carry-forward outperforms direct Forecast for all four diagnostic models; frozen-evidence replay links a shared component of this reversal to Forecast-oriented policies retrieving a smaller share of recent evidence. Even with exact historical activity, future-specific updating remains limited, with only GPT-5.5 plus reopened Search slightly surpassing EWMA. Fine-tuning on realised outcomes improves Qwen3-4B’s forecast Spearman correlation by 0.105 on held-out fields at later origins, with gains also on change-rich episodes.
[NLP-26] NOPE-HYPE: A Structured Simulation Workflow for Robust Speech-to-Text Across Diverse Acoustic Environments
【速读】： 该论文旨在解决语音到文本翻译系统在多样化声学环境下的鲁棒性问题，现有实际部署流程缺乏可控工具以系统性地探索不同环境条件对模型性能的影响。当前大规模语音模型对未见声学条件仍表现出敏感性，主要由于训练数据无法覆盖真实场景中全部的声学变化。为此，本文提出NOPEHYPE——一种结构化的训练工作流，其核心在于结合可控制的环境模拟器、基于功率谱密度（Power Spectral Density, PSD）模板的覆盖最优环境降维方法，以及对模拟器参数旋钮的小规模可解释性超参数搜索。该方案的关键创新在于通过模拟生成噪声实现与均衡真实噪声训练相当的性能表现，在Whisper和SeamlessM4T模型上验证了其有效性，并构建了具有原则性的环境原型集，同时基于27次系统的超参数实验，识别出适用于多数场景的实用默认模拟器配置。
链接: https://arxiv.org/abs/2609.10058
作者: Niramay M. Patel,Bibek Behera,Raksha Sharma
机构: IISER Bhopal (印度科学与工程研究所博帕尔分校); IIT Bombay (印度理工学院孟买分校); IIT Roorkee (印度理工学院鲁尔基分校)
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Robust speech-to-text translation systems should perform reliably across diverse acoustic conditions, yet practical pipelines lack controllable tools for systematic environment exploration. Large speech models remain sensitive to unseen acoustic conditions, as training data rarely cover the full range of real this http URL present NOPEHYPE, a structured training workflow that combines a controllable environment simulator, coverage-optimal environment reduction on Power Spectral Density (PSD) templates, and a small, interpretable hyperparameter search over simulator knobs. We show that simulator-generated noise achieves performance comparable to balanced realnoise training across Whisper and SeamlessM4T models, provide principled environment prototype sets, and identify practical default simulator configurations from a structured 27-run hyperparameter sweep.
[NLP-27] OntologyAligner: Ontology-Aligned Retrieval and Hierarchy-Guided Large Language Model Reranking for Biomedical Ontology Normalization
【速读】： 该论文旨在解决生物医学本体（Biomedical Ontology）归一化中因词汇变异及层级概念间细微差异导致的概念边界模糊问题，从而实现自由文本表达到标准化概念的准确映射。其解决方案的关键在于提出OntologyAligner这一三阶段框架：首先通过本体对齐检索生成候选概念集，其次利用大语言模型（Large Language Model, LLM）进行候选重排序以提升语义匹配精度，最后采用选择性层次引导精炼机制进一步优化结果，有效融合了语义理解与结构约束。该方法在人类表型本体（Human Phenotype Ontology, HPO）归一化任务上达到88.78%的宏平均Top-1准确率和86.75%的微平均Top-1准确率，显著优于现有基线，并通过构建包含13,390个样本的PhenoNormBench统一基准验证了其有效性与可迁移性，展现出良好的泛化能力与鲁棒性。
链接: https://arxiv.org/abs/2609.10055
作者: Jie Song,Zhichuan Xu,Ziyu Lu,Meng Xiao,Cheng Bi,Yuxin Zhang,Xin Zheng,Xiaoran Li,Qiongfang Cao,Hao Yang,Bairong Shen
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 4 figures
Abstract:Biomedical ontology normalization maps free-text expressions to standardized concepts, enabling consistent integration and analysis of biomedical data. This task remains challenging because lexical variation and subtle distinctions among hierarchically related concepts can obscure concept boundaries. We present OntologyAligner, a three-stage framework that combines ontology-aligned retrieval, large language model candidate reranking, and selective hierarchy-guided refinement. We also construct PhenoNormBench, a unified benchmark comprising 13,390 samples from seven Human Phenotype Ontology datasets. OntologyAligner achieved state-of-the-art performance on HPO normalization, with 88.78% Macro Top-1 Accuracy and 86.75% Micro Top-1 Accuracy, exceeding the strongest baseline by 4.85 and 5.07 percentage points, respectively. Ablation analyses showed complementary contributions from all three stages, and sensitivity analyses demonstrated stability across candidate-set sizes and model backbones. Applications to MONDO, MEDIC, and NCBITaxon further established portability to other ontologies. OntologyAligner offers a generalizable framework for accurate mapping of biomedical text to structured ontology concepts. PhenoNormBench and the code are publicly available at this https URL.
[NLP-28] Direct Diversity Optimization for Diverse Successful Trajectories in Preference Post-Training EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）代理在序列决策任务中，通过轨迹级结果标签进行后训练时，难以有效保留同一决策状态下的多种成功策略路径的问题。核心挑战在于现有方法对多路径成功策略的覆盖能力不足，导致模型泛化性受限。其解决方案的关键是提出一种离线后训练方法——直接多样性优化（Direct Diversity Optimization, DDO），该方法由两个核心组件构成：分枝树收集（Divergence-Tree Collection, DTC）与参考相对目标奇数目标（Reference-Relative Target-Odds Objective, RTO）。其中，DTC通过构建以相同决策状态为根的分支集合，实现状态对齐的多样化路径采样；RTO则引导模型在多个成功替代路径中学习与参考策略相比的目标概率分布，从而增强对多样成功策略的建模能力。实验表明，DDO在BabyAI、BabaIsAI和WebShop三个基准任务上均实现了最优的任务成功率与成功策略覆盖率，并在局部动作替换后的恢复率方面表现最佳，显著优于仅基于成功样本的模仿学习和解码阶段的多样性控制方法。
链接: https://arxiv.org/abs/2609.10052
作者: Junwon Ko,Dong-Jae Lee,Minchan Kwon,Sunghyun Baek,Junmo Kim
机构: KAIST(韩国科学技术院); Daejeon, Republic of Korea(大田，大韩民国)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to EMNLP 2026 Main Conference. 19 pages, 11 figures
Abstract:LLM agents for sequential decision tasks are often post-trained with trajectory-level outcome labels, but such labels provide little supervision for preserving multiple successful branches from the same decision state. We study this problem as successful strategy coverage: how broadly a model realizes distinct successful strategies under a fixed rollout budget. We present Direct Diversity Optimization (DDO), an offline post-training method that combines Divergence-Tree Collection (DTC) with the Reference-Relative Target-Odds Objective (RTO). DTC constructs state-aligned branch sets rooted at shared decision states, and RTO trains the model to match reference-relative targets over successful alternatives. DDO achieves the strongest task success and successful strategy coverage among the compared post-training methods across BabyAI, BabaIsAI, and WebShop. It also achieves the highest recovery rate after local action replacement and higher task success and coverage than successful-only imitation and decoding-time diversification controls.
[NLP-29] MedDeID enables locally governed clinical-text de-identification from real or synthetic training data
【速读】： 该论文旨在解决临床笔记中包含个人身份识别信息（PII）导致数据难以用于科研和医疗人工智能（AI）研究的问题，特别是在数据无法离开机构的场景下。其核心挑战在于如何在保障隐私的前提下实现高质量的去标识化（de-identification），同时避免依赖真实敏感数据进行模型训练。解决方案的关键在于提出MedDeID框架，该框架采用本地部署（on-premises）模式，集成内部标注、合成笔记生成、模型训练、推理、伪匿名化及评估全流程；通过使用真实数据或合成数据进行训练，实现了高精度的识别与低误删率——在荷兰医院基准测试中，基于真实数据训练的紧凑型Transformer模型检测到98.9%的标识性文本，且仅误删0.24%非标识性内容；而纯合成数据训练的模型也达到96.1%的检测率。此外，在外部英文合成基准上，无真实文本训练的模型仍可实现99.7%和98.9%的字符级识别率，表明该工作流具备跨语言迁移能力。因此，MedDeID提供了一种可在本地可控环境下，利用真实或合成数据实现高效、安全去标识化的可行路径。
链接: https://arxiv.org/abs/2609.10049
作者: Stig Hellemans,Tom Stroobants,Elyne Scheurwegs,Pieter Meysman,Philippe G. Jorens,Kris Laukens
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 64 pages total: 32-page main manuscript with 4 figures, followed by 32-page Supplementary Information
Abstract:Clinical notes contain personally identifiable information (PII), restricting reuse for research and medical AI, especially when data cannot leave an institution. We developed MedDeID, an on-premises framework combining in-house annotation and synthetic-note generation with model training, inference, pseudonymisation and evaluation. On an independently annotated, adjudicated 300-note Dutch hospital benchmark, a hospital-trained compact transformer detected 98.9% of identifying text while redacting 0.24% of text outside annotated identifiers; a synthetic-only counterpart detected 96.1%. On 100 primary-care notes, the synthetic-trained model achieved higher recall than the hospital-trained model (90.3% versus 87.0%) and greater robustness to identifier-format perturbations. An English instantiation trained without real text detected 99.7% and 98.9% of annotated identifier characters on two external synthetic benchmarks. These results demonstrate transfer of the workflow to another language, but not clinical English performance. MedDeID provides a route to locally governed de-identification using real or synthetic training data.
[NLP-30] Deterministic Prompting for Speaker-Stable Low-Resource Greek TTS INTERSPEECH2026
【速读】： 该论文旨在解决低资源语言（如现代希腊语）在文本转语音（TTS）系统中因缺乏高质量标注语料而导致性能显著下降的问题。其核心挑战在于如何在仅有少量清洁语音数据的情况下，构建接近人类水平的单说话人语音合成系统。解决方案的关键在于提出一套数据清洗流程：利用WhisperX进行语音-文本对齐并筛选高质量数据，从而将播客录音转化为适用于TTS训练的可用数据集；在此基础上，采用基于提示（prompt-based）的多语言模型Parler-TTS（880M）进行微调，并通过引入确定性提示（deterministic prompts）替代由大语言模型生成的风格提示，有效缓解了推理阶段的说话人风格漂移问题；进一步地，仅使用3.5小时单说话人数据训练一个说话人特定的低秩适配器（LoRA），仅更新约5%的参数，即可稳定保留说话人身份特征。实验结果显示，该系统在词错误率（WER）为10.7%（高于自动语音识别基线2.9个百分点）、语音自然度评分（MOS-I）达4.00（接近真人语音的4.36）、说话人一致性评分（MOS-C）达4.24（接近真人水平4.30）的表现下，证明了在有限高质量数据条件下实现接近人类水平的单说话人希腊语TTS是可行的。
链接: https://arxiv.org/abs/2609.10022
作者: Georgios Syllas,Efthymios Georgiou,Kosmas Kritsis,Alexandros Potamianos
机构: 未知
类目: ound (cs.SD); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Interspeech 2026
Abstract:Modern TTS systems approach human quality for high-resource languages but degrade when clean speech data is scarce. Modern Greek exemplifies this, lacking the curated corpora behind state-of-the-art synthesis. We propose a data curation recipe that transforms audiobook recordings into TTS-ready data via WhisperX alignment and filtering. Then we fine-tune Parler-TTS (880M), a prompt-based multilingual model whose pre-training encodes phonetic priors transferable to Greek. During development, we find that LLM-generated style prompts introduce speaker drift at inference. Replacing them with deterministic prompts resolves this, and a speaker-specific LoRA stage trained on 3.5 h of single-speaker data anchors identity while updating ~5% of parameters. Our system achieves WER 10.7% (2.9 above the ASR floor), MOS-I 4.00 (vs. 4.36 human speech), and near-human speaker consistency (MOS-C 4.24 vs. 4.30), showing that robust single-speaker Greek TTS is achievable with limited curated data.
[NLP-31] MetroLLM -Bench: Evaluating Language Models as Transit Kiosk Runtimes
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）作为城市轨道交通自助查询终端（transit kiosk）策略层（policy layer）时的可靠性与实用性问题，尤其关注其在复杂、多模态真实场景下的任务执行能力。核心挑战在于模型需在结构化工具调用与动态交互中准确完成包括路径规划、票价计算、运营中断处理、无障碍服务支持及对抗性输入应对等十一类任务，并输出可机器解析的终端状态（terminal state），以确保系统级可用性。解决方案的关键在于构建一个覆盖六座真实地铁系统（站点数37至414不等）、包含955个案例的综合性基准测试集MetroLLM-Bench，其设计包含两层级评估体系：第一层为14项确定性评分组件（Tier 1），强调逻辑正确性与工具调用准确性；第二层为8项语义质量评分组件（Tier 2），其中6项采用语言模型判别器进行评估，以捕捉自然语言响应的合理性与完整性。研究进一步通过参数高效微调（PEFT）训练出4B规模的Qwen 3.5学生模型，在未使用全量推理资源的情况下，于保留测试集上超越GPT-5.6的两个层级（91.3 vs. 90.6 and 90.0），并达到与GPT-5.4全量推理相当的表现（91.4），且仅需2.6 GB Q4_K_M量化内存。此外，研究揭示了模型规模收益递减现象（9B与27B学生未带来额外提升），以及不同模型尺寸下PEFT增益随规模增长而下降的趋势，凸显策略适应、复合场景建模、无障碍服务与时间推理等维度仍是当前模型的主要瓶颈。最终，该工作不仅提供了一个可复现的评测框架与开源资源，也为生成式AI在关键基础设施人机交互系统中的部署提供了实证依据与优化路径。
链接: https://arxiv.org/abs/2609.10016
作者: Remco Hendriks(Continker)
机构: Continker(Continker)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 23 pages, 5 figures, 10 tables. Code and data at this https URL (tag paper-v1.2); DOI https://doi.org/10.5281/zenodo.21893944
Abstract:We introduce MetroLLM-Bench, a 955-case benchmark for testing language models as the policy layer of a transit kiosk. It covers six real metro systems, ranging from 37 to 414 stations, and eleven categories that include routing, fare calculation, disruptions, accessibility, and adversarial input. In each case, the model must call structured tools and submit a machine-renderable terminal state containing an outcome, a per-ticket fare quote when applicable, and a kiosk action. Fourteen deterministic scoring components form Tier 1; eight semantic-quality components form Tier 2, six of which use a language-model judge. We report Tier 1 and the combined score of both tiers. A stratified 75/25 split reserves 717 cases for training-data generation and 238 for held-out evaluation. We evaluate twenty-six models from six vendors, of which twenty-three are ranked. On the held-out partition, a 4B Qwen 3.5 student trained through parameter-efficient fine-tuning (PEFT) exceeds both GPT-5.6 tiers on Tier 1 (91.3 against 90.6 and 90.0) and matches GPT-5.4 full at maximum reasoning effort (91.4), with a 2.6 GB Q4_K_M footprint. Larger 9B and 27B students provide no further Tier 1 improvement over the 4B student at this training scale. Across the four Qwen sizes, the PEFT gain over the corresponding base model decreases from +7.03 points at 2B (three training seeds) to -0.91 at 27B; every seed shows the same direction at every size. A deterministic rule-based baseline reaches 84.6 on Tier 1, with the remaining language-model advantage concentrated in policy adaptation, compound scenarios, accessibility, and temporal reasoning. Muse Glimmer 30B leads the composite ranking, and serving configuration alone moves the Qwen 3.5-to-3.8 comparison by 2.7 Tier 1 points. The benchmark, harness, reproduction guide, and fine-tuned students are released at this https URL. Comments: 23 pages, 5 figures, 10 tables. Code and data at this https URL (tag paper-v1.2); DOI https://doi.org/10.5281/zenodo.21893944 Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) Cite as: arXiv:2609.10016 [cs.LG] (or arXiv:2609.10016v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.10016 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-32] SalamandraTA at WMT 2026 Terminology Shared Task: Hard Examples Are Better Teachers
【速读】： 该论文旨在解决术语感知翻译（terminology-aware translation）中因标准微调方法效率低下而产生的问题：在大多数训练样本中，模型的输出本就与术语表一致，导致这些样本无法有效学习如何遵循术语规范。其解决方案的关键在于仅保留模型生成结果与术语表相矛盾的样本进行训练，从而显著提升模型对术语规范的遵守能力。通过构建一种双向合成数据管道过滤出此类关键样本，并将其融入指令微调数据集，所提出的SalamandraTA-7b-instruct v3.0系统在WMT26术语共享任务中实现了94.2%的术语准确率和74.6 chrF++的翻译质量，优于绝大多数参赛系统，且无需复杂强化学习框架，仅依赖监督微调即可达成卓越性能。
链接: https://arxiv.org/abs/2609.09999
作者: Xixian Liao,Maite Melero
机构: Barcelona Supercomputer Center(巴塞罗那超级计算中心)
类目: Computation and Language (cs.CL)
备注: To appear at Proceedings of the Eleventh Conference on Machine Translation (WMT26)
Abstract:Terminology-aware translation asks for more than a correct translation: the output must use the exact terms a glossary prescribes. The standard recipe, fine-tuning on glossary-annotated translation pairs, hides an inefficiency: for most examples the glossary prescribes exactly what the model would have produced anyway, so they teach nothing about following a glossary. We therefore keep only the examples where the model’s own translation contradicts the glossary. In a controlled study at fixed data volume, this selection alone raises term accuracy from 78.7% to 89.9%. The filtered data, built by a two-way synthetic pipeline on open models, is part of the instruction-tuning mixture of our public release SalamandraTA-7b-instruct v3.0, which, used exactly as released and wrapped in a document-level inference pipeline, forms the BSC submission to the WMT26 Terminology Shared Task Track 1. At the official WMT26 evaluation, our system achieves 94.2% term success at 74.6 chrF++, with only two of the twenty-two submissions outperforming it on both metrics. On last year’s benchmark, it also surpasses our GRPO-based system, despite being trained solely with ordinary supervised fine-tuning.
[NLP-33] Stable Answers Unfinished Reasoning : Why Self-Consensus Is Not a Safe Early-Exit Signal
【速读】： 该论文旨在解决生成式推理模型（Generative AI）在推理过程中高昂的计算成本问题，具体聚焦于通过“自一致（self-consensus）”策略——即反复探测单一推理轨迹的当前答案，直至多个探测结果达成一致后停止——来降低推理所需的令牌（token）消耗。其核心问题是：是否存在一种既安全又能显著节省令牌的自一致规则，并且该规则能否一经选定便在不同场景中重复使用。研究发现，经过预先注册的3,520种共识规则筛选，在两个模型和三个基准测试上的冻结轨迹重播实验中，均无法通过预先设定的三项验收标准；而采用边界置信度控制（DEER）方法则能全部通过。根本原因在于信号本质的误解：共识仅表明在固定探测流程下当前答案保持稳定，而非推理过程已终止，即存在“共识-终止鸿沟（consensus-termination gap）”。据此停止将导致非终态答案被提前截断。即使在仍可节省32%令牌的规则下，约每九次停止中有一次会基于模型自身后续放弃的答案进行决策，且多数此类过早停止阻碍了模型本可能做出的修正。扩大一致性窗口并未消除该问题，其错误占比趋于稳定在7%左右，而此时节省率已降至8%。通过对探测措辞的调整及人工标注错误分类体系分析，进一步揭示：达成一致的答案常为模型尚未确定的占位符（placeholder）。因此，以一致性作为停止信号的失败并非因阈值不够严格，而是因其持续测量的是错误的对象——即未收敛的中间状态，而非真正的推理终点。
链接: https://arxiv.org/abs/2609.09989
作者: Yunxiang Mo,Donghao Zhao,Hejia Geng
机构: HKUST(香港科技大学); University of Oxford(牛津大学)
类目: Computation and Language (cs.CL)
备注: 21 pages, 9 figures, 10 tables. Yunxiang Mo and Donghao Zhao contributed equally. Code and data will be released at this https URL
Abstract:A natural way to cut reasoning-model inference cost is to repeatedly probe a single partial trajectory for its current answer and stop once probes agree – self-consensus. We ask whether any such rule is both safe and token-saving, and whether one can be selected once and reused. A preregistered sweep of 3,520 consensus rules, replayed on frozen trajectories from two models and three benchmarks, clears none of three acceptance gates fixed in advance; the frontier reproduces on a held-out split and on two unseen models – while a boundary-confidence control (DEER) swept through the same pipeline clears all three. The reason lies in the signal: agreement establishes that the current answer persists under a fixed probing procedure, not that the reasoning has terminated – a consensus-termination gap. Stopping on it commits non-terminal answers. At a rule still saving 32% of the tokens, one stop in nine fires on an answer the trajectory itself later abandons, and most of those stops cut off a correction it would otherwise have made. Widening the agreement window does not remove them: the share levels off near 7%, and by then the saving has fallen to 8%. Probe re-wording and a hand-labelled error taxonomy show the agreed answer is often a placeholder the model had not settled on. Used on its own as the stop signal, agreement fails not because it is insufficiently strict, but because it repeatedly measures the wrong object.
[NLP-34] VLX-VR: An Agent ic-Aware Video Reasoning Model
【速读】： 该论文旨在解决现实世界视频理解中多模态证据（视觉、音频、文本及时间信息）分布分散且动态变化，而传统方法依赖固定上下文与单次推理所导致的适应性不足问题。其核心挑战在于如何在观测不完整、模糊或冲突时实现自适应的证据获取与推理决策。解决方案的关键在于提出一种基于“思考-记忆-观察”循环框架的代理感知型视频推理模型VLX-VR，通过引入可动态调用读写记忆的操作机制，使模型能够在每一步自主判断所需证据、更新记忆状态并融合新观测，进而决定是否继续推理或输出结果。该模型利用包含视频数据和智能体轨迹的多模态数据集，结合强化学习进行训练，以优化证据获取策略、记忆使用效率与推理终止时机。实验表明，VLX-VR在MINERVA基准上达到78.79%的准确率，跨不同视频时长组的准确率方差仅为2.97，展现出优异且稳定的性能；同时，其96.20%的推理轨迹与参考标准一致，约75.80%的样本同时满足答案正确性和证据可追溯性，验证了其在复杂推理任务中的有效性与鲁棒性，但计数、状态变化、因果推理及空间感知仍存在挑战。
链接: https://arxiv.org/abs/2609.09985
作者: Sheng Li,Peng Liu,Qianqian Zhang,Tiancheng Zhao
机构: Om AI Research
类目: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages
Abstract:Real-world video understanding requires integrating visual, audio, textual, and temporal evidence distributed across a video. Yet many pipelines use a fixed video context and single-pass inference, limiting adaptive evidence acquisition when observations are incomplete, ambiguous, or conflicting. We present VLX-VR, an agentic-aware video reasoning model trained within a video reasoning framework defined by a Think–Memory–Observation loop. At each step, VLX-VR determines the needed evidence, invokes read_memory or write_memory, incorporates the returned Observation, and decides whether to continue or produce the task output. We train VLX-VR with multimodal data, including videos and agent trajectories, using reinforcement learning to learn evidence acquisition, memory use, and termination. On MINERVA, VLX-VR achieves state-of-the-art performance among the models included in our comparison, with 78.79% accuracy. Under the original three duration groups, its accuracies are 76.70%, 78.73%, and 80.92%, with a cross-duration accuracy variance of 2.97~ \mathrmpp^2 . On correctly answered samples, 96.20% of VLX-VR’s reasoning traces are consistent with the MINERVA reference reasoning traces and the evidence described by them, while approximately 75.80% of all evaluated samples satisfy both answer correctness and this evidence-grounded trace criterion. These results show strong performance and broadly stable behavior across durations, while counting, state changes, causal reasoning, and spatial perception remain challenging.
[NLP-35] Multi-Functional Embedding Models for Funder Name Disambiguation in Scientific Publication Records
【速读】： 该论文旨在解决科研资助机构名称（funder name）在大规模文献分析中因拼写变体、翻译差异、缩写及粒度不一致导致的命名消歧难题，这一问题严重制约了对科研经费历史分配与分布模式的准确理解。其解决方案的关键在于构建一个面向多语言、多功能的资助机构名称消歧框架：首先通过整合研究组织注册库（ROR）、Web of Science（WoS）和Crossref开放资助者注册表（OFR）构建训练数据集；其次采用基于对比损失（Contrastive Loss）与多重负样本排序损失（Multiple Negatives Ranking Loss）的多任务学习策略，微调Sentence Transformer、Gemma与Qwen3系列的开源嵌入模型，实现高精度匹配（准确率超0.90），显著优于通用大语言模型（如GPT-5.2、Claude-Sonnet-4.6、Gemini-2.5-Flash）；针对未被ROR收录的资助机构，进一步构建名称间相似性网络并识别聚类结构，有效拓展了覆盖范围。该框架具备可复用性，支持多种模型架构与数据集，且通过低成本训练数据生成与多任务学习提升了效率与泛化能力，为跨领域科研资助分析提供了技术基础。
链接: https://arxiv.org/abs/2609.09984
作者: Kanyao Han,Zhiwen You,Jinseok Kim,Jana Diesner
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Understanding the historical allocation and distribution of research funding advances our knowledge of how scientific research is supported across fields, institutions, and regions. However, large-scale analyses are hindered by the lack of comprehensive funder name disambiguation solutions, as funder names often exhibit spelling variations, translations, abbreviations, and inconsistent levels of granularity. In this paper, we present a framework for developing multilingual, multi-functional funder name disambiguation models and demonstrate its application to research publications in biodiversity conservation. To construct a training dataset, we integrated the Research Organization Registry (ROR), which provides unique identifiers for research organizations, with two publication datasets: the Web of Science (WoS) and the Crossref Open Funder Registry (OFR). We used multi-task learning with Contrastive Loss and Multiple Negatives Ranking Loss to fine-tune three open-weight embedding models from the Sentence Transformer, Gemma, and Qwen3 families. The best-performing models achieved accuracy above 0.90 when matching WoS funder names to ROR identifiers, outperforming general-purpose LLMs, including GPT-5.2, Claude-Sonnet-4.6, and Gemini-2.5-Flash, by more than 0.1. For funder names not indexed in ROR, we constructed a similarity network among funder names and identified clusters within it. Finally, we analyzed the disambiguation results and highlighted challenges arising from limited knowledge of smaller funders and funders from non-English-speaking countries. This work provides a reusable framework for funder name disambiguation with potential applicability across different model architectures and datasets, featuring cost-effective training data creation and multi-task learning and disambiguation.
[NLP-36] owards Stress-Aware Sentence-Level Filipino G2P With Weakly-Supervised ByT5 Fine-Tuning
【速读】： 该论文旨在解决菲律宾语（Filipino）在句法层面进行音素转换（G2P）时面临的挑战，即由于缺乏标注有音素转写的标准句级数据，导致传统基于单字的G2P模型难以有效捕捉重音等韵律特征。其核心问题在于如何在无现成句级音素标注的情况下，构建高质量的句级音素数据以支持更精准的G2P建模。解决方案的关键在于利用大语言模型（LLM）辅助的标注流程，结合Wiktionary等开源资源，构建三个句级菲律宾语G2P数据集，并在此基础上对预训练的ByT5模型进行微调。该方法显著提升了模型性能，在手动校正的测试集上将词错误率（CER）降至2.50%，音素错误率（PER）降至0.54%，远优于基础模型的19.74% PER。此外，模型在识别菲律宾语主要重音类别方面表现良好，但在处理“malumi”类同音异义词时仍存在困难，表明该方法具备在句法层面实现同音词消歧的潜力。
链接: https://arxiv.org/abs/2609.09974
作者: Lorenz Bernard Marqueses,Paulo Grane Gabriel Silva,Chastine Cabatay,Ericson Adler Tan,Ann Franchesca Laguna
机构: 未知
类目: Computation and Language (cs.CL)
备注: Accepted at the 10th International Conference on Natural Language Processing and Information Retrieval (NLPIR 2026), Nara, Japan
Abstract:Grapheme-to-phoneme conversion (G2P) refers to the task of converting a sequence of graphemes to a corresponding sequence of phonemes. While Filipino G2P is fairly straightforward due to its shallow orthography, the inclusion of prosodic features such as stress adds a layer of complexity that requires sentence-level context instead of single-word inputs. However, sentence-level data for Filipino typically do not include phoneme transcriptions, posing a challenge for training G2P models. As such, we investigate how to obtain sentence-level phoneme data for Filipino using available data and compare the resulting models with multilingual word-level G2P as well as measure how accurately they predict stress marker position for Filipino. We propose fine-tuning a ByT5-based model, pre-trained on multilingual word-level G2P data, on three sentence-level G2P datasets annotated with an LLM-assisted pipeline guided by data from Wiktionary. This approach produces models that perform well on the G2P task, achieving at best around 0.54% PER and 2.50% CER, a significant decrease compared to base model PER at around 19.74%, on a manually-corrected test set. The model is able to correctly classify most of the main stress classes in Filipino, but struggles particularly with malumi words. We show that a ByT5-based model performs well at sentence-level Filipino G2P and offers strong potential for Filipino homograph disambiguation.
[NLP-37] 5-Dialects-BN: Unmasking the Impact of Transliteration on Bangla Dialectal LLM s EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在低资源语言及方言多样性场景下性能显著下降的问题，尤其聚焦于孟加拉语（Bangla）这一全球使用人数第六多的语言中，其区域方言缺乏系统性标注数据与评估基准的现状。现有资源主要针对标准孟加拉语（Standard Bangla），导致地方方言无法有效支持方言感知型自然语言处理系统的开发与评估。为此，本文提出了首个多标注的孟加拉语方言基准数据集——5-Dialects-BN，该数据集首次实现了罗马化转写（Romanized transliteration）、方言文本、标准孟加拉语、英文翻译以及主观性标签（subjectivity label）在五个主要方言（吉大港、巴里萨尔、诺阿哈利、锡尔赫特和朗普尔）间的对齐。数据集包含6,000条人工标注样本，由母语者及语言学本科生共同标注并交叉验证，确保方言真实性和语义保真度。其关键贡献在于构建了一个标准化、多维度的跨方言标注框架，支持方言识别、方言到标准语归一化、机器翻译、主观性分类以及多语言大模型的参数高效微调（如LoRA）等任务，为低资源、方言敏感的自然语言处理研究提供了可复现的评估基础。
链接: https://arxiv.org/abs/2609.09964
作者: Md Mahir Jawad,Galib Mahmud Jim,Rafid Ahmed,Mir Sazzat Hossain,Md Fahim,Md Farhad Alam Bhuiyan
机构: BRAC University(布拉克大学); Penta Global Limited; University of Central Florida(中佛罗里达大学); Center for Computational Data Sciences, Independent University, Bangladesh(独立大学计算数据科学中心)
类目: Computation and Language (cs.CL)
备注: 31 pages, 18 figures, 26 tables. Accepted to EMNLP 2026 (Main Conference)
Abstract:Large Language Models (LLMs) have achieved remarkable progress across natural language processing (NLP) tasks, yet their capabilities degrade sharply for low-resource languages and dialectally diverse settings. Bangla, the world’s sixth most spoken language, exemplifies this gap: existing resources overwhelmingly target Standard Bangla, leaving its regional dialects without the benchmarks needed to develop or evaluate dialect-aware systems. We address this gap with 5-Dialects-BN, the first multi-annotation Bangla dialect benchmark to align Romanized transliteration with dialectal text, Standard Bangla, English, and subjectivity labels across five regional varieties. The dataset comprises 6,000 manually annotated entries spanning five major dialects: Chittagong, Barisal, Noakhali, Sylhet, and Rangpur (Chittagong 1,900; Noakhali 1,500; Sylhet 1,200; Barisal 700; Rangpur 700), reflecting natural online availability. Each entry is enriched with five aligned annotations: the original dialectal text, a Romanized transliteration, an English translation, a Standard Bangla translation, and a subjectivity label (subjective vs. objective). Annotations were produced and cross-validated by native speakers and undergraduate linguistics students to ensure dialectal authenticity and semantic fidelity. The resulting resource supports a diverse suite of tasks, including dialect identification, dialect-to-standard normalization, machine translation, subjectivity classification, and parameter-efficient fine-tuning (e.g., LoRA) of multilingual LLMs. By providing a standardized, multi-annotation benchmark, 5-Dialects-BN enables principled evaluation of LLMs on dialectally diverse Bangla and lays a foundation for further research in low-resource, dialect-aware NLP.
[NLP-38] Improving Cross-Lingual Token Representations by Adding a Pinch of SALT
【速读】： 该论文旨在解决现有跨语言句子编码器（cross-lingual sentence encoder）在训练阶段以句级对齐为目标，却广泛应用于词元级任务（如幻觉检测和序列标注）所导致的表征不匹配问题。其核心解决方案是提出一种轻量级后训练方法SALT，通过向已有的句子编码器注入片段级（span-level）监督信号，从而有效提升词元级表示能力。关键创新在于利用片段级标注作为额外监督信号，弥补了传统句级训练在细粒度语义建模上的不足，实验表明SALT在五个多语言词元级基准上于四项取得最优表现，同时提升了句级任务中的跨语言检索与分类性能，验证了片段级监督在优化词元与句级表示方面的有效性。
链接: https://arxiv.org/abs/2609.09953
作者: Guillem Ramírez
机构: ILCC, University of Edinburgh (爱丁堡大学交互式语言与计算中心)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Cross-lingual sentence encoders enable scalable transfer across hundreds of languages, powering applications such as translation mining and zero-shot learning in low-resource settings. Although trained for sentence-level alignment, they are increasingly also applied to token-level tasks such as hallucination detection and sequence tagging, exposing a mismatch between training and usage. We propose SALT, a lightweight post-training method that improves token representations by injecting span-level supervision into existing sentence encoders. Across five multilingual token-level benchmarks, SALT achieves the best overall results on four of them, outperforming alternative fine-tuning strategies and competitive encoders. It also improves sentence-level performance on cross-lingual retrieval and classification tasks. These results demonstrate that span-level supervision is an effective signal for improving both token and sentence representations.
[NLP-39] Vague2Detect: Handling Ambiguous Prompts in Knowledge-Based Open-World Detection
【速读】： 该论文旨在解决现有目标检测模型在处理功能型或模糊性提示（vague prompts）时的语义对齐问题，尤其针对传统模型如YOLO受限于固定类别列表、而开放词汇模型如YOLO-World在将模糊语言与目标物体正确关联时频繁出现误匹配的缺陷。其解决方案的关键在于提出一种名为Vague2Detect的混合式检测流水线：首先通过微调的Sentence-BERT从结构化的家庭知识库（Knowledge Base, KB）中检索候选对象；随后由YOLO-World验证这些候选对象在图像中的存在性。对于知识库之外的提示，引入大语言模型（GPT-3.5-turbo）生成候选描述，并动态扩展知识库以覆盖新概念。实验表明，在家庭场景基准测试中，仅使用YOLO-World的模糊提示成功率（Vague Prompt Success Rate, VPSR）仅为32%，而Vague2Detect将该指标提升至61%（高精度），并进一步增至85%（结合GPT回退机制）。该方法的核心优势在于通过融合语言理解与视觉-语义匹配，实现对任务驱动型模糊查询的精准语义接地。
链接: https://arxiv.org/abs/2609.09949
作者: Ibrohimjon Muminov(1),Jihie Kim(1) ((1) Dongguk University, Seoul, South Korea)
机构: Dongguk University (东国大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 15 pages, 4 figures, 3 tables. Code: this https URL
Abstract:Real-world detectors must often interpret functional or ambiguous prompts, yet conventional models such as YOLO remain restricted to fixed class lists. Even open-vocabulary models like YOLO-World frequently misalign vague language with the intended objects. Building on our prior work Commonsense-Guided Open-World Object Detection Using LLMs and Visual-Semantic Matching, we address YOLO-World’s limitations in grounding task-driven queries. We propose Vague2Detect, a hybrid pipeline in which a fine-tuned Sentence-BERT retrieves candidates from a structured household Knowledge Base (KB), and YOLO-World verifies their presence in the image. For prompts outside the KB, a large language model (GPT-3.5-turbo) generates candidate descriptions, dynamically expanding the KB to cover novel concepts. On a benchmark of household scenes using custom images and an Open Images V7 subset, YOLO-World alone achieves only 32% Vague Prompt Success Rate (VPSR), the ability to map ambiguous queries to correct detections. In contrast, Vague2Detect improves performance to 61% VPSR with high precision, and up to 85% when augmented with GPT fallback.
[NLP-40] Contrastive Projection: Reading Transformer Internals by Differencing Logit Lenses
【速读】： 该论文旨在解决生成式模型（Generative AI）中内部状态在词元空间（token space）中的可解释性问题：尽管通过逻辑值透镜（logit lens）读取单个隐藏状态的操作简便，但在中间层其信号极易被模型对几乎所有输入都预测的通用词元所主导，导致结果不可靠。为克服这一局限，论文提出的关键解决方案是“差分读取”（reading the difference）——即对两个结构相近的提示（prompt）的隐藏状态进行相减，并通过未嵌入矩阵（unembedding）投影，从而消除共有的语义成分，凸显二者之间的差异。这一操作等价于通过逻辑值透镜读取一个RepE/ActAdd类的控制向量（steering vector）。研究构建了一个无需训练的追踪框架，可在每个位置、子层和注意力头上进行读取，并对设计的基线平均化处理。该方法成功追踪到Phi-2模型中复合名词的MLP-注意力链路径，经激活修补验证后确认；且在三种不同架构中，仅通过读出与探测而非修补即可复现相同区分特征。该方法能够识别真实实体与虚构实体之间的语义差异，以及隐喻作为跨领域映射集合的本质，而非单一的“修辞性”特征。跨种子对照实验进一步表明，尽管不同初始化网络的词元表示高度异构（前10名重叠仅为0.08），但核心语义区分仍稳定出现，说明计算在词元空间的表现形式具有网络特异性，而其所揭示的语义区分则具有普遍性。
链接: https://arxiv.org/abs/2609.09902
作者: Olli Tuomi
机构: Evident Solutions Oy
类目: Computation and Language (cs.CL)
备注: 27 pages, 34 tables. Code and data: this https URL
Abstract:Reading a transformer’s internal states in token space is easy to do and hard to trust: a logit lens on a single hidden state is dominated, at intermediate layers, by the generic tokens the model would predict for almost any input. We read the difference instead. Subtracting two closely matched prompts’ hidden states and projecting through the unembedding cancels the shared component and surfaces what separates them, an operation equivalent to reading a RepE/ActAdd steering vector through a logit lens. Built into a training-free tracer that reads at every position, sub-layer, and head and averages over designed baselines, it traces a compound- noun MLP-attention chain in Phi-2, confirmed there by activation patching, with the same distinction recovered across three architectures by readout and probe rather than by patching; it reads what retrieval surfaces for real versus fictional entities, and reads metaphor as a set of domain-to-domain mappings rather than a single figurativity feature. A cross-seed control marks the boundary: across five networks differing only in initialization, the same distinction surfaces as almost entirely different tokens (top-10 overlap 0.08). What a computation looks like in token space is network-specific; the distinction it draws is not
[NLP-41] Deep and shallow biases in language models
【速读】： 该论文旨在解决大语言模型在生成回答时存在重复选择同一答案的问题，尤其关注这种集中现象是否源于模型固有的稳定偏好，还是仅由特定提示（prompt）的表述方式所引发的临时偏差。其核心挑战在于区分“深层偏见”（Deep biases）——即模型在多种情境重述下仍持续表现出的稳定偏好——与“浅层偏见”（Shallow biases）——即仅依赖于初始提示措辞而产生的临时性倾向。解决方案的关键是提出一种新的偏见深度评分（bias depth score），该指标同时衡量模型在直接提示下的首选响应强度及其在情景重构（scenario reframing）后是否依然保持不变。实验结果表明，在4,442个观点类提示和四种大语言模型中，仅有约四分之一的集中偏好能在重述后持续存在，这些持久性偏好被定义为“深层偏见”，且更可能源自预训练阶段，并在监督微调（SFT）过程中得以保留。相比之下，浅层偏见则更容易通过持续微调或基于提示的去偏策略消除。因此，偏见深度评分能够有效将模型的稳定学习偏见与提示词敏感的伪象区分开来，克服了传统单次提示评估方法的混淆问题。
链接: https://arxiv.org/abs/2609.09901
作者: An Vo,Vy Tuong Dang,Khai-Nguyen Nguyen,Emilio Villa-Cueva,Thamar Solorio,Anh Totti Nguyen,Daeyoung Kim
机构: MBZUAI; University of Michigan (密歇根大学); KAIST (韩国科学技术院); University of Virginia (弗吉尼亚大学); Auburn University (奥本大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models often repeatedly select the same answer even when many alternatives are plausible. Prior work treats this concentration as bias, but it does not distinguish stable model preferences from responses that depend on a particular prompt wording. We introduce a bias depth score that measures both how strongly a model prefers its top answer under direct prompting and whether that answer survives scenario reframing. Across 4,442 opinion prompts and four large language models, only about a quarter of the concentrated preferences survive reframing. We call these persistent cases Deep biases, and the remaining prompt-dependent cases Shallow biases. Our results show that Deep biases are more often inherited from pretraining and preserved through SFT. Under both continued fine-tuning and prompt-based debiasing for diversity, Deep biases are consistently harder to remove than Shallow biases. Bias depth therefore separates stable learned biases from prompt-wording artifacts that single-prompt metrics conflate. Code, models, and data are available at this http URL.
[NLP-42] Strangers to Themselves: What Language Models Say About Themselves Is Generic
【速读】： 该论文旨在解决生成式 AI 模型在自我报告行为预测中的可信度问题，即模型对其自身行为的描述是否真正反映其内在认知或仅是对外部行为模式的泛化推断。核心问题是：当要求模型预测自身在特定情境下的行为表现时，其所提供的答案究竟源于对“自我”的真实自知（self-knowledge），还是基于对通用人工智能助手行为规律的抽象理解？解决方案的关键在于将“自我认知”转化为可验证的预测测试——通过设计九项行为评估任务，测量模型在不同条件下的实际行为，再让模型预测这些行为的发生率，并与移除自我指涉的控制组（如针对“具备能力的AI代理一般”进行提问）进行对比。研究发现，直接自述的预测相关性极低（r = +0.04），即使提供具体行为实例，预测能力提升也有限（至+0.24）；而关于“通用能力强的AI代理”的问题却能达到相似甚至更高的预测准确性（+0.28），且其他模型对自己行为的预测同样能有效预测目标模型的行为。此外，模型规模的提升并未带来显著的自我特异性预测增强，表明性能提升更可能源于对通用行为模式的更好建模，而非自我意识的深化。第一人称表述虽未提高准确性，但存在明显的美化偏差，倾向于低估自身有害行为。最后，基于自身行为记录的微调虽可实现特定情境下的自我预测，但会改变原始行为模式，且泛化能力差。因此，结论为：当前大模型的行为预测主要反映其对通用智能助手行为的理论理解，叠加一种自我偏袒倾向，而非真正意义上的“自我知识”。
链接: https://arxiv.org/abs/2609.09899
作者: Phil Blandfort,Urja Pawar
机构: Predictably Weird(预测性奇怪); Independent(独立)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Computers and Society (cs.CY)
备注:
Abstract:Language models can fluently describe how they would behave: whether they would cave to pushback, misuse a tool, or lie under pressure. Is that description actually about the model speaking? We turn self-knowledge into a prediction test. Across nine behavioral evaluations, we measure how a model behaves under different conditions, ask it to predict those rates, and compare its predictions with controls that remove the self from the question. We find that: (i) Direct self-report is weak (r = +0.04), and even showing the model the exact items only raises prediction to +0.24. Crucially, the same item-informed question about “capable AI agents in general” does just as well (+0.28), while other models’ answers about themselves predict the target model at least as well as its own. (ii) Frontier scale does not detectably change this pattern: any gains in prediction are not self-specific, and are consistent with a better theory of how AI assistants behave rather than better self-knowledge. (iii) First-person framing does have one robust effect: it shifts reports in the flattering direction, understating harmful behavior relative to the same question about a generic agent. (iv) Finetuning on a model’s own behavioral record can teach narrow self-predictions, but it also changes the behavior being predicted and the gains do not transfer broadly. The practical implication is simple: asking a model what it would do mostly reveals a theory of AI assistants in general, plus a favorable bias, rather than privileged knowledge of that model.
[NLP-43] Leverag ing Fine-grained Error Correction in Korean Speech Recognition for Consultation Services
【速读】： 该论文旨在解决低资源语言（特别是韩语）在真实复杂场景下自动语音识别（ASR）文本后编辑中因缺乏高质量标注语料和专用纠错方法而导致的错误修正难题。现有纯文本后编辑方法在低资源语言上表现受限，主要由于缺乏针对文本级纠错任务设计的标注数据与有效模型架构。为此，研究者构建了首个面向对话级ASR错误修正的大型韩语基准数据集DasanCallDial，包含1,974段真实的客服对话及115,460个话语，覆盖真实业务场景中的语音转录误差。其核心解决方案是提出一种基于检测器门控的上下文跨度修正框架（Detector-Gated Contextual Span Correction, DCSC），该框架采用两阶段机制：首先通过编码器-基础检测器实现细粒度的词级别错误定位，再由基于语言模型的校正器对跨度级别（span-level）的错误进行精细化修正；同时引入对话级上下文增强策略，使模型能够利用话语历史信息进行语义消歧。通过多粒度联合建模，该方法显著提升了低资源环境下韩语ASR文本的纠错性能，有效克服了通用大语言模型（LLM）在资源匮乏场景下的局限性。
链接: https://arxiv.org/abs/2609.09889
作者: Yonghyun Jun,Jimin Lee,Hwan Chang,Dongho Shin,Seolah Kim,Hwanhee Lee
机构: Chung-Ang University (中央大学); Korea Local Information Research Development Institute (韩国地方信息研究开发院); SK intellix
类目: Computation and Language (cs.CL)
备注: Published in Engineering Applications of Artificial Intelligence
Abstract:Automatic Speech Recognition (ASR) technology is fundamental to customer service automation and large-scale transcription. However, even advanced ASR models exhibit inevitable errors in complex real-world environments such as call center conversations. When privacy restrictions preclude audio access, error correction must rely on text-based post-editing. Existing text-only approaches face significant challenges in low-resource languages, mainly due to a critical scarcity of annotated corpora and tailored correction methodologies. For Korean, this resource gap is particularly pronounced, as existing resources are predominantly designed for ASR training rather than text-based error correction. To address this, we introduce DasanCallDial, the first large-scale Korean benchmark dataset specifically curated for dialogue-level ASR error correction. Derived from genuine call center interactions, it comprises 1,974 dialogues with 115,460 utterances. Leveraging this resource, we propose Detector-Gated Contextual Span Correction (DCSC), a text-only post-editing framework for error-sparse Korean speech recognition transcripts. DCSC combines an encoder-based detector that first performs token-level error detection, followed by a language model-based corrector trained to rectify fine-grained span-level errors. Additionally, we employ dialogue-level context augmentation to enable the model to leverage discourse history for disambiguation. By employing multi-level granularity, our method achieves state-of-the-art performance, effectively overcoming the limitations of general LLMs in low-resource settings.
[NLP-44] When Does Defendant Statement Matter? A Study of Bias and Persuasion in LLM -Simulated Jurors EMNLP2026
【速读】： 该论文旨在解决生成式 AI（Generative AI）在模拟普通法陪审团审判中的行为机制问题，特别是探讨被告在法庭上的陈述如何影响由大语言模型（LLM）模拟的陪审员决策。其核心挑战在于理解情感诉求、意识形态偏见与背景契合度等因素在陪审团判断中的作用。解决方案的关键在于构建了一个名为JuryBench的基准数据集，涵盖美国刑事法中的争议性刑事案件，并系统性地设计具有不同社会背景的被告角色，使其在法庭上发表带有不同情感强度或反驳力度的陈述；同时，模拟了具有多样化意识形态谱系的陪审员群体。通过对20个前沿大语言模型进行测试，共获得43.2万条判决结果及推理依据，量化分析了陪审裁决严厉程度的变化。研究发现：情感化陈述可能适得其反，引发陪审员对被告诚信的质疑；被告与陪审员之间的背景契合度是比其他单一因素更强且显著的影响因素，陪审员普遍对异质背景被告更严厉，对同质背景被告更宽容；此外，陪审员的意识形态也显著影响判决严重性。这些结果揭示了使用大语言模型模拟陪审团推理的潜力与风险，强调需对其应用进行严谨评估。
链接: https://arxiv.org/abs/2609.09887
作者: Cho-Ying Wu
机构: Bosch AI Research(博世人工智能研究院)
类目: Computation and Language (cs.CL); Computers and Society (cs.CY)
备注: Accepted to EMNLP 2026
Abstract:LLMs have been used to simulate human decision-making in professional settings, yet their behaviors in common-law jury trials remain unexplored. We study when and how a defendant’s courtroom statement affects LLM-simulated jurors, focusing on persuasion, ideological bias, and background-based affinity. To support the analysis, we introduce JuryBench, a benchmark containing controversial criminal cases in U.S. criminal law. In each case, a defendant can claim various plausible justifications to support acquittal or reduced liability. We fix the base case and design defendants of different backgrounds, who give courtroom statements with varying emotional appeal or rebuttal. Jurors with diverse ideological profiles across the spectrum are simulated. We examine 20 frontier LLMs, resulting in a total of 432K decisions and rationales, and quantify changes in verdict severity. Our findings show that LLM-jury simulation echoes many human-jury findings. First, emotional persuasion can be detrimental, since jurors may perceive it as evidence of guilt or inconsistency. Next, we show that background fit between jurors and defendants is a stronger and significant factor than other isolated factors, and that jurors are in general harsher toward opposite-background defendants and lenient toward same-background ones. Finally, we find that juror ideology also strongly shapes severity judgments. These findings highlight both the promise and risks of using LLMs to model jury reasoning and call for careful evaluation. The data and code are available at this https URL
[NLP-45] S3-Bench: Evaluating Speech Interaction Models as Scientific Voice Assistants
【速读】： 该论文旨在解决多模态大语言模型（Multimodal Large Language Models, MLLMs）在科学领域语音交互中表现不足的问题。尽管现有模型在通用语音助手任务中展现出卓越性能，但在涉及专业术语、缩略语使用习惯及符号表达口语化等复杂科学语境下，其表现仍不充分。为系统评估这一问题，论文提出S^3-Bench框架，涵盖10个主要学科领域，包含用于语音问答的知识集和用于多轮渐进式对话的对话集。该框架通过将完整对话回合分解为语音识别、感知理解、知识调用与推理、响应发音四个阶段，系统刻画了现有方法在科学语音交互中的共性挑战与性能权衡。实验结果揭示了当前模型在用户适应性以及生成准确、全面且高效响应方面的持续局限性，凸显了提升科学领域专用语音交互能力的关键需求。
链接: https://arxiv.org/abs/2609.09852
作者: Heyang Liu,Jiayi Huang,Wenyang Xiao,Ziyang Cheng,Lixin Zhang,Zhen Liu,Miao He,Ronghua Wu,Qunshan Gu,Yanfeng Wang,Yu Wang
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:The advance of multimodal large language models (MLLMs) has fundamentally reshaped the paradigm of human-computer interaction, especially speech interaction models capable of seamless conversations. Despite remarkable performance as general voice assistants, their performance in specialized domains remains underexplored, particularly in scientific areas. Scientific interactions introduce formidable challenges, involving rare technical terminology, spoken norms of abbreviations, and the natural verbalization of symbolic special expressions. In this paper, we introduce S ^3 -Bench, a systematic evaluation framework covering 10 major disciplines, consisting of a Knowledge set for speech question-answering and a Dialogue set for multi-turn progressive interactions with simulated user agents. By decomposing a complete atomic turn into stages of speech recognition, perception, knowledge utilization with reasoning, and response pronunciation, we systematically characterize the common challenges and performance tradeoffs of existing approaches. Furthermore, experiments on multi-turn interactions reveal persistent limitations in user adaptation and the generation of accurate, comprehensive, and efficient responses.
[NLP-46] HyperTrace: Hypothesis-Based Preference Tracing for Online LLM Personalization EMNLP2026
【速读】： 该论文旨在解决个性化语言模型在用户偏好难以显式表达且随交互逐步显现的背景下，如何有效融合长期偏好与短期话题需求的问题。现有无训练（training-free）方法依赖历史记录或检索记忆，但在处理长短期偏好冲突时表现不佳。其解决方案的关键在于提出一种名为HyperTrace的无训练框架，将在线个性化建模为对潜在用户偏好的追踪过程；该框架通过维护可解释的自然语言假设（涵盖短期意图与长期偏好），并利用基于大语言模型（LLM）的代理选择模型，以类似粒子滤波（SMC）的重加权机制动态更新这些假设。通过跨对话轮次和会话持续迭代更新，实现无需参数更新的个性化响应生成。实验在PRISM和PersonaMem-v2数据集上验证了该方法在响应一致性、偏好预测准确性和个人画像稳定性方面均优于强基准模型，证明了对潜在用户偏好进行追踪在实现鲁棒个性化中的有效性。
链接: https://arxiv.org/abs/2609.09835
作者: Jianzhi Shen,Keyu Mao,Minghao Shao,Chuanyang Jin,Yusong Wang,Ailiang Lin,Kotaro Funakoshi,Manabu Okumura,Tianmin Shu,Muhammad Shafique
机构: Johns Hopkins University(约翰霍普金斯大学); Institute of Science Tokyo(东京科学研究所); NYU Tandon(纽约大学坦登工程学院); NYU Abu Dhabi(纽约大学阿布扎比分校)
类目: Computation and Language (cs.CL)
备注: Accepted to Findings of EMNLP 2026
Abstract:Personalized language models aim to adapt responses to individual users, whose preferences are often latent and revealed gradually through interaction. Existing training-free methods rely on stored histories or retrieved memories, but they often struggle to reconcile long- term preferences with short-term topic-specific needs. To address this issue, we propose HyperTrace, a training-free framework that formulates online personalization as latent preference tracing. HyperTrace maintains interpretable natural-language hypotheses over short-term intent and long-term preferences, and updates them through an SMC-style reweight process using an LLM-based surrogate choice model. By updating these hypotheses across turns and sessions, HyperTrace enables personalization without parameter updates. Experiments on PRISM and PersonaMem-v2 show that HyperTrace improves response alignment, preference prediction, and profile consistency over strong online baselines, demonstrating the effectiveness of tracing latent user preferences for robust personalization. Code and scripts are available in the repository: this https URL.
[NLP-47] How Frag ile Is Safety Alignment at Frontier Scale? A Single-Direction Attack on a 320B MoE
【速读】： 该论文旨在解决生成式 AI（Generative AI）在混合专家模型（Mixture-of-Experts, MoE）架构中，传统定向消融（directional ablation）攻击失效的问题。经典方法依赖于对密集模型中单一残差流（residual stream）的“拒绝方向”进行投影，从而削弱模型的拒绝能力，但其有效性此前仅在参数量约700亿以下的密集模型上得到验证。随着前沿模型如GLM-5.3-Flash（320B参数、288个路由专家、四路超连接残差结构、块FP8量化）采用分段化、稀疏化的专家架构与量化权重，残差流不再为单一张量，导致原有攻击范式难以直接适用。研究发现，尽管该攻击在架构层面仍有效，但其关键作用机制已发生根本性迁移：拒绝能力的削弱并非来自单一模块的编辑，而是多个组件——注意力层、密集前馈层及路由专家层——协同干预的结果，三者联合编辑可使拒绝分数下降0.776，其中74%的效应仅在联合干预下显现。而传统基于模块名称匹配的攻击路径仅捕获了其中0.066的效应，因而无法在MoE模型中有效触发。此外，攻击效果具有方向特异性，随机正交方向的消融不影响拒绝行为；且在所有编辑操作后，特定类别（暴力、性内容、仇恨言论）相关的子空间仍保留显著拒绝能力，表明拒绝机制具有高度任务聚焦的冗余结构。研究最终报告了该方法在七项有害内容基准上实现41–89个百分点的性能下降，且未观测到模型通用能力退化，并明确了其在复杂架构中的作用边界。
链接: https://arxiv.org/abs/2609.09793
作者: Yi Shi,Tanyu Chen,Kai Shen
机构: Continuum AI
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 20 pages, 14 tables
Abstract:Directional ablation removes an aligned language model’s ability to refuse by projecting a single “refusal direction” out of the weights that write the residual stream. It needs no gradient-based training and no optimization, only a few hundred contrastive prompts, which makes it the canonical white-box attack on open-weight alignment. However, it has been established only on dense models up to roughly 70B parameters. We study whether it survives the shift to frontier mixture-of-experts (MoE) models whose residual streams are no longer a single tensor and whose weights ship quantized. We apply it to GLM-5.3-Flash (320B parameters, 288 routed experts, a four-wide hyper-connection residual, block-FP8). The attack survives the architecture, but what it reaches is no longer where a reader of the original recipe would look for it. Editing the attention, dense and routed-expert writers on their own removes 0.039, 0.016 and 0.148 of refusal respectively; editing all three together removes 0.776. As a result, 74% of the effect exists only under the joint intervention. The part the conventional recipe reaches by module-name matching accounts for 0.066 of that 0.776, which is why it fails silently on an MoE. The effect does not follow from removing just any direction: ablating a random direction orthogonal to it leaves refusal unchanged. A category-concentrated residue survives every edit we tried: subspaces fitted on violence, sexual content and hate leave measurable refusal at every rank from 1 to 12. We report the method, the 41-89 percentage-point reductions it achieves across seven harmful benchmarks with no detected change in capability, and the boundary where it stops.
[NLP-48] MUCnoHARM@GermEval Shared Task 2026: Retrieval-based In-Context Learning for Defamatory Offences and Where It Falls Short
【速读】： 该论文旨在解决在线平台上诽谤性言论（defamatory offences under §§ 185-187 StGB）的自动检测问题，尤其关注具有刑事相关性的社交媒体内容识别。其核心挑战在于如何在少样本（few-shot）场景下有效提升模型对刑事相关言论的识别能力。解决方案的关键在于探索基于检索的上下文学习（Retrieval-based In-Context Learning, RetICL）策略，通过引入相关法律案例或示范文本以增强模型判断能力。然而研究发现，尽管检索增强策略在理论上具备优势，实际效果仅略优于随机示范，且不及经过优化的静态示范集；更重要的是，模型选择显著影响性能，远超其他系统设计因素。此外，现有模型普遍存在过度预测刑事相关性的问题，同时仍遗漏26%-57%的真实刑事相关帖子，表明其更适合用于初步筛选（triage）而非完全自主的内容审核。
链接: https://arxiv.org/abs/2609.09791
作者: Kristin Gnadt,Maximilian Meidinger,Matthias Aßenmacher
机构: Central Office for Information Technology in the Security Sector (ZITiS), Munich, Germany; Department of Statistics, LMU Munich, Germany; Munich Center for Machine Learning (MCML), Germany
类目: Computation and Language (cs.CL)
备注: accepted at GermEval Workshop on Harmful Content Detection @ KONVENS 2026
Abstract:With hate speech being ubiquitous online, automatic detection is crucial, in particular when it comes to criminally relevant social media posts. We study a variety of retrieval-based in-context learning (RetICL) strategies for detecting defamatory offences under §§ 185-187 StGB (the subject of GermEval 2026 Subtask 4). Few-shot prompting beats zero-shot, but retrieval-based approaches offer only marginal gains over random demonstrations, and even fall behind an optimised static set of demonstrations. Providing concrete legal knowledge helps, yet model choice outweighs every other system choice. Models over-predict criminal relevance while still missing 26-57% of criminally relevant posts, suiting them for triage rather than autonomous moderation.
[NLP-49] LogiScope-VQA: Benchmarking Vision-Language Models for Logistics Hazard Identification in Industrial Scenarios
【速读】： 该论文旨在解决大型多模态模型（Large Multimodal Models, LMMs）在工业仓库场景中实际部署时，因缺乏真实工业数据而难以实现人类专家级的风险感知、理解与推理能力的问题。其核心挑战在于现有数据受限于商业保密协议，导致模型训练与评估缺乏真实世界物流环境的充分覆盖。为此，研究者构建了首个面向物流作业场景的基准数据集LogiScope-VQA，包含2,476张图像和2,918段视频，以及由人工标注者精心设计并验证的10,274个视觉问答（VQA）样本。该数据集基于18类核心工业对象和20种风险类型，设计了39项子任务，涵盖工业元素感知、仓储知识理解与潜在风险推理三大主题，并引入动态思维预算配置与双维度风险偏见分析机制，以系统评估LMMs在复杂场景下的表现。实验结果表明，即使是最先进的闭源模型（如GPT-5.5、Gemini-3.1-Pro、Claude-Opus-4.7），在综合感知—理解—推理链路下的风险识别能力仍显著落后于人类水平，揭示了当前模型在跨模态协同推理方面存在巨大提升空间。此外，研究还发现普遍存在的安全偏见问题，严重影响模型在真实工业场景中的可部署性。解决方案的关键在于通过高保真、结构化的真实工业数据集推动对模型在风险认知能力方面的系统性评估与改进。
链接: https://arxiv.org/abs/2609.09790
作者: Hanjing Zhou,Mingze Yin,Ying Lian,Jun Ma,Chang-Yu Hsieh,Yanbing Zhou
机构: Cainiao Group, Alibaba Group(菜鸟集团，阿里巴巴集团); Zhejiang University(浙江大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Large Multimodal Models (LMMs) large-scale deployment in industrial warehouse settings specifically necessitates that models exhibit human-expert-level hazard-oriented perception, understanding, and reasoning capabilities. However, the scarcity of real industrial data, tightly coupled to commercial terms, significantly hampers further advancement. To bridge this gap, we curate LogiScope-VQA to investigate the practical applicability of mainstream LMMs in real-world logistics operations. LogiScope-VQA comprises 2,476 images and 2,918 videos primarily sourced from real-world logistics parks, along with 10,274 VQAs meticulously curated and validated by human annotators. Grounded in 18 core objects and 20 risk types, we devise 39 subtasks aligned with three principal themes: industrial element perception, warehouse knowledge understanding, and potential risk reasoning. Furthermore, we incorporate dynamic thinking-budget configurations and dual-dimensional risk bias analyses to elucidate the properties of LMMs. Extensive experiments unveil that even powerful proprietary models, including GPT-5.5, Gemini-3.1-Pro, and Claude-Opus-4.7, exhibit a significant gap relative to human performance. The unique challenge of jointly integrating perception, understanding, and reasoning for hazard identification poses substantial headroom for further improvement on LogiScope-VQA. We additionally reveal the pervasive security bias issue that impedes LLMs’ practical deployment in real-world settings. The industrial dataset is publicly available under the CC BY-NC-SA 4.0 license.
[NLP-50] ROAM: Robust Organization of Atomic Memories for Agents through Semantic Relations
【速读】： 该论文旨在解决长期语言模型智能体在跨交互过程中依赖外部记忆时所面临的原子记忆（atomic memory）冗余、重叠与冲突问题。现有方法通常由大语言模型（LLM）管理器直接执行记忆的增删改写操作，将语义理解、存储决策与内容生成耦合于单一易出错的流程中。为克服这一局限，论文提出一种关系引导的框架ROAM，其核心在于利用原子性进行记忆管理的同时支持更丰富的回答时表征。ROAM通过分类新输入与已存储原子对的关系（独立、等价、方向性包含或冲突），将观察结果组织为具有主动主角色（Primary）和辅助证据角色（Evidence）的结构，并通过融合机制整合互补信息与时间变化，生成紧凑且可能非原子化的视图。仅主角色视图用于回答检索，从而避免冗余或过时原子之间的竞争。实验表明，ROAM在多种模型与评估设置下可提升答案准确率最高达29.8个百分点；消融实验验证了不同关系分类与融合机制的互补优势，且机制分析显示关键源召回率提高15.6点，混淆词占比降低11.5点，同时在不同规模管理器下保持鲁棒性。
链接: https://arxiv.org/abs/2609.09778
作者: Jianjie Zheng,Peng Lai,Sijie Cheng,Jiehui Zhao,Lei Yang,Guanhua Chen
机构: Southern University of Science and Technology (南方科技大学); Tsinghua University (清华大学); RayNeo.AI; Deepexi Technology Co. Ltd. (深盒科技有限公司)
类目: Computation and Language (cs.CL)
备注: 21 pages, 4 figures, 10 tables
Abstract:Long-term language-model agents rely on external memory across interactions. Atomic memories are particularly useful: their fine-grained semantic boundaries enable precise retrieval and direct comparison between observations. Yet accumulating atoms inevitably become redundant, overlapping, or conflicting. Existing methods often ask an LLM manager to add, update, delete, or rewrite memories directly, coupling semantic interpretation, storage decisions, and content generation in one error-prone operation. We introduce ROAM, a relation-guided framework that uses atomicity for management while allowing richer answer-time representations. ROAM classifies incoming–stored atom pairs as independent, equivalent, directionally subsuming, or conflicting, then organizes observations into active Primary and supporting Evidence roles. Fusion subsequently combines complementary details and temporal changes into compact, potentially non-atomic views. Only Primary views are retrieved for answering, preventing redundant or outdated atoms from competing independently. Across models and evaluation settings, ROAM improves answer accuracy by up to 29.8 percentage points. Ablations show complementary benefits from different relations and consistent gains from fusion beyond role organization. Mechanism analysis further finds 15.6-point higher answer-critical source recall and an 11.5-point lower confounder-token share. ROAM remains robust across manager scales.
[NLP-51] SymbolicLight V2: Hybrid Neuromorphic Architecture and Sparse Execution for Low-Energy Language Inference
【速读】： 该论文旨在解决生成式 AI（Generative AI）在低功耗边缘设备上高效推理的问题，特别是在资源受限场景下实现高吞吐量与低能耗的平衡。其核心挑战在于传统密集计算模式在处理长序列时产生的冗余计算和数据移动开销，导致能效瓶颈。解决方案的关键在于提出 SymbolicLight V2 的混合类脑架构：通过稀疏事件计算（sparse event computation）与连续状态处理（continuous-state processing）的协同设计，引入分级有符号事件（graded signed events）及无 Softmax 局部注意力机制，在保持模型表达能力的同时显著减少无效计算与内存访问。具体实现中，利用数字定点算术在 Alveo U50C FPGA 上部署 194M 参数模型，并采用稀疏整数执行优化 ARM CPU 性能；通过主动行权重聚合与有效状态键值（KV）加载等技术，将解码吞吐量从 474.6 提升至 643.2 tokens/s，同时使每生成令牌的总卡级能耗降低 27.6%，并揭示了空闲状态能耗占主导（82.8%），凸显短延迟对能效的正面影响。实验表明，尽管当前版本性能略低于同预算的稠密模型，但该架构通过事件稀疏性有效映射到计算与数据移动的省略，为后续专用 V2 实现提供了可扩展的能量效率提升路径。
链接: https://arxiv.org/abs/2609.09772
作者: Ting Liu
机构: SymbolicLight Research(符号光研究); Foshan, Guangdong, China(广东省佛山市)
类目: Computation and Language (cs.CL)
备注: 22 pages, 8 figures, 11 tables
Abstract:SymbolicLight V2 combines sparse event computation with continuous-state processing in a hybrid neuromorphic language architecture. Extending V1’s spike-gated dual paths, it adds graded signed events at further projections and softmax-free local attention. We implement the 194M-parameter model on an Alveo U50C FPGA using digital fixed-point arithmetic and on an ARM CPU using sparse integer execution. Across three same-checkpoint FPGA implementations at 175 MHz, active-row weight gathering and valid-state KV loading raise decode throughput from 474.6 to 643.2 tokens/s for a 32-token prefix and 128 outputs. Estimated gross card energy falls from 0.06087 to 0.04407 J per generated token, a 27.6% reduction. Complete-request energy, including prefill, falls by 24.4-27.7% across three prefix lengths. An independent idle split attributes 82.8% of gross card energy to loaded idle, explaining the benefit of shorter token latency. Against the recorded RTX 5090 compiled-FP32 baseline, integer FPGA execution uses 89.1% less estimated card energy during short-context decode; arithmetic precisions differ, and the GPU baseline is not the lowest-energy tested configuration. On four Cortex-A76 cores of a ROCK 5T, complete requests reach 65.4 tokens/s at 9.80 W and 0.151 J per generated token at the adapter’s AC input. These results connect event sparsity to omitted computation and data movement. The mechanisms also support other dedicated V2 implementations: increasing throughput by a greater factor than active power lowers energy per generated token. Evaluation holds the deployed checkpoint fixed; its quality trails a same-budget dense control, so the results do not establish equal-quality efficiency.
[NLP-52] Fine-Tuning a KV Cache Concatenation-Aware Model or Recomputing KV Caches? Why Not Both?
【速读】： 该论文旨在解决检索增强生成（Retrieval-Augmented Generation, RAG）系统中因大量检索片段拼接导致输入上下文过长所引发的时延问题，特别是预填充阶段（prefill workload）耗时增加带来的首个词元时间（Time to First Token, TTFT）延长问题。尽管已有方法通过复用预计算的键值缓存（Key-Value Cache, KV cache）有效降低长上下文输入下的TTFT，但其对生成质量的影响尚不明确。本文提出一种综合解决方案，其关键在于：（i）在微调过程中显式考虑KV缓存拼接对模型行为的影响，使模型适应长上下文输入；（ii）选择性地重新计算部分KV缓存，以在保持响应质量的同时减少冗余计算。实验结果表明，该方法在RULER基准测试中，针对124k token的输入，相比仅重新计算全部KV缓存的基线，RULER得分提升9.7分，同时相较全注意力机制将TTFT降低了80%。
链接: https://arxiv.org/abs/2609.09768
作者: Fumihiko Tachibana,Daisuke Miyashita,Jun Deguchi
机构: Kioxia Corporation(铠侠公司)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:In Retrieval-Augmented Generation (RAG) systems, a large number of retrieved chunks are concatenated to form the input context so that users can receive high-quality responses based on external knowledge. As a result, the input context length increases substantially, leading to a larger prefill workload and, in turn, a longer time to first token (TTFT). While previous works that reuse precomputed key-value (KV) caches effectively reduce TTFT for long-context inputs, it remains unclear whether response quality is preserved when the input context becomes very long. In this paper, we propose a combined approach that (i) fine-tunes the model while taking KV cache concatenation into account and (ii) selectively recomputes a subset of the KV caches. By applying both techniques, we demonstrate improved accuracy for long-context inputs. Experiments on the RULER benchmark show that, for a 124k-token input, our method improves the RULER score by 9.7 point over the baseline that recomputes KV caches only. Moreover, TTFT is reduced by 80% compared with full attention.
[NLP-53] CARRE: Counterfactual Action Retrieval and Reason Evaluation for Explainable Churn Prescription
【速读】： 该论文旨在解决传统流失预测模型（churn model）仅能识别高风险客户，却无法明确推荐具体可行的保留措施及其合理性的问题。其核心解决方案是提出CARRE（Counterfactual Action Retrieval and Reason Evaluation）框架，该框架包含三个关键阶段：基于检索增强的候选保留策略生成、考虑成本的反事实评分，以及利用大语言模型（LLM）进行结构化流失原因诊断与个性化解释生成。通过在IBM电信客户流失数据集上的实验验证，CARRE在313个高风险测试样本中相较原始SHAP基线实现79.8%的平均预测流失风险降低，较成本控制版SHAP+Cost基线提升80.4%，且成本归一化效率高出10.5%；同时，在136例按原因分层的评估样本上，基于诊断驱动的提示优化使弱标签一致性从79.4%提升至90.4%，表明推理质量显著改善。此外，由跨厂商LLM判别器对135条生成解释进行评分，平均得分介于4.02至5.00之间（满分5），且经确定性审计未发现66个可验证用户画像陈述之间的逻辑矛盾。消融实验进一步表明，检索顶数k=5在候选覆盖度与下游推理一致性之间取得最佳平衡。研究结果表明，检索、基于模型的反事实评分与语言生成可在原型流失处方流水线中分离设计并协同评估，为可解释、可操作的客户保留决策提供了系统性方法。
链接: https://arxiv.org/abs/2609.09766
作者: MinJoo Kim,SanJin Park,SeungHwan Cho
机构: Hanyang University (汉阳大学)
类目: Computation and Language (cs.CL)
备注: 14pages, 1 figure, Accepted at Workshop on 5th End-to-End Customer Journey Optimization at the International Conference on Knowledge Discovery and Data Mining
Abstract:Churn models typically identify high-risk customers but do not specify which feasible retention action should be considered or why that action is appropriate. We present CARRE (Counterfactual Action Retrieval and Reason Evaluation), a three-stage framework that combines retrieval-augmented candidate generation, cost-aware counterfactual scoring, and large language model (LLM) reasoning. CARRE retrieves a predefined catalog of retention actions, estimates model-predicted churn-risk changes under explicit feature transformations, and generates a structured churn reason and a profile-grounded explanation for the selected action. On the IBM Telco Customer Churn dataset, CARRE achieves 79.8% greater mean model-predicted risk reduction than the plain SHAP baseline and 80.4% greater reduction than the cost-controlled SHAP+Cost baseline across 313 high-risk test cases; its cost-normalized efficiency is 10.5% higher than that of plain SHAP. On a 136-case reason-stratified evaluation sample, diagnosis-driven prompt refinement increases weak-label agreement from 79.4% to 90.4%, with no auxiliary-plan constraint violations; because the same sample was used for error diagnosis and re-evaluation, the post-refinement result is not an independent estimate of generalization. For 135 explanations generated using the pre-refinement v2 reason outputs, two cross-vendor LLM judges assign mean scores ranging from 4.02 to 5.00 out of 5, although one judge saturates on actionability, and a deterministic audit finds no contradictions among 66 verifiable profile claims. Retrieval ablations show that k=5 provides the best evaluated compromise between high candidate coverage and downstream reasoning agreement in this dataset. These results illustrate how retrieval, model-based counterfactual scoring, and language generation can be separated and jointly evaluated in a prototype churn-prescription pipeline.
[NLP-54] SocialRL: Refining LLM s Social Intelligence through Multi-turn Reinforcement Learning and Reward Design
【速读】： 该论文旨在解决当前生成式对话系统在多轮交互中因采用单轮强化学习与稀疏结果奖励而导致的短视策略问题，即难以有效平衡任务目标达成与人际关系维护之间的矛盾。其核心解决方案是提出SocialRL——一种面向多轮对话的强化学习框架，关键在于：首先，采用基于PPO的多轮强化学习机制，将延迟的最终结果奖励回传至每一对话回合，实现长程规划能力；其次，设计包含目标推进、关系契合度、语境连贯性等六个维度的过程奖励体系，并引入动态生成的细粒度评分标准及阶段感知的权重调度策略，分别在对话初期侧重关系建立、中期注重目标进展、后期实现目标与关系的均衡收尾。实验表明，SocialRL在多个社交对话基准上相较基线模型平均提升9.2个百分点的目标达成率，验证了其在合成与真实社交场景、常规与挑战性任务中的有效性。
链接: https://arxiv.org/abs/2609.09764
作者: Jianing Wang,Xintao Wang,Aili Chen,Jie Shi,Hongcheng Guo,Jun Gao,Wenxuan Zhao,Chengkun Lang,Yuanli Guo,Yanghua Xiao
机构: Fudan University (复旦大学); Hello Group
类目: Computation and Language (cs.CL)
备注: 30pages 2figures
Abstract:Social intelligence enables agents to read social context, infer intent, and adapt over sustained dialogue. As language models become autonomous collaborators, it is central to building effective and trustworthy human-AI interaction. Existing reinforcement learning methods optimize single-turn utterances and sparse outcome rewards, producing short-sighted policies that struggle to manage goal-relationship tensions across multi-turn interactions. We propose SocialRL, a multi-turn reinforcement learning framework addressing both challenges. First, we apply multi-turn reinforcement learning using PPO that propagates delayed outcome rewards back to each turn, enabling long-horizon planning. Second, we design six process reward dimensions capturing the goal-relationship trade-off, including goal advancement, relational attunement, contextual coherence, etc. A reward model dynamically generates fine-grained scoring criteria for each dimension, while a stage-aware weight schedule prioritizes relationship-building in early turns, goal advancement mid-way, and balanced closure late. Across multiple social-dialogue benchmarks, SocialRL improves Goal Achievement by an average of 9.2 percentage points over the corresponding Base models. These results demonstrate the effectiveness of SocialRL across synthetic and real social scenes, as well as standard and challenging social scenarios.
[NLP-55] Can Artificial Intelligence Support Healthcare and Mental Health Through Early Cyberbullying Detection ? The Impact of Emotion-Aware AI on Proactive Online Safety
【速读】： 该论文旨在解决网络欺凌（cyberbullying）及相关有害在线互动对医疗系统、心理健康及公众福祉日益加剧的负面影响问题。其核心挑战在于如何在保障计算效率的前提下，实现对敏感内容的高精度、上下文感知的早期识别与预警。解决方案的关键在于提出CareGuard框架，通过融合零样本语义标注（zero-shot semantic labeling）与微调后的Transformer模型（如BERT、DistilBERT和RoBERTa），实现跨类别敏感内容的鲁棒分类；同时引入情感感知过滤机制与基于余弦相似度的语义筛选策略，优先处理语义相关且情绪显著的内容，从而在医疗导向的监测场景中显著降低无效计算开销，兼顾检测准确性与系统运行效率。实验结果表明，CareGuard在基准数据集上实现了良好的性能平衡，具备在心理健康监测与在线安全应用中规模化部署的潜力。
链接: https://arxiv.org/abs/2609.09735
作者: Hamed Jelodar,Amir Firouzi,Yen-Wu Lo,Maryam Tanha,Sajjad Dadkhah
机构: University of New Brunswick(新不伦瑞克大学); Northeastern University(东北大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Healthcare systems, mental health, and public well-being are increasingly affected by cyberbullying and harmful online interactions. This paper presents CareGuard, an early-warning framework designed to support healthcare-driven mental health protection and proactive online safety through the detection of cyberbullying-related content using advanced natural language processing techniques. CareGuard integrates zero-shot semantic labeling with fine-tuned transformer-based models, including BERT, DistilBERT, and RoBERTa, to enable robust and context-aware classification across sensitive cyberbullying categories. To improve efficiency and reduce unnecessary computation in healthcare-oriented monitoring settings, the framework incorporates an emotion-aware filtering mechanism alongside cosine similarity-based semantic screening, allowing the system to focus on semantically relevant and emotionally salient content. Experimental results on benchmark datasets demonstrate that CareGuard effectively balances detection accuracy and computational efficiency, highlighting its potential for scalable deployment in healthcare systems, mental health monitoring, and online safety applications.
[NLP-56] StreamAlign: Streaming Text-Aligned Speech Tokenization EMNLP2026
【速读】： 该论文旨在解决现有文本对齐语音分词方法在实时流式处理与音素粒度保持方面的关键问题。传统方法依赖离线自动语音识别（ASR），导致两大局限：一是必须等待完整语句才能进行分词，无法支持实时流式处理；二是ASR与大语言模型（LLM）之间存在词汇不匹配，使语音分词粒度从子词降级至词级别，损失了声学细节。为此，论文提出StreamAlign框架，其核心在于通过结合字符级RNN-Transducer对齐与词级ASR引导实现在线语音-文本对齐，在缓解ASR-LLM词汇不匹配的同时保持高识别准确率。该方案引入主动词边界分类器，提前预测词边界，将分词延迟从560 ms降低至270 ms，显著提升实时性。在LibriSpeech数据集上，StreamAlign实现了最低的词错误率（WER）和最高的统一语音质量评分（UTMOS）。进一步地，基于StreamAlign单元训练的流式语音语言模型（StreamAlign-SLM）在语音续写任务中表现优于其他端到端语音语言模型，并在SALMon和语音StoryCloze任务中展现出最强的整体一致性。
链接: https://arxiv.org/abs/2609.09719
作者: Kang-wook Kim,Jinyoung Park,Jinsoo Kim,Sehun Lee,Sang Hoon Woo,Gunhee Kim
机构: Seoul National University (首尔国立大学); University of California, Berkeley (加州大学伯克利分校); KRAFTON; Georgia Institute of Technology (佐治亚理工学院)
类目: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注: Findings of EMNLP 2026. Project page: this https URL
Abstract:Text-aligned speech tokenization methods have emerged to better align speech tokens with LLM token spaces, enabling more effective utilization of pretrained LLMs. However, they rely on offline automatic speech recognition (ASR), leading to two key limitations: (i) the need for complete utterances before tokenization, precluding real-time streaming, and (ii) vocabulary mismatch between ASR and LLMs, which reduces acoustic granularity from the subword to the word level. We introduce StreamAlign, a text-aligned speech tokenization framework that enables streaming tokenization for real-time speech-text joint modeling. StreamAlign performs online speech-text alignment by combining character-level RNN-Transducer alignment with word-level ASR guidance, mitigating ASR-LLM vocabulary mismatch while preserving recognition accuracy. A proactive word boundary classifier anticipates word completion at chunk boundaries, reducing tokenization latency from 560 ms to 270 ms. On LibriSpeech, StreamAlign achieves the lowest WER and highest UTMOS among evaluated tokenizers. Furthermore, StreamAlign-SLM, a spoken language model trained on StreamAlign units, outperforms other end-to-end spoken language models in speech continuation while achieving the strongest overall consistency on SALMon and spoken StoryCloze.
[NLP-57] Scaling E-Commerce Attribute Extraction with Parallel Decoding
【速读】： 该论文旨在解决电子商务目录中产品属性信息杂乱无章、缺乏结构化的问题，导致难以识别对消费者购买决策具有区分意义的关键属性，并实现大规模属性值提取。传统属性值抽取（Attribute Value Extraction, AVE）系统对所有属性一视同仁，生成冗长且不一致的属性集合，无法准确反映消费者用于区分产品的实际因素。为此，本文提出一种两阶段大语言模型（Large Language Model, LLM）流水线：第一阶段通过一个轻量级的LLM发现每个产品类别的紧凑、可排序的购买区分性属性模式（purchase-discriminative attributes），构建高质量的属性架构；第二阶段利用微调后的紧凑型大模型（Qwen3-4B）结合超并行解码（Hyper-Parallel Decoding, HPD）技术，高效精准地从目录文本中提取属性值。该方案在保持85%提取准确率（与基础模型相当）的同时，将推理成本降低92%，实现了生产级部署。其核心创新在于通过分阶段优化，构建了基于类别级别的结构化产品知识库，为下游知识密集型应用提供一致、可比的属性表示。
链接: https://arxiv.org/abs/2609.09716
作者: Nikhita Vedula,Dushyanta Dhyani,Bryan Wang,Shervin Malmasi
机构: Amazon.com, Inc.(亚马逊公司)
类目: Computation and Language (cs.CL)
备注: Accepted to 11th Workshop on Automated Knowledge Base Construction
Abstract:Customers rely on specific product attributes to compare products and make purchasing decisions, but e-commerce catalogs are messy and unstructured, making it difficult to identify which attributes matter most and extract them at scale. Standard Attribute Value Extraction (AVE) systems treat all attributes equally, producing large, inconsistent attribute sets that do not reflect the factors consumers use to differentiate products. We introduce a two-stage LLM pipeline that first discovers a compact, ranked schema of purchase-discriminative attributes for each product category, then extracts their values from catalog text using a fine-tuned compact LLM (Qwen3-4B) with Hyper-Parallel Decoding (HPD). This pipeline achieves 85% extraction accuracy, on par with the foundational LLM it was distilled from, while reducing inference costs by 92% over foundational LLMs, enabling production-scale use for product discovery and catalog enrichment. The resulting category-level structured representations effectively constitute automatically constructed product knowledge bases, providing consistent, comparable attributes across varied product categories that can ground downstream knowledge-intensive applications.
[NLP-58] When Auditors Fabricate: Batch-Size Degradation and Confident Hallucination in LLM Detection of Planted Document Contamination
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在文档质量审计中作为自动检测工具的可靠性问题，特别是其对人为植入错误（planted errors）的识别能力。研究发现，尽管在小规模任务中模型表现尚可，但在大规模处理时检测性能急剧下降，且失败模式并非简单回避，而是产生虚假报告——即模型会生成虚构的、符合语境但实际不存在的“错误”（如“心灵感应松鼠”、“量子驱动烤面包机”），从而导致误导性结论。关键问题是：当前大语言模型在批量处理文档时无法保持一致性与真实性，其错误检测能力随输入规模增加而退化，并以制造虚假证据的形式表现，而非拒绝响应。因此，解决方案的关键在于实施严格的系统约束：限制批处理规模、直接向源文本注入验证内容，并对每一项报告结果进行机械式比对验证，以确保审计过程的可信赖性与可追溯性。
链接: https://arxiv.org/abs/2609.09696
作者: Karan Parekh,Sanjana Pendyala Ravinder,Sana Mhapsekar,Medina Maloku
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 8 pages, 3 tables. Preprint also deposited at Zenodo, doi: https://doi.org/10.5281/zenodo.21939088
Abstract:Large language models are increasingly proposed as automated auditors of document quality, yet their reliability as detectors of planted errors is poorly characterised. We construct a contaminated corpus of 150 academic papers spanning supply chain management and medical research, injecting 450 known contaminants of three types: typographical corruption, semantic reversal, and absurd out-of-context insertion. We then evaluate Google Gemini 3.0 Pro’s ability to recover a 180-contaminant answer-key subset across 60 documents under three prompting regimes of increasing scale: single document, small batch, and large batch. Detection holds at small scale and then collapses: 50% recovery on single documents, 60% on small batches, and 2.8% on large batches. The failure mode at scale is not abstention but fabrication. Rather than reporting incomplete processing, the model produced confident findings including invented contaminants of its own, absurdities such as “telepathic squirrel” and “quantum-powered toaster” that mimic the style of the planted material but do not appear in any document. Detection also varies by contamination type: absurd insertions were recovered at 75% in completed evaluations, while semantic reversals and typographical corruptions were each recovered at only 50%. The corruptions most likely to occur in the wild, plausible ones, are the ones most often missed. We conclude that LLM document auditing degrades not gracefully but deceptively, and outline the harness such systems require: bounded batch sizes, direct content injection, and mechanical verification of every reported finding against source text.
[NLP-59] Looped GPT -BERT: Trading Parameters for Computation in Small Language Modeling
【速读】： 该论文旨在解决在训练数据有限的情况下，如何在不依赖大规模参数量的前提下提升语言模型性能的问题。其核心挑战在于如何在参数资源受限的场景下实现与大模型相当甚至更优的泛化能力与任务表现。解决方案的关键在于提出一种基于深度参数共享（depth-wise parameter sharing）的循环式架构——Looped GPT-BERT，通过在少量物理层上进行多次循环遍历（recurrent traversals），以重复利用少量参数来模拟深层网络的表达能力。实验表明，该方法在仅使用12.18M参数的4×12模型中，实现了与更大规模基线模型（如GPT-2、GPT-BERT）相当的性能，尤其在BLiMP和GLUE等语言学及下游任务上表现优异。然而，消融实验也揭示了该设计的内在局限性：尽管循环计算有助于提升训练效果并保持对特定语言任务的强表现力，但受限于物理层数较少，导致模型的表示空间受限，从而在部分任务上出现性能下降。因此，该方案的核心创新在于通过参数复用与循环结构，在低参数量条件下有效扩展模型的“计算深度”，但其有效性受制于表示能力的边界。
链接: https://arxiv.org/abs/2609.09691
作者: Tingshuo Fan,Hongtao Mu,Tianyu Zhou,Hansen Liu,Tao Ji
机构: Fudan University (复旦大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 10 pages, 2 figures
Abstract:When training data are limited, increasing parameter count is not the only way to improve language-model performance. A small parameter set, when repeatedly applied, can also deliver comparable performance. We study Looped GPT-BERT in the BabyLM 2026 Strict-small setting, combining GPT-BERT’s masked next-token and causal language-modeling objectives with depth-wise parameter sharing. We train on a preprocessed 7.48M-word English corpus and compare objective ratios, non-looped and looped architectures, and loop counts. Our final 4\times12 model uses four physical layers for twelve recurrent traversals and contains 12.18M parameters. The BabyLM 2026 leaderboard reports an Overall Average of 35.42 and an NLP Average of 48.48. Compared with public BabyLM 10M Strict-small GPT-2 and GPT-BERT baselines, it achieves comparable performance on selected linguistic and downstream metrics, including BLiMP and GLUE, with fewer parameters. The loop ablations show that additional recurrent computation can improve training and preserve strong performance on selected linguistic tasks, whereas poorer performance on other tasks may reveal an inherent limitation of the looped design: using only a few physical layers restricts the model’s representational space.
[NLP-60] Which Medical Questions Deserve Rationales? Perturbation-Sensitive Selection for Robust QA
【速读】： 该论文旨在解决在有限标注预算下，如何从已标注的医学问答数据中高效选择最具价值的样本进行推理过程（rationale）监督的问题。传统方法关注于哪些问题应被标注，而本文聚焦于在固定令牌预算约束下，应为哪些已有答案标签的问题分配额外的高质量推理链标注资源。其核心解决方案是提出一种基于均方根鲁棒性采样优先级（Root-Mean-Square Robustness-based Sample Prioritization, RMS-RSP）的方法，该方法仅对推理链中的标记（rationale tokens）施加隐藏状态扰动，并通过衡量真实答案与最佳干扰项之间的置信度差距变化来评估样本的重要性。实验结果表明，尽管在原始设置下RMS-RSP未显著提升平均准确率（仅在AfriMed-QA上取得1.44个百分点的统计显著提升），但在对答案选项进行三次重排后，其在鲁棒准确率和语义一致性方面分别平均提升了1.91和2.85个百分点，且在所有五个数据集上均呈一致改善趋势。这表明，基于推理链局部边界敏感性的筛选策略虽不能普遍提高模型精度，但能有效识别出有助于提升模型对语义等价格式变换不变性的监督信号。
链接: https://arxiv.org/abs/2609.09684
作者: Yuexin Wu,Dayou Yu,Vasile Rus
机构: University of Memphis(孟菲斯大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Medical question-answering datasets often contain answer labels, whereas high-quality rationales remain scarce, noisy, or costly to validate. This changes the acquisition question: rather than asking which questions should be labeled, we ask which already-labeled questions should receive rationale supervision under a fixed token budget. We study an offline version of this problem in which candidate rationales are visible to the selector but withheld from downstream training unless selected. We propose root-mean-square Robustness-based Sample Prioritization (RMS-RSP), which perturbs hidden states only at rationale tokens and measures the resulting shift in the gold-versus-best-distractor margin. Across five medical QA datasets, MedGemma-4B-IT, three training seeds, ten budgeted non-RSP selectors, and an unbudgeted full-supervision reference, RMS-RSP provides a deliberately qualified result. Its locked-budget accuracy is 60.61% on average versus 60.08% for Random, with a statistically resolved gain only on AfriMed-QA (+1.44 points). Its full-budget accuracy area is not better than Random. However, after three answer-option reorderings, RMS-RSP improves robust accuracy and semantic consistency by 1.91 and 2.85 points on average, respectively, with the same direction on all five datasets. Training on every pool rationale raises macro accuracy to 63.74%, but consumes 29–254 times more rationale tokens and does not uniformly improve robustness. These findings do not establish universal accuracy gains; they instead suggest that rationale-local boundary sensitivity can identify supervision that improves invariance to semantically equivalent formatting changes.
[NLP-61] X2-NativeCursor: Native-Token Text Progress Tracking for Incremental-Text Streaming Codec TTS
【速读】： 该论文旨在解决增量式文本流式语音合成（Incremental-text streaming text-to-speech, TTS）中语音进度追踪的难题，特别是在需要同步高亮、中断处理和对话历史更新的应用场景下。传统方法依赖完整的音频输出或在流式过程中引入额外的声学处理以实现对齐，存在延迟高或计算开销大的问题。其核心解决方案是提出X2-NativeCursor，一种轻量级的进度观测器，能够在不修改TTS生成器的前提下，仅基于解码前的原生语音标记（native speech tokens）进行实时进度跟踪。该方法通过归一化策略将已发声的标签与原始文本片段精确映射，并结合文本编码器与原生标记编码器构建局部匹配模块，以估计当前语音位置；再通过独立的输出规则将可修正的位置估计转化为永不回退的光标（cursor）。实验表明，在80毫秒前瞻条件下，该方法的均方绝对误差仅为0.151个汉字，显著优于采用320毫秒前瞻的在线波形基线（1.253字符），且对齐的实时因子从0.3598降低至0.0180。在更严格的自动对齐参考下，该方法仍保持较低的追踪误差，验证了其鲁棒性。此外，该方法成功适配于Qwen3-TTS与CosyVoice2两种不同骨干网络，仅需为每个模型训练独立的观测器即可实现良好性能，代码已公开。
链接: https://arxiv.org/abs/2609.09677
作者: Zehan Liu,Carl Chen,Rime Wen,Kaiqi Fu,Altman Lin,Shawn Qin,Lights Shi,Roy Gan,Hao Wang,Qian Wang
机构: X Square Robot
类目: Computation and Language (cs.CL)
备注:
Abstract:Incremental-text streaming text-to-speech (TTS) needs online text progress tracking for synchronized highlighting, interruption handling, and dialogue-history updates. Input text arrives before it is spoken, so text arrival alone cannot indicate speech progress. Existing waveform-based alignment requires complete audio or adds acoustic processing during streaming. We propose X2-NativeCursor, a lightweight observer that tracks progress from native speech tokens before waveform decoding without changing the TTS generator. Its normalization plan links spoken labels to their original-text spans. Text and native-token encoders feed a local matcher that estimates the current label position. A separate output rule converts revisable position estimates into a cursor that never moves backward. Mean absolute error against an automatic reference is 0.151 Chinese characters with 80-ms lookahead, versus 1.253 characters with 320-ms lookahead for an online waveform baseline. Alignment real-time factor also decreases from 0.3598 to 0.0180 relative to this baseline. Lower tracking error is retained under a second automatic alignment reference. We evaluate X2-NativeCursor on Qwen3-TTS and validate its adaptation to CosyVoice2 by training a separate observer for each backbone. Code is publicly available at this https URL.
[NLP-62] SEA-SpeechBench: A Large-Scale Multitask Benchmark for Speech Understanding Across Southeast Asia EMNLP2026
【速读】： 该论文旨在解决当前语音理解评估框架严重依赖英语、导致东南亚（SEA）语言普遍被忽视的问题。其核心挑战在于缺乏一个大规模、多任务、覆盖多种语音理解能力的基准测试，尤其在低资源语言上的表现评估不足。为此，研究者提出了SEA-SpeechBench，据我们所知，这是首个面向11种东南亚语言的大规模多任务评估基准，包含97,194个样本、99个评估集及597小时精心筛选的音频数据。该基准涵盖9项多样化任务，分为三类：语音处理（自动语音识别、语音翻译、口语问答）、副语言分析（情绪、性别、年龄、说话人识别）以及一项新提出的时序理解维度——支持长达3分钟音频序列中的时间戳内容查询与时间定位。关键解决方案在于采用原生东南亚语言与英语双语提示（multilingual prompting），以真实反映用户与语音-语言模型交互的场景。实验表明，主流开源与专有系统在时序理解、情绪识别和语音翻译等任务上表现均不理想，且在低资源语言如缅甸语和泰米尔语中，性能较英语下降高达41个百分点。这一发现揭示了现有模型在跨语言泛化与复杂语义理解方面的显著局限性，强调了推动包容性模型开发的紧迫性。
链接: https://arxiv.org/abs/2609.09672
作者: Jingyi Liao,Wenyu Zhang,Zhuohan Liu,Yingxu He,Geyu Lin,Xunlong Zou,Shuo Sun,Syed Ali Redha Alsagoff,Ai Ti Aw
机构: Institute of Advanced Intelligence and Computing, A*STAR(新加坡科技研究局先进智能计算研究所); Nanyang Technological University(南洋理工大学); Center for AI Safety(人工智能安全中心)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026
Abstract:The rapid advancement of audio and multimodal large language models has unlocked transformative speech understanding capabilities, yet evaluation frameworks remain predominantly English-centric, leaving Southeast Asian (SEA) languages critically underrepresented. We introduce SEA-SpeechBench, to the best of our knowledge, the first large-scale multitask benchmark that evaluates speech understanding in 11 SEA languages through 97,194 samples across 99 evaluation sets and 597 hours of curated audio data. Our benchmark comprises 9 diverse tasks across 3 categories: speech processing (automatic speech recognition, speech translation, spoken question answering), paralinguistic analysis (emotion, gender, age, speaker recognition), and temporal understanding, a novel dimension featuring timestamped content queries and temporal localization within extended audio sequences up to 3 minutes. We implement multilingual prompting in both native SEA languages and English to reflect user interactions with audio-language models. Evaluation of leading open-source and proprietary systems reveals marked performance gaps. Across all models, performance remains underwhelming on temporal understanding, emotion recognition, and speech translation. Prompting in low-resource languages such as Burmese and Tamil lags behind English by up to 41 percentage points. Our findings expose critical model limitations and underscore the need for inclusive model development. The SEA-SpeechBench benchmark is available at this https URL.
[NLP-63] PELM: Power Efficient On-Device LLM Inference with Speculative Decoding and Dynamic Voltage Frequency Scaling
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在资源受限的移动与边缘设备上进行本地部署时面临的高计算负载、功耗及散热挑战。现有方法虽采用动态电压频率调节（Dynamic Voltage and Frequency Scaling, DVFS）等硬件级调控策略以降低能耗与发热，但在热约束严苛场景下仍表现不足。其核心问题在于传统方法仅依赖单一维度的频率调优，未能充分挖掘推理过程中的能效优化空间。本文提出PELM（Power-Efficient LLM Inference Management），通过引入两个面向任务特性的新型控制维度——推测解码（speculative decoding）与可变验证深度（variable verification depth），扩展了优化维度，使系统能够在保证生成质量的前提下动态调整计算开销。该方案利用生成式AI（Generative AI）中非所有令牌均需全深度推理的关键洞察，实现更精细的能效管理。实验表明，在多种硬件平台与数据集上的评估中，PELM相较当前最优功率管理方法实现了最高达23.1%的加速比和52.4%的能量消耗降低，同时维持相近的任务性能表现。
链接: https://arxiv.org/abs/2609.09662
作者: Weisi Yang,Stephen Xia
机构: Northwestern University (西北大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Operating Systems (cs.OS); Performance (cs.PF)
备注: Accepted to ACM/IEEE SenSys’26
Abstract:Deploying Large Language Models (LLMs) directly on mobile platforms at the edge is gaining traction due to a myriad of benefits, such as increased privacy, personalization, and reduced latency. However, LLMs have heavy computational requirements, which are difficult for resource-constrained mobile and edge platforms to fulfill. In addition to limited compute resources, mobile and edge systems often have a compact form factor and lack physical mechanisms to dissipate heat generated from high processor usage rates (e.g., fans) to prevent throttling and reduced processing power, which LLMs can easily cause. To mitigate these effects, prior works have proposed various power governing strategies, such as dynamic voltage and frequency scaling (DVFS), for reducing power and heat generation for heavy computational tasks on mobile platforms. Recently, DVFS methods tailored for mobile LLMs have also been proposed. However, these methods mostly focus on optimizing hardware parameters and processor frequencies, and they fall short under some thermally constrained scenarios. Drawing from recent advances in machine learning, we identify and take advantage of the key insight that not all tokens require full-depth inference to maintain high-quality generation. Motivated by this, we present PELM, a solution that augments traditional DVFS processor frequency tuning with two additional workload-specific knobs: 1) speculative decoding and 2) variable verification depth to expand the optimization space to multiple dimensions for more power efficient on-device LLM inference. In extensive evaluations across hardware platforms and datasets, PELM demonstrates superior performance compared to state-of-the-art power governing methods, with up to 23.1% speedup and 52.4% reduction in energy consumption, while maintaining comparable task performance. The source code is available at this https URL.
[NLP-64] CityPlanner: A Sandbox Agent for Executable Urban Planning EMNLP
【速读】： 该论文旨在解决城市规划这一现实世界中的空间优化问题，其核心挑战在于如何在大规模候选方案中，基于成本、服务品质等实际目标，选择可行的行动策略。现有优化与强化学习方法虽在固定任务形式下表现有效，但普遍依赖于特定任务的表示方式和约束处理机制，缺乏通用性与可执行性。为此，论文提出 \emphCityPlanner——一个面向可执行城市规划的沙盒-智能体框架。其关键创新在于引入 \emphUrbanSandbox，一个统一的基于文件的环境，使智能体能够读取任务文件、生成规划、运行评估器，并根据可执行反馈迭代修正决策。为提升学习效率，进一步提出原子任务强化学习（atomic-task reinforcement learning），将长周期沙盒轨迹分解为两个阶段：\emphBuildPlan（初始建设）与 \emphImprovePlan（基于反馈的优化），实现分步规划与渐进改进。实验表明，CityPlanner在真实世界基准上显著优于启发式方法、任务专用强化学习及通用大语言模型（LLM）智能体基线；消融实验验证了 UrbanSandbox、原子任务强化学习以及迭代部署机制的关键贡献。
链接: https://arxiv.org/abs/2609.09578
作者: Wentao Zhang,Jingyuan Wang,Zetong Zhou,Yifan Yang,Wenrui Wang
机构: Beihang University (北京航空航天大学); MIIT Key Laboratory of Data and Decision Intelligence (工业和信息化部数据与决策智能重点实验室)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: EMNLP Under Review
Abstract:Urban planning is a real-world spatial optimization problem that requires selecting feasible actions from large candidate spaces under practical objectives such as cost and service quality. Existing optimization and reinforcement learning methods are effective for fixed formulations, but often depend on task-specific representations and constraint handling. We propose \emphCityPlanner, a sandbox-agent framework for executable urban planning. CityPlanner introduces \emphUrbanSandbox, a unified file-based environment where agents inspect task files, generate plans, run evaluators, and revise decisions based on executable feedback. To make learning tractable, we further propose atomic-task reinforcement learning, which decomposes long sandbox trajectories into \emphBuildPlan for initial construction and \emphImprovePlan for feedback-based refinement. Experiments on a real-world benchmark show that CityPlanner consistently outperforms heuristic, task-specific RL, and general LLM-agent baselines. Ablations verify the contributions of UrbanSandbox, atomic-task RL, and iterative deployment. We release the code and dataset at this https URL
[NLP-65] Reproducing Omitted Temporal Expressions in Japanese News for Retrieval-Augmented Applications EMNLP2026
【速读】： 该论文旨在解决日文新闻文章中常出现的时间表达省略（如仅提及月份或日期）在独立文本索引与检索增强生成（RAG）系统中引发的时间错位及大语言模型（LLM）解释不稳定的难题。其核心解决方案是基于发布日期作为外部上下文，通过规则驱动的流水线jaROTE，对省略的时间表达进行具体化还原为明确的日期或时间区间，从而实现基于发布时间的时序归一化。该方法结合了成熟的时间表达提取与归一化技术，并依据对日文新闻的详尽人工分析进行设计，在两个新闻语料库上的实验表明，jaROTE不仅性能优异，且在保持高效、低成本的同时，表现可与大语言模型相媲美。进一步实验证明，该时间还原机制显著提升了受限于时间条件的词汇检索效果，验证了以发布日期为基准的时序归一化在日文新闻检索中的实际价值。
链接: https://arxiv.org/abs/2609.09569
作者: Tomoaki Yasuda,Shotaro Ishihara
机构: Kagawa University(香川大学); Nikkei Inc.(日经公司)
类目: Computation and Language (cs.CL)
备注: EMNLP 2026 Industry Track
Abstract:News articles often contain omitted temporal expressions, such as day-only or month-only mentions, which must be interpreted with reference to the publication date. When such articles are indexed or processed as standalone text in search and retrieval-augmented generation (RAG) systems, these omissions can cause temporal mismatches and unstable interpretation by large language models. We focus on reproducing omitted temporal expressions as concrete dates or intervals using the publication date as external context before the articles are indexed for search and RAG applications. Specifically, building on established temporal-expression extraction and normalization techniques and informed by a manual analysis of Japanese news articles, we propose jaROTE, a rule-based pipeline for Japanese news. Experiments on two news corpora demonstrate that jaROTE achieves high performance, and remains competitive with LLMs while providing a fast, low-cost pipeline. We further show that temporal reproduction improves time-constrained lexical retrieval, demonstrating the practical value of publication-date-grounded normalization for Japanese news retrieval.
[NLP-66] owards Automatic Evolution Tree Generation from Citation Graphs
【速读】： 该论文旨在解决当前生成式AI（Generative AI）领域中方法演化谱系自动构建面临的两大核心挑战：一是传统综述类研究在面对爆炸式增长的论文发表速度时难以有效扩展；二是现有分类体系归纳方法普遍存在“仅关注末端节点”（leaf-bound）和“忽略时间维度”（time-agnostic）的问题，导致过渡性论文被错误归入成熟类别，并引发祖先与后代节点之间的拓扑倒置。其解决方案的关键在于提出EvoTree——一种分阶段的框架，通过解耦概念骨架学习与时间精调过程实现更准确的演化建模：首先利用基于分布的层次聚类与图感知编码器构建稳定且结构合理的分类骨架；随后在单调路径约束下进行时间敏感的微调，将边缘论文精准重新锚定至内部节点；最后通过大语言模型（LLM）对概念进行语义标注而不改变整体拓扑结构。该方法在首个涵盖11个AI子领域的标注基准上表现卓越，不仅在标准化互信息（NMI）和引文方向准确性方面超越所有基线模型，还在概念纯度与边缘论文识别能力上展现出显著优势，是唯一具备实质性边缘论文检测能力的方法。
链接: https://arxiv.org/abs/2609.09561
作者: Zexing Zhao,Yuntong Hu,Liang Zhao
机构: Georgia Institute of Technology(佐治亚理工学院); Emory University(埃默里大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Surveys remain the primary way researchers grasp the lineage of methods within an AI subfield, but they scale poorly against the current rate of publication. Existing taxonomy-induction methods are largely leaf-bound and time-agnostic; they tend to force transitional papers into mature leaves and can create topological inversions between ancestors and descendants. We propose EvoTree, a staged framework that decouples conceptual backbone learning from temporal refinement: a graph-aware encoder with distribution-based hierarchical clustering yields a stable taxonomy backbone; temporal fine-tuning then re-attaches marginal papers to internal nodes under monotonic-path constraints; a final LLM pass labels concepts without altering the topology. We release the first annotated benchmark for this task across 11 AI subfields. EvoTree attains the highest NMI and citation-direction accuracy among all baselines and the best concept purity on the annotated benchmark, and is the only method with non-trivial marginal-paper detection on the annotated set.
[NLP-67] BuzzASR: A Swarm of 100 Monolingual Speech Recognition Models EMNLP2026
【速读】： 该论文旨在解决大规模多语言自动语音识别（ASR）模型在低资源语言上性能不佳的问题。尽管以Whisper为代表的端到端Transformer架构已显著提升ASR性能，但其高度多语言特性导致对训练数据覆盖不足的语言表现欠佳。为此，本文提出了一种大规模语言适配方案——BuzzASR，通过在102种语言（覆盖FLEURS数据集）上进行单语数据微调，显著提升了模型在低资源语言上的识别准确率。其关键解决方案包括：一是大规模应用简单的单语数据微调策略；二是引入更复杂的语言适配方法，结合单语分词器替换与仅文本的增强数据微调。实验表明，BuzzASR在77种语言上优于Whisper-large-v3，平均字符错误率（CER）降低超过2.8倍；在27种语言上达到开源系统最优水平。此外，所提出的分词器替换策略使压缩率（每标记字符数）平均提升3.3倍，最高达21.7倍。研究团队已公开所有模型、代码及详细结果。
链接: https://arxiv.org/abs/2609.09554
作者: Shivam Singh,Aditya Yadavalli,Catherine Arnett,Alex Warstadt
机构: UC San Diego(加州大学圣地亚哥分校); EleutherAI(艾莱瑟人工智能)
类目: Computation and Language (cs.CL)
备注: Accepted at EMNLP 2026. Models: this https URL ; Project page: this https URL
Abstract:We introduce BuzzASR, a collection of language-specialized fine-tuned Whisper models adapted for automatic speech recognition (ASR) in 102 languages. Large end-to-end Transformer-based ASR models such as Whisper have revolutionized ASR, but most prominent models are highly multilingual. As a result, these models often perform poorly on languages less well-represented in their training set. While it has long been known that effective language adaptation can be achieved through simple fine-tuning on monolingual data, this strategy has only been applied to a small number of languages. We massively scale up this simple approach to 102 languages covered in the FLEURS dataset, while also implementing a more complex language adaptation strategy that integrates monolingual tokenizer replacement and data augmentation using text-only fine-tuning. BuzzASR models outperform Whisper-large-v3 on 77 out of 102 languages, reducing character error rates (CER) by a factor of over 2.8 on average. Our models achieve state-of-the-art CER among open-source systems on 27 of 102 languages on the combined FLEURS and Common Voice test set. Our tokenizer replacement strategy yields an average 3.3x improvement in compression rate (characters per token) over Whisper’s multilingual BPE, with gains of up to 21.7x. We release all models, code, and detailed results: this https URL
[NLP-68] EFM: Token-Efficient Faithful Modeling for Structured Data
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在关键领域应用中的两大核心挑战：令牌效率（token efficiency）与忠实性（faithfulness）。为协同应对这两项约束，作者提出TEFM（Token-Efficient Faithful Modeling）框架，专用于关键领域中的结构化数据解析。其解决方案的关键在于：通过将冗长的结构化观测数据压缩为紧凑的“行为代码”（Behavioral Code）令牌，显著降低令牌消耗并实现极小的信息损失；同时，引入双保真度目标函数，联合优化代码级重构与预测级忠实性，从而识别出基于输入数据的最小充分特征子集，确保推理过程的可解释性与可靠性。实验结果表明，TEFM在多个领域数据集及不同模型架构（Qwen3、Gemma-2、Phi-4）上均实现了与现有方法相当的分类准确率，同时在临床和安全领域分别仅保留约1%和2%的令牌，且生成的解释具有高度忠实性。
链接: https://arxiv.org/abs/2609.09552
作者: Zhichao Hou,Lingdao Sha,Xueyu Mao,Yang Liu,Peijie Qiu,Rui Song
机构: North Carolina State University(北卡罗来纳州立大学); Amazon Web Services(亚马逊网络服务)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:In this paper, we solve two fundamental obstacles in applying LLMs to critical domains: token efficiency and faithfulness. To address both constraints jointly, we present TEFM (Token-Efficient Faithful Modeling), a framework designed for structured data analysis in critical domains. TEFM achieves token efficiency by compressing lengthy structured observations into compact Behavioral Code tokens, dramatically reducing token consumption with minimal information loss. Moreover, TEFM enables faithful rationalization through a dual-fidelity objective that jointly optimizes code-level reconstruction and prediction-level fidelity, identifying minimal sufficient feature subsets grounded in input data. Comprehensive experiments across various domain datasets and model backbones (Qwen3, Gemma-2, Phi-4) show that TEFM achieves competitive classification accuracy with dramatic token reduction (approximately 1% token retention in clinical and 2% in security domains) while producing faithful rationales.
[NLP-69] An Efficient and Effective Agent ic Group Shilling Attack on Recommender Systems ICDM2026
【速读】： 该论文旨在解决推荐系统在面对洗白攻击（shilling attacks）时的脆弱性问题，尤其是现有攻击方法因依赖特定目标微调或固定伪造用户模板，导致适应性差、易被检测的缺陷。其解决方案的关键在于提出一种名为代理群体攻击系统（Agentic Group Attack System, AGAS）的协同式攻击框架：由一个中央协调者（Coordinator）统一调度多个可动态切换角色的工作者代理（worker agents），这些代理能够根据环境反馈自适应地调整策略，在不同受害者系统间灵活切换身份与行为模式，以实现对目标物品的高效推广。该系统通过动态策略优化、角色轮换机制以及对抑制信号的响应能力，显著提升了攻击隐蔽性与效率，同时有效维持了良性推荐质量，并削弱了主流检测器的识别能力。研究结果表明，未来推荐系统的防御需具备应对自适应、协同化攻击的能力，而不仅局限于防范孤立的虚假用户注入。
链接: https://arxiv.org/abs/2609.09551
作者: Quoc Viet Nguyen,Trinh Pham,Viet Huynh,Hongzhi Yin,Quoc Viet Hung Nguyen,Bay Vo,Thanh Tam Nguyen
机构: Griffith University (格里菲斯大学); Edith Cowan University (埃迪斯科文大学); The University of Queensland (昆士兰大学); HUTECH University (胡志明市技术大学)
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL)
备注: Accepted by ICDM 2026
Abstract:Recommender systems have become core infrastructure for modern online platforms, personalizing content at scale and strongly influencing what users see, click on, and purchase. However, this dependence on user interaction also exposes them to shilling attacks, where malicious actors can inject fake profiles to distort item rankings and control visibility. Existing attacks often rely on target-specific fine-tuning or fixed profile templates, making them either difficult to adapt to different victims or easier to detect. To overcome these limitations, we propose the Agentic Group Attack System (AGAS), a coordinated shilling framework where a central Coordinator directs a group of role-switching worker agents to adaptively promote a target item across different victim families. The Coordinator dynamically adjusts the strategy when progress stalls or suppression signals increase, while workers pursue a shared objective and switch between active and inactive roles to avoid repetitive patterns. Under the same attack budgets and evaluation protocols, AGAS consistently surpasses strong baselines in target promotion while better preserving benign recommendation quality, weakening representative detectors, and achieving higher efficiency than prior attacks. These findings also emphasize that defending recommender systems may require mechanisms that can handle adaptive shilling campaigns, not just isolated fake-profile injections. Our code is available at this https URL.
[NLP-70] From Fixed Keys to Readable Schemas: Small Language Models for Vehicle Agent Function Calls
【速读】： 该论文旨在解决车载助手在严格内存与延迟约束下，如何将自然语言请求准确映射为车辆功能调用的问题。核心挑战在于小语言模型（SLMs）在资源受限设备上部署时，函数表面（function surface）的表示方式对模型性能与泛化能力具有决定性影响。解决方案的关键在于对比两种函数表示方法：专用功能标记（Functional Token, FT）与提示中嵌入函数模式（Schema-in-Prompt, SIP）。FT通过固定标记实现紧凑推理，但仅能处理训练时见过的功能，无法泛化；而SIP通过在提示中直接提供函数结构信息，虽增加上下文长度与推理开销，却具备对未见功能的泛化能力。实验基于包含79个车辆功能的9,822条单轮样本数据集，在四个参数量从270M到1.7B的SLM上进行对比，结果表明：在已见功能上，模型规模提升带来的收益有限，0.6B模型表现最佳；而在未见功能上，FT表现归零，而SIP随规模增长显著提升；对于超出范围的请求，FT可能错误调用不存在功能，而SIP则更可靠地拒绝。理论分析进一步揭示了SIP实现泛化的机制及其因长上下文导致的推理成本上升原因。研究结论指出，函数表面的表示方式，而非单纯扩大模型规模，才是决定SLM在车载功能调用任务中能力边界与失效模式的核心因素。
链接: https://arxiv.org/abs/2609.09476
作者: Hamed Jafarzadeh Asl,Yuanhao Yu,Vahid Partovi Nia
机构: Huawei Noah’s Ark Lab, Canada(华为诺亚方舟实验室，加拿大); Polytechnique Montreal (蒙特利尔理工学院)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:In-vehicle assistants must translate natural-language requests into accurate vehicle function calls under strict memory and latency constraints, making small language models (SLMs) attractive for on-device deployment. For such models, a key design choice is how the available function surface is presented. Two approaches are to represent each function with a dedicated Functional Token (FT) or provide function schemas directly in the prompt. FTs enable compact inference but are restricted to functions learned during training, whereas Schema-in-Prompt (SIP) can generalize to unseen functions at the cost of longer prompts and higher inference overhead. We introduce a benchmark of 9,822 single-turn examples spanning 79 vehicle functions derived from Android Automotive, including held-out functions and requests requiring refusal. We compare both approaches under matched fine-tuning across four SLMs from 270M to 1.7B parameters. On functions seen during training, scaling provides limited benefit: the 270M model can match the 1.7B model, while the strongest overall performance occurs at 0.6B. On held-out functions, FT achieves zero accuracy by construction, whereas SIP generalizes and improves substantially with scale. On out-of-scope requests, FT can invoke an unavailable function it was trained to emit, while SIP more reliably refuses based on the functions offered. This flexibility comes with higher memory use and latency. Our theoretical analysis explains how SIP enables generalization and why longer schema contexts increase inference cost. Overall, function-surface representation, rather than model scale alone, determines the capabilities and failure modes of SLM-based vehicle function calling.
[NLP-71] Edu-QuRating: Multi-Dimensional Educational Data Curation with Distilled Pairwise Judgements
【速读】： 该论文旨在解决现有教育数据过滤方法中将教育价值简化为单一标量属性所带来的局限性，尤其在高密度教育内容的数据集中，这种简化难以满足多样化应用对学习材料质量的精细化需求。为此，论文提出Edu-QuRating——一种多维度教育数据评分与筛选的流水线框架。其核心创新在于构建了面向教育场景的多维评价标准（rubrics），利用大语言模型（LLM）作为评判者对采样文档对进行成对偏好标注，并通过偏好蒸馏生成可复用的Edu-QuRater模型，能够对文本片段在多个教育维度上进行独立评分。实验表明，最优Edu-QuRater在六项教育标准上对保留的GPT-4.1-mini成对判断的平均准确率达0.917。进一步应用显示：在小语言模型预训练中，基于Edu-QuRating筛选的3.22亿条FineWeb-Edu-Fortified数据显著提升了模型在九个基准测试中的综合表现，尤其在特定任务上增益明显；在后训练阶段，将Edu-QuRater得分作为奖励项引入GRPO（Generalized Reward-based Policy Optimization）框架，结合答案结构奖励，生成的响应在教学质量和指令遵循方面均优于Qwen3-4B基线模型。因此，该方案的关键在于通过多维度、可复用的教育质量评估机制，实现对教育数据的精细化筛选与优化利用。
链接: https://arxiv.org/abs/2609.09425
作者: Oliver G. B. Garrod,Robin A. A. Ince,Meng Liu,Mohamed Huti,Moritz Boos,Amy Waldock,Dominic Andrews,Paul Atherton
机构: Fab AI
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Educational data filters have become a practical way to improve language-model pre-training, but most filters treat educational value as a single scalar property. This may be too broad for some applications, especially if the data set already features a high density of educational material. Useful learning material needs to be accurate, engaging, well structured, and appropriate for the intended audience and application (e.g. learner- vs teacher-facing). Following QuRating (Wettig et al. 2024), we introduce Edu-QuRating: a pipeline for multi-dimensional educational data scoring and curation. Edu-QuRating defines education-specific rubrics, uses an LLM judge to label sampled document pairs and distills those pairwise preferences into reusable Edu-QuRaters, which can score individual text chunks on a set of educational criteria. Across two sequence-classification base models and six educational criteria, the best Edu-QuRater recovers held-out GPT-4.1-mini pairwise judgements with mean accuracy 0.917. We then apply the resulting scorers in two applications. First, we investigate the potential of Edu-QuRaters for corpus filtering to improve pretraining of small language models. We scored 322.25M FineWeb-Edu-Fortified documents to obtain a filtered pre-training mixture. In matched single-run pre-training comparisons, models trained with Edu-QuRating-based mixtures reached higher observed aggregate accuracy across nine benchmarks than the FineWeb-Edu baseline, with gains concentrated in particular tasks. Second, we used Edu-QuRater scores as reward terms for GRPO post-training. In held-out pairwise judge evaluations, combining Edu-QuRater and answer-structure rewards produced responses preferred to the Qwen3-4B base model on both pedagogical quality and instruction following.
[NLP-72] Benchmarking Hybrid Deep Research Across Database Querying and Web Search
【速读】： 该论文旨在解决多模态信息融合背景下自主智能体在复杂现实任务中面临的“跨系统证据传递”难题，即如何在非结构化文本（如开放网络）与结构化数据（如关系型数据库）之间有效迁移并保持约束条件的完整性。现有基准测试通常孤立评估单一模态的能力，忽视了真实场景中混合使用多种信息源时的关键挑战——即在不同系统间传递证据时维持逻辑一致性与精确性。为此，作者提出HybridDeepResearch，作为首个要求同时调用网页搜索与SQL查询以生成可验证答案的深度研究基准。其核心创新在于构建了380个依赖工具的任务，基于LiveSQLBench-Base-Lite数据库和公开网络语料库，涵盖三种推理模式（SQL2S、S2SQL与Parallel），并通过自动化验证与人工评审确保质量。实验表明，即使最先进的模型（如GLM-5.2、Claude-Sonnet-4.6和GPT-5）在困难子集上也仅达到约50%-54%的Pass@8性能，且定向推理远难于并行交叉推理，凸显当前智能体系统在结构化与非结构化信息空间间无缝衔接且不丢失约束条件方面仍存在重大瓶颈。解决方案的关键在于设计一个综合性、可验证的多模态评估框架，以揭示并推动智能体在异构数据环境下的协同推理能力发展。
链接: https://arxiv.org/abs/2609.09410
作者: Ruofan Wu,Peiran Xu,Xiaolong Li,Fan Shu,Soyoung Yoon,Yite Wang,Xiaodong Yu,Boyi Liu,Feng Yan,Debiao Li,Yuxiong He,Zhewei Yao
机构: University of Houston (休斯顿大学); University of California, Los Angeles (加州大学洛杉矶分校); The University of Hong Kong (香港大学); Seoul National University (首尔国立大学); Snowflake AI Research
类目: Computation and Language (cs.CL)
备注:
Abstract:While autonomous agents have made significant strides in “deep research” by iteratively navigating the open web to synthesize information, real-world problem-solving is rarely confined to a single environment. Complex analytical tasks inherently require agents to weave together evidence from both ambiguous unstructured text (e.g., the open web) and highly precise structured data (e.g., relational databases). However, existing benchmarks evaluate these modalities in isolation, failing to capture the critical “handoff” - the ability to preserve constraints when moving evidence between systems. We introduce HybridDeepResearch, to our knowledge the first deep-research benchmark that requires both web search and SQL to form a complete, verifiable answer. The benchmark contains 380 tool-dependent tasks grounded in LiveSQLBench-Base-Lite databases and public web corpora, validated through automated checks and human review, and covering three reasoning patterns: SQL2S, S2SQL, and Parallel. Evaluations across proprietary and open-weight models under various agentic scaffolds reveal that even state-of-the-art models like GLM-5.2, Claude-Sonnet-4.6 and GPT-5 achieve only about 50-54% Pass@8 on the hard subset. Notably, results show that directional reasoning is substantially more difficult than parallel intersection, highlighting that bridging structured and unstructured information spaces without losing constraints remains a major open challenge for agentic systems. Code and datasets are publicly available at GitHub (this https URL) and Hugging Face (this https URL).
[NLP-73] Do LLM s Make More Mistakes If They Do Not Believe the Input Data?
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在检索增强生成或数据到文本系统中因幻觉（hallucination）或对上下文事实的误解释而导致的忠实性（faithfulness）问题。其核心挑战在于模型在面对与自身参数化知识存在冲突的外部上下文时，如何保持对所提供信息的忠实性，即“上下文-记忆冲突”（context-memory conflict）。解决方案的关键在于通过设计多语言、多类型语料（包括真实、反事实和虚构的RDF三元组）来系统评估模型在不同语言环境（英语、捷克语、斯洛伐克语及上索布语）下的忠实性表现，并借助人工标注样本验证模型判断的一致性。研究发现，在人类标注样本中，上下文-记忆冲突效应较弱，反事实输入仅比真实输入略低0.05分（1–5评分尺度），表明模型对非主流语言和局部知识的依赖程度可能被高估；同时揭示了选择不当的模型评判者（LLM judge）会显著夸大上下文-记忆冲突的强度，从而影响对模型忠实性的准确评估。
链接: https://arxiv.org/abs/2609.09363
作者: Peter Kochelka,Aleš Manuel Papáček,Vojtěch Dvořák,Ondřej Dušek
机构: Charles University, Faculty of Mathematics and Physics (查理大学数学与物理学院); Institute of Formal and Applied Linguistics (形式与应用语言学研究所); Prague, Czech Republic (捷克共和国, 布拉格)
类目: Computation and Language (cs.CL)
备注: 16 pages, 2 figures, to be published in INLG 2026
Abstract:Large language models (LLMs) are prone to hallucinating or misinterpreting facts, which impairs their usability in retrieval-augmented generation or data-to-text systems. We analyse how faithfulness of LLMs to provided context depends on how plausible they perceive the context to be (context-memory conflict). To better identify error patterns, we make use of the increased difficulty of non-English and low-resource language text generation and input data based on local knowledge, only partially captured in models’ parametric knowledge. We let the models generate text in English, Czech, Slovak and Upper Sorbian from factual (FA), counterfactual (CFA) and fictional (FI) RDF triples containing local Czech and Slovak data. Contrary to our expectations, we observe only a weak context-memory conflict on the human-annotated sample. For Kimi K3 as an LLM judge, which agrees well with human annotations on the sample, counterfactual inputs receive only slightly lower faithfulness scores than factual ones (-0.05 on a 1-5 scale). We also find that a suboptimal choice of LLM judge would lead to overestimating the strength of the context-memory conflict.
[NLP-74] Auditable Emergency Triage for Maternal and Newborn Care in India
【速读】： 该论文旨在解决在大规模医疗咨询场景中，基于大语言模型（LLM）的紧急情况分诊系统因缺乏可解释性与可审计性而导致的临床应用障碍。核心问题在于：传统方法依赖模型对症状的隐式判断，导致错误分析困难、提示工程变更需重新评估全量数据，且未显式建模临床决策流程。解决方案的关键在于将分诊过程分解为两个可审计阶段：首先由临床专家制定的术语体系驱动的LLM提取标准化症状与患者背景信息；随后通过确定性规则引擎依据临床决策树识别紧急情形。该架构显著提升了召回率（从0.565增至0.810）和F1分数（从0.606增至0.702），且结构化规则贡献了主要性能提升。更重要的是，该设计实现了端到端可追溯性，使临床专家可独立审查各环节的输出并快速添加新规则，无需重跑昂贵的评估流程，从而构建起高效、安全的持续迭代机制。
链接: https://arxiv.org/abs/2609.09356
作者: Shobhit Jagga,Aman Dalmia,Niharika Priyadarshini,Neelima Devadas,Amrita K Prasen,Nikhil Nalin,Santhosh SJ,Sreeram Nurani Ramasubramanian,Muhammed Afeer K,Anubhav Arora
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: First three authors contributed equally
Abstract:At Noora Health, our nurses answer more than 50,000 medical queries per month on our WhatsApp-based service that provides caregivers with on-demand support. Their most time-critical task is emergency triage: deciding which queries need immediate in-person attention. To support them, we built a system that uses a large language model (LLM) to classify whether a message is an emergency and provide a rationale for interpretability. But the system was opaque: analyzing mistakes meant reading reasoning chains for each message, which is infeasible at our scale. Prompt changes meant re-running a full evaluation to prevent regressions, which was both costly and operationally challenging. Clinicians follow a decision tree to make this call, but it was never documented or passed to the model, which relied on a flat list of danger signs. To address these issues, we decomposed triage into two steps: an LLM extracts canonical symptoms and patient context from the query using a clinician-authored vocabulary, and a deterministic rule engine captures the scenarios that indicate an emergency. We show that the new system raised recall from 0.565 to 0.810 and F1 from 0.606 to 0.702, with structured rules driving most of the accuracy gains while the decomposition provides auditability: clinical experts can inspect each stage of the new system to see whether the query was mistranslated, symptoms were incorrectly extracted, patient context was wrongly inferred, or the necessary rules were missing. They can add new rules independently without causing regressions and avoid running costly evaluations. Since deployment, the new system has triaged 152,421 patient queries and flagged 28,535 (18.7%) as emergencies. The over-escalation rate has been 17.8%, without any increase in missed emergencies. Clinicians have also added 48 new rules since deployment, evidence of the faster correction loop we set out to build.
[NLP-75] SWORD: Wikidata-based Distortions Reveal Hidden Cross-Lingual Inconsistencies in LLM Factual Error Rejection
【速读】： 该论文旨在解决当前大型语言模型（LLM）在多语言场景下评估中普遍存在的核心问题：现有标准基准测试主要关注模型是否能选出正确答案，而未能有效衡量其对事实信息的真实理解能力。为此，研究提出Systematic Wikidata-based Object-Relation Distortion（SWORD）基准，通过在八种广泛使用的语言中对Wikidata三元组进行受控扰动（如随机实体替换或语义上合理的属性选择），生成语法正确但事实错误的陈述，以系统性地检验模型在跨语言环境下识别事实错误的一致性。其解决方案的关键在于采用基于知识库的语义扭曲机制，揭示传统评估方法无法捕捉的深层缺陷——模型在面对语义合理但错误的事实陈述时反而表现优于荒谬的随机替换，表明其依赖于分布外的统计熟悉度而非真实事实验证；同时，模型在东亚语言上的性能显著下降，跨语言差距可达28个百分点（相对下降49%），凸显多语言事实推理能力的非对称性，而这一现象被聚合准确率指标所掩盖。
链接: https://arxiv.org/abs/2609.09349
作者: Sanghyeok Park,Minji Kang,Hosung Kwak,Jinhyuk Yun
机构: Soongsil University (中央大学); KAIST (韩国科学技术院)
类目: Computation and Language (cs.CL)
备注: 20 pages, 12 figures, 6 tables (including appendix)
Abstract:Modern LLMs demonstrate impressive multilingual performance, yet standard benchmarks primarily reward selecting correct answers rather than evaluating genuine factual understanding. We introduce Systematic Wikidata-based Object-Relation Distortion (SWORD), a benchmark that evaluates whether models consistently reject factual errors across languages. SWORD generates syntactically well-formed but factually incorrect statements in eight widely spoken languages through controlled perturbations of Wikidata triples, ranging from random entity substitutions to semantically plausible property-based selections. Our distortion-based evaluation surfaces two critical insights that remain entirely obscured by conventional benchmarks. First, models counterintuitively achieve higher accuracy on semantically plausible distortions than on nonsensical random substitutions, suggesting reliance on distributional familiarity rather than genuine factual verification. Second, models exhibiting comparable baseline accuracy across languages show substantial performance degradation specifically on (East) Asian languages when presented with distorted statements, with cross-lingual performance gaps reaching up to 28 percentage points (49% relative reduction) in some models. These findings demonstrate that multilingual factual reasoning involves asymmetric capabilities that aggregate accuracy metrics systematically obscure.
[NLP-76] Osprey: Target-agnostic Pre-training Makes Stronger Drafters in Speculative Decoding EMNLP2026
【速读】： 该论文旨在解决生成式 AI（Generative AI）在大语言模型（Large Language Model, LLM）推理加速过程中，基于推测解码（speculative decoding）技术所面临的接受率（acceptance rate）脆弱性问题。现有方法中，推测生成器（drafter）通常针对单一目标模型进行训练，仅在特定分布下有效，一旦工作负载发生偏移，其性能便会急剧下降，这与现代大模型通过大规模预训练获得广泛泛化能力的趋势形成鲜明反差。论文提出的关键解决方案是引入 Osprey 框架，其核心在于将可直接使用的预训练小型语言模型（small language model, SLM）作为初始骨架，通过目标无关的预训练实现通用能力复用，从而将每种目标模型的适配过程简化为轻量级微调步骤，避免了传统方法中对每个目标模型重复进行端到端训练的高成本。为实现这一目标，Osprey 面临两大挑战：一是小型模型通常远比延迟敏感的推测生成器所能承受的深度更深；二是必须在保持原始预训练计算能力的同时，使生成器能够接收目标模型的隐藏状态并输出目标词汇表中的标记。为此，Osprey 采用浅层剪枝构建基础架构，通过目标无关的下一个词预测任务恢复其语言建模能力，并结合词汇对齐、零初始化的 QKV 扩展以及从目标模型输出分布进行知识蒸馏等策略完成对具体目标的适应。实验表明，单个预训练的 Osprey 骨干可在多个目标模型间迁移使用，在 Qwen3-8B、Llama-3.3-70B-Instruct 及 MiniMax-M2.5 等模型上分别提升平均接受长度 16.1%、21.2% 和 22.7%，同时实现 17.5% 的每秒生成令牌数增益，尤其在域外和多语言数据上表现更优。
链接: https://arxiv.org/abs/2609.09338
作者: Fengxiang Bie,Yuqing Jian,Yifan Yu,Zhongzhu Zhou,Zelei Shao,Ben Athiwaratkun,Shuaiwen Leon Song,Chenfeng Xu,Xiaoxia Wu,Tianyi Zhang
机构: Together AI; The University of Sydney (悉尼大学); University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校); The University of Texas at Austin (德克萨斯大学奥斯汀分校)
类目: Computation and Language (cs.CL)
备注: Accepted at EMNLP 2026. 21 pages, 4 figures
Abstract:Speculative decoding is critical for accelerating LLM inference. However, the speedup is fragile: drafters are typically trained against a narrow distribution for a single target model, and their acceptance rate collapses under workload shifts. This is a striking inversion of modern LLM development, where target models are valued precisely for the broad generalization they acquire through large-scale pretraining. We argue that the natural remedy, pretraining, has been hard to apply to drafters because existing recipes are target-specific: the drafter consumes the target’s hidden states and is distilled on the target’s logits, so pretraining must be repeated for each target. We introduce Osprey, which instead bootstraps drafters from off-the-shelf pretrained small language models, treating broad pretraining as a reusable, target-agnostic asset and reducing per-target work to a lightweight adaptation step. Realizing this requires overcoming two challenges: small LMs are far deeper than a latency-bound drafter can afford, and their pretrained computation must remain intact while the drafter learns to ingest target hidden states and emit tokens in the target’s vocabulary. Osprey addresses both by pruning to a shallow backbone, restoring its language-modeling capability with target-agnostic next-token pretraining, and adapting it to each target through vocabulary alignment, zero-initialized QKV expansion, and distillation from the target model’s output distribution. Empirically, a single pretrained Osprey backbone transfers across targets and improves mean acceptance length by 16.1% for Qwen3-8B, 21.2% for Llama-3.3-70B-Instruct, and 22.7% for the 229B MiniMax-M2.5 (with 17.5% higher tokens per second), with the largest gains on out-of-domain and multilingual data. Our code is available at this https URL.
[NLP-77] StochBench: A Domain-Specific Benchmark for Stochastic Processes in Lean
【速读】： 该论文旨在解决当前大语言模型在形式化定理证明领域中存在的基准测试样本代表性不足的问题，尤其是现有主流基准（如国际数学奥林匹克竞赛IMO和普特南数学竞赛Putnam）多集中于竞赛类数学问题，难以反映实际应用领域的复杂性与专业性。针对这一缺陷，论文提出StochBench——一个基于Lean 4的新型基准，包含450道研究生级随机过程（stochastic processes）问题，涵盖有限与可数马尔可夫链、更新过程、随机游走、鞅、停时、排队论、布朗运动、随机微积分、弱收敛以及泊松过程与连续时间马尔可夫过程等核心主题，并为每道题提供自然语言来源描述。该基准填补了Mathlib中随机过程领域的空白，更真实地反映了领域特定的应用数学需求。其关键解决方案在于构建一个兼具高抽象层次多样性与实际应用背景的高质量、大规模领域专用基准，并通过基于Opus 4.8的智能体在每题15分钟时限下实现34.9%（157/450）的证明成功率，验证了该基准对先进形式化证明系统仍具足够挑战性。
链接: https://arxiv.org/abs/2609.09264
作者: Idan Davidovich,Debargha Ganguly,Vikash Singh,Vipin Chaudhary
机构: Case Western Reserve University (凯斯西储大学)
类目: Computation and Language (cs.CL); Logic in Computer Science (cs.LO)
备注:
Abstract:Leading benchmarks for formal theorem proving with large language models are small collections drawn from competition math, such as the IMO and Putnam, that poorly represent field-specific applications. We introduce StochBench, a Lean 4 benchmark of 450 graduate stochastic-processes problems at varying abstraction levels, each paired with its natural-language source. Addressing a field underrepresented in Mathlib, it covers finite and countable Markov chains, renewal processes, random walks, martingales, stopping times, queues, Brownian motion, stochastic calculus, weak convergence, and Poisson and continuous-time Markov processes. Our Opus 4.8-based agent achieves a 34.9% proof rate (157/450) under a 15-minute per-problem limit. StochBench better represents domain-specific applied mathematics while remaining challenging for advanced provers.
[NLP-78] In RAG We Trust? Measuring Robustness of Retrieval-Augmented Generation Under Document Poisoning
【速读】： 该论文旨在解决检索增强生成（Retrieval-Augmented Generation, RAG）系统在面对外部检索内容被篡改时的脆弱性问题，即当部分检索到的文本被恶意污染后，生成模型可能传播虚假信息。其核心解决方案的关键在于评估一个小型量化模型（Llama 3.1 8B）在不同污染策略下的鲁棒性表现，具体测试了三种篡改方式：实体替换（entity swap）、数值替换（number swap）和否定（negation），并系统分析其对模型准确率的影响。实验结果表明，在未污染情况下模型准确率为77.9%，而当所有三段检索内容均被污染时，准确率下降至43.5%；其中实体替换导致正确答案翻转的比例最高，数值篡改在污染占比低于一半时影响较小，但一旦超过阈值则出现显著性能跃降。值得注意的是，模型在攻击下主要表现为拒绝生成（abstain），而非主动编造新谎言，且支持性语义重叠度反而降低，表明其对不可靠输入具有一定的自我抑制机制。研究虽为小规模、基于粗粒度自动标注的初步测量，但揭示了不同污染策略的差异性效应，提示未来需通过更精细的解码控制与人工验证以进一步验证结论。
链接: https://arxiv.org/abs/2609.09243
作者: Iliano Fasolino
机构: University of Milan (米兰大学)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 6 figures. Preprint also available on Zenodo: this https URL
Abstract:Retrieval-augmented generation (RAG) grounds a language model in retrieved documents, which reduces hallucination but creates a new attack surface: if retrieved text is tampered with, the model may repeat the falsehood. We study how much a small quantized model, Llama 3.1 8B, degrades when a fraction of its retrieved context is poisoned. Three corruption strategies are tested, entity swap, number swap, and negation, each applied to zero, one, two, or three of the three retrieved passages, over a factorial sweep of 588 runs on a fact-checking task built from FEVER. Accuracy falls from 77.9% on clean context to 43.5% when all three passages are corrupted. Entity swap flips the largest share of answers that were correct on clean context. Number-based corruption stays flat while poisoned passages are a minority and jumps once they form a majority, a pattern we re-check with query-level bootstrap intervals. The model rarely invents new falsehoods; its dominant reaction is to abstain, and a lexical overlap proxy of unsupported generation falls under attack rather than rising. The study is a small-scale measurement with coarse automated labels; we treat the strategy contrasts as suggestive until decoding is controlled and stronger adjudication is in place.
[NLP-79] Distribution-Consistent Inference for Dynamic Sparse Mixture-of-Experts EMNLP2026
【速读】： 该论文旨在解决在稀疏化专家混合模型（Sparse Mixture-of-Experts, SMoE）中，推理时动态降低激活专家数量（dynamic top-k routing）所引发的分布偏移（distributional shift）问题。尽管此类方法可在不重新训练的情况下减少计算开销，但现有方法忽视了因偏离训练阶段路由配置而导致的输出表示分布变化，进而导致下游任务性能下降。其核心问题在于：减少激活专家数会系统性地增大输出的均方根（RMS）尺度与方差，造成表示失配（representation mismatch），从而损害模型表现。为解决这一可校正的偏差，作者提出轻量级的层间分布对齐（Layer-wise Distribution Alignment, LDA）机制，通过利用各层的校准统计量，在推理时对低路由开销下的表示进行分布对齐，使其逼近默认路由配置下的分布。实验表明，LDA在多种SMoE大语言模型、基准测试和路由策略下，能有效恢复因分布偏移造成的性能损失，同时保持稀疏推理的高效性，且引入的额外计算开销可忽略不计。
链接: https://arxiv.org/abs/2609.09241
作者: Dohyeon Kim,Bedionita Soro,Sung Ju Hwang
机构: KAIST(韩国科学技术院); DeepAuto.ai
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted to Findings of EMNLP 2026
Abstract:Mixture-of-Experts (MoE) architectures have emerged as a powerful paradigm for scaling model capacity while preserving efficient inference in large foundation models. However, most MoE models use a fixed top- k expert selection policy, assigning the same expert budget to every token even when fewer experts may be sufficient. Inference-time dynamic top- k routing can reduce computation without retraining, but existing methods often overlook the distributional shift caused by deviating from the training-time routing configuration. We show that reducing the number of activated experts consistently increases the RMS scale and variance of SMoE outputs, inducing a representation mismatch that contributes to downstream performance degradation in addition to the loss of expert capacity. To address this correctable component, we propose Layer-wise Distribution Alignment (LDA), a lightweight inference-time correction that uses layer-wise calibration statistics to align reduced-routing representations with the default configuration. Across multiple SMoE LLMs, benchmarks, and routing strategies, LDA recovers much of the performance lost induced by the distributional shift under reduced routing while preserving sparse-inference efficiency with negligible overhead.
[NLP-80] Subagents vs Agent Agent Skills: Executing Reusable Knowledge for Long-Horizon Agentic Tasks
【速读】： 该论文旨在解决语言模型智能体在执行长时程任务时，如何有效利用可复用的知识库（即“技能包”）以提升任务完成能力的问题。现有方法通常将技能包中的指令加载至主智能体的上下文窗口中并依赖其自主执行，但随着任务复杂度增加，上下文窗口信息累积导致推理质量显著下降，从而限制了性能表现。本文提出一种替代方案：将技能包作为子智能体（subagent）调用，而非直接注入主上下文。通过为每个子任务创建独立的、全新的上下文窗口来执行子任务，避免了上下文污染与信息过载问题。研究发现，当技能包具备清晰的输入-输出契约（input-output contract），且其内部指令编码了实现契约所需的程序性知识时，子智能体执行方式显著优于传统上下文加载模式。尽管该方法引入了额外的通信开销（需更多标记用于主智能体与子智能体间的协调），但整体性能优势明显。因此，研究结论表明，可复用知识的价值不仅取决于其内容质量，更关键的是其组织结构与调用机制的设计。
链接: https://arxiv.org/abs/2609.09233
作者: Wasu Top Piriyakulkij,Rachel Lawrence,Alicia Curth,Sushrut Karmalkar,Niranjani Prasad
机构: Cornell University(康奈尔大学); Microsoft Research Cambridge(微软研究院剑桥)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:How can language model agents effectively leverage libraries of reusable knowledge to solve long-horizon tasks? Recent work has increasingly focused on agent skills: reusable capabilities represented as skill packages, i.e., multi-file bundles containing instructions, scripts, and other resources that help agents perform specific tasks. Agent skills are typically executed by loading their skill instructions into an agent’s context and relying on the agent to follow them. As task horizons grow, however, this approach becomes increasingly brittle, because reasoning quality degrades as more information accumulates in the context window. We investigate an alternative approach in which skill packages are instead invoked as subagents. Rather than loading skill instructions into the main context, subagent execution spawns fresh context windows dedicated to solving individual subtasks. We show that subagent execution outperforms agent-skill execution when skill packages expose clear input-output contracts and their instructions encode the procedural knowledge needed to fulfill those contracts. The tradeoff is additional communication overhead, as extra tokens are required to coordinate between the main agent and its subagents. Our results show that the benefit of reusable knowledge depends not only on its content, but also on how it is organized and invoked.
[NLP-81] MLLM s Hallucinate when Information Distribution Drifts in Synergy Heads
【速读】： 该论文旨在解决多模态大语言模型（Multimodal Large Language Models, MLLMs）中存在的幻觉问题，该问题严重制约了模型在实际应用中的可靠性。现有基于注意力机制的缓解方法主要依赖于间接信号（如注意力权重），无法准确反映导致幻觉生成的实际信息分布偏移。本文提出一种名为HEAL（Head-level information disentanglement and calibration）的新方法，其核心在于通过因果噪声干预筛选出因果冗余的注意力头，并利用反事实差分法（counterfactual Difference-in-Differences）对剩余注意力头内的信息分布进行解耦，将其分类为四类。研究发现，幻觉的发生源于“协同头”（synergy heads）中信息分布偏离健康均衡状态，而非模态特异性头的数量或强度。基于此洞察，HEAL在协同头的值向量中注入动态信息校准因子，主动调节视觉-语言依赖关系，引导输出分布向真实证据靠拢。实验结果表明，HEAL在多个MLLM上均能有效降低幻觉率，提供了一种简洁且可解释的提升模型可信度的路径。
链接: https://arxiv.org/abs/2609.09206
作者: Meng’en Qin,Junye Chen,Jucheng Liu,Youlu Xing,Song Wang,Ruize Han
机构: Shenzhen University of Advanced Technology (深圳先进科技学院)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Multimodal Large Language Models (MLLMs) often struggle with hallucinations, thus hindering their reliable practical applications. Existing attention-based mitigation methods mainly rely on indirect signals (e.g., attention weights) that fail to accurately reflect the actual information shift underlying hallucination generation. In this paper, we propose HEAL, Head-lEvel information disentAnglement and caLibration for identifying and mitigating hallucinations. HEAL first employs causal noise intervention on multi-head outputs to filter out causally redundant heads. Subsequently, it disentangles information distribution within the remaining heads via the counterfactual Difference-in-Differences, categorizing heads into four types. Through analysis, we observe: hallucinations happen when information distribution drifts away from a healthy equilibrium in synergy heads, not strongly correlated with the quantity or strength of modality-specific heads. Motivated by this insight, HEAL injects dynamic information calibration factors into the value vectors of synergy heads, and actively regulates visual-language dependencies, steering the output distribution towards factual evidence. Extensive experiments demonstrate that HEAL effectively reduces hallucinations across multiple MLLMs, offering a simple and interpretable pathway to enhance model trustworthiness.
[NLP-82] Agent icGen: Reward-Guided Agent ic Video Generation for Advertising
【速读】： 该论文旨在解决广告视频生成中长期存在的“生成质量与商业效果脱节”问题，即现有生成式AI模型虽能基于多模态输入合成逼真视频片段，但缺乏对广告策略优化及基于线上业务反馈迭代改进的机制。其核心挑战在于如何将广告视频生成从单纯的视觉合成任务，转化为可被在线业务指标（如点击率、转化率）引导的可优化推理过程。解决方案的关键在于提出AgenticGen框架——一种基于奖励驱动的智能体范式，将广告视频生成分解为可训练的两个推理阶段：策略选择（strategy selection）与初稿生成（draft generation），从而显式暴露可被在线反馈监督的优化目标。该框架通过融合累积的线上业务反馈构建性能导向奖励，并结合符合人类质量标准的规则型奖励，利用直接偏好优化（DPO）与渐进式奖励策略优化（GRPO）分阶段优化策略，实现从用户偏好到生成质量的双重提升。实验表明，该方法在抖音广告系统中的在线A/B测试中，相较监督微调基线，显著提升了2.72%的点击率（CTR）、2.63%的转化率（CVR）以及9.61%的广告价值（Advv）。
链接: https://arxiv.org/abs/2609.09187
作者: Xingyuan Bu,Chengru Song,Hao Zhou,Tao Zhou,Dong Li,Wei Li,Shilong Li,Hao Shi,Yongxin Guo,Donghao Zhou,Qiangpeng Yang,Shilei Wen
机构: 未知
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Advertising video generation is not only a video synthesis task, but also a product-conditioned reasoning problem whose success is measured by online business metrics. Recent video foundation models can generate realistic clips from multimodal conditions, yet they do not optimize how a product should be transformed into an effective advertisement or how future generation should be improved from online business feedback. To close this loop, we propose AgenticGen, a reward-guided agentic framework that decomposes advertising video generation into two trainable reasoning stages, strategy selection and draft generation, thereby exposing optimization targets that online business feedback can supervise. AgenticGen learns a performance-based reward from accumulated online feedback and a complementary rubric-based reward aligned with human quality standards, then uses them to supervise policy optimization. DPO first moves the agentic policies toward online preferences, and GRPO further refines both stages with process and outcome rewards. Offline experiments validate the reward models and successive policy optimization. Online A/B experiments in the TikTok advertising system show that AgenticGen after DPO and GRPO improves CTR by 2.72%, CVR by 2.63%, and Advv by 9.61% over the SFT baseline.
[NLP-83] X-CoSD: Communication-Efficient Cross-Vocabulary Collaborative Speculative Decoding
【速读】： 该论文旨在解决分布式大语言模型（LLM）推理框架中协同推测解码（CoSD）存在的两大核心问题：一是现有方法假设小型语言模型（SLM）与大型语言模型（LLM）共享词汇表，限制了模型异构性；二是残差重采样机制导致设备与边缘服务器间需频繁传输完整的词元分布，造成显著通信开销。其解决方案的关键在于提出跨词汇表协同推测解码（X-CoSD），通过混合重采样（HR）机制将残差重采样过程分解至设备端的共用词汇区域和服务器端的LLM独有词汇区域，仅需在共用词汇区域传输分布信息，从而实现无损且通信高效的推理。进一步提出的X-CoSD-E变体采用服务器重采样与设备验证（SR-DV）策略，仅由服务器发送候选替换词及其概率，由设备本地完成验证，进一步降低通信负载。理论证明二者均能保持服务器LLM的原始分布，实验表明其在显著提升生成速度的同时，仍可维持与服务器LLM相当的生成质量。
链接: https://arxiv.org/abs/2609.09166
作者: Jaeduk Lee,Wan Choi
机构: Seoul National University (首尔国立大学)
类目: Computation and Language (cs.CL); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
备注:
Abstract:This paper investigates collaborative speculative decoding (CoSD), a distributed large language model (LLM) inference framework in which an on-device small language model (SLM) drafts candidate tokens and a server LLM verifies them. Existing CoSD methods assume a shared vocabulary between the SLM and the LLM and incur substantial communication load because residual resampling requires token distribution exchange between the user device and the edge server. To address these limitations, we propose cross-vocabulary CoSD (X-CoSD), a lossless and communication-efficient CoSD framework for heterogeneous SLM-LLM vocabularies. X-CoSD is built on hybrid resampling (HR), which splits residual resampling across the common-vocabulary region on the device and the LLM-only region on the server, so that distribution transmission is required only for the common-vocabulary region. We further propose X-CoSD-E, an enhanced variant based on server resampling with device verification (SR-DV), in which the server sends only replacement candidates sampled from the server LLM and their corresponding probabilities for local verification at the device. We prove that both X-CoSD and X-CoSD-E preserve the server LLM distribution, and experiments show that they significantly improve token generation speed while maintaining generation quality comparable to that of the server LLM.
[NLP-84] Less Is Personal: Learning Minimal Sufficient User Profiles for Personalized Language Models
【速读】： 该论文旨在解决生成式AI（Generative AI）在个性化任务中因冗余或无关用户历史记录引入而导致的上下文开销过大与个性化效用下降的问题。现有个性化语言模型通常采用固定长度的检索记录拼接方式，忽视了用户行为序列中潜在的冗余性与相关性差异，导致不必要的上下文成本和性能退化。其解决方案的关键在于提出一种“最小充分个性化”（minimal sufficient personalization）框架——即在保证个性化效果的前提下，构建最短且最优序的用户行为配置（profile）。为此，论文提出了ENOUGH方法：通过迭代式地添加行为记录或触发停止信号（STOP），自适应生成可变长度的个性化配置；在离线阶段，采用受约束的反事实搜索（bounded counterfactual search）联合评估下游性能增益、用户特异性与分词成本，生成长周期优化目标，并将其蒸馏为多头值控制器（multi-head value controller），该控制器具备显式的排序与停止监督能力；推理时，控制器以轻量级决策选择并排序记录，仅在停止后调用一次冻结的生成器，从而实现高效、精准的个性化生成。实验表明，ENOUGH在六项个性化任务上显著优于主流启发式与检索增强基线，在保持个性化效用的同时，有效降低了上下文开销。
链接: https://arxiv.org/abs/2609.08180
作者: Minghang Liu,Qiang Qiu,Yuanzhuo Wang,Huawei Shen,Xueqi Cheng
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 21 pages
Abstract:Retrieval-augmented personalization enables large language models to produce more accurate and preference-aligned outputs using relevant records retrieved from user histories. Personalized language models typically prepend a fixed number of retrieved user records, even when additional history is redundant, harmful, or unrelated to a user’s distinctive behavior. We study minimal sufficient personalization: constructing the least costly ordered profile for each input while preserving the utility achievable from a retrieved candidate pool. We introduce ENOUGH, a method that iteratively appends behavioral records or emits STOP to construct profiles with adaptive lengths. Offline, bounded counterfactual search evaluates profile prefixes by jointly considering downstream gains, user specificity, and token costs. The resulting long-horizon targets are distilled into a multi-head value controller with explicit ranking and stopping supervision. At inference, the controller selects and orders records through lightweight decisions, and the frozen generator is invoked once after stopping. Extensive experiments on six personalized tasks demonstrate that ENOUGH consistently outperforms strong heuristic and retrieval-augmented baselines in both effectiveness and efficiency, achieving minimal sufficient profiles that preserve personalization utility while reducing unnecessary context costs.
[NLP-85] From Plausible to Actionable: A Position on LLM Self-Explanations
【速读】： 该论文旨在解决生成式人工智能中大语言模型（Large Language Models, LLMs）自解释（self-explanations）的可信性与有效性问题。尽管模型生成的解释在表面上具有高度可读性和逻辑连贯性，但其是否真实反映模型内部推理过程仍存疑，这构成了可解释人工智能（Explainable Artificial Intelligence, XAI）领域的重要挑战。论文指出，当前主流评估范式仅关注解释的合理性（plausibility）和忠实性（faithfulness），难以全面衡量自解释的实际价值。其解决方案的关键在于拓展评估维度，引入“可操作性”（actionability）作为核心评价标准——即自解释不仅应合理且接近真实推理，更需能够支持多元利益相关者在复杂决策场景中做出明智判断与有效行动。通过这一视角重构评估框架，论文为推动大语言模型理性化能力在实际应用中的落地提供了理论指导与实践路径。
链接: https://arxiv.org/abs/2607.15957
作者: Elize Herrewijnen,Benedetta Muscato,Gizem Gezici,Fosca Giannotti
机构: University of Utrecht (乌得勒支大学); National Police Lab AI (荷兰警察国家实验室人工智能); Scuola Normale Superiore (意大利高等师范学校); University of Pisa (比萨大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 5 pages
Abstract:Large Language Models (LLMs) can generate natural language explanations that rationalize their own decisions, a phenomenon commonly referred to as self-explanations. Such explanations have emerged as a promising direction for explainable artificial intelligence (XAI), particularly for interpreting LLM behavior. However, while self-explanations often appear plausible, whether they faithfully reflect a model’s underlying reasoning process remains an open question. In this opinion paper, we argue that self-explanations can be highly plausible, questionably faithful, and yet highly actionable. From a traditional XAI perspective, we identify the limitations of standard evaluation protocols for LLM-generated self-explanations and propose practical guidelines for assessing their plausibility and this http URL, we argue that evaluation should extend beyond these criteria to actionability, highlighting applications of LLM rationalization capabilities that support informed decision-making and appropriate action across diverse stakeholders.
[NLP-86] Every Activation Boosted: Scaling General Reason er to 1 Trillion Open Language Foundation
【速读】： 该论文旨在解决大模型在推理能力扩展过程中面临的计算效率与推理性能难以兼顾的核心问题，尤其针对传统密集型模型在参数规模扩大时出现的资源开销剧增与推理质量提升不匹配的瓶颈。其解决方案的关键在于构建一个基于统一混合专家（Mixture-of-Experts, MoE）架构的系列化推理导向语言基础模型——Ling 2.0，通过高稀疏性激活机制、跨尺度一致性设计以及基于经验缩放定律的高效训练范式，实现从数十亿到一万亿参数规模的可扩展推理能力。核心创新包括：采用具有多任务感知（MTP）能力的高稀疏MoE结构以提升推理效率；引入面向推理的数据构造与中段思维链（CoT）激活策略；结合强化学习驱动的微调方法（如DFT与Evo-CoT）优化推理路径；并实现全规模FP8精度训练与细粒度异构流水线协同，最终在万亿参数级别上达成推理准确率与计算效率的帕累托前沿，验证了稀疏激活若与推理目标精准对齐，可有效支撑可扩展、高效率的智能推理系统。
链接: https://arxiv.org/abs/2510.22115
作者: Ling Team,Ang Li,Ben Liu,Binbin Hu,Bing Li,Bingwei Zeng,Borui Ye,Caizhi Tang,Changxin Tian,Chao Huang,Chao Zhang,Chen Qian,Chenchen Ju,Chenchen Li,Chengfu Tang,Chilin Fu,Chunshao Ren,Chunwei Wu,Cong Zhang,Cunyin Peng,Dafeng Xu,Daixin Wang,Dalong Zhang,Dingnan Jin,Dingyuan Zhu,Dongke Hu,Fangzheng Zhao,Feifan Wu,Feng Zhu,Gangshan Wang,Haitao Zhang,Hailin Zhao,Hanxiao Zhang,Hanzi Wang,Hao Qian,Haoyi Yu,Heng Zhang,Hongliang Zhang,Hongzhi Luan,Huirong Dong,Huizhong Li,Jia Li,Jia Liu,Jialong Zhu,Jian Sha,Jianping Wei,Jiaolong Yang,Jieyue Ma,Jiewei Wu,Jinjing Huang,Jingyun Tian,Jingyuan Zhang,Jinquan Sun,Juanhui Tu,Jun Liu,Jun Xu,Jun Zhou,Junjie Ou,Junpeng Fang,Kaihong Zhang,Kaiqin Hu,Ke Shi,Kun Tang,Kunlong Chen,Lanyin Mei,Lei Liang,Lei Xu,Libo Zhang,Lin Ju,Lin Yuan,Ling Zhong,Lintao Ma,Lu Liu,Lu Yu,Lun Cai,Meiqi Zhu,Mengying Li,Min Chen,Minghao Xue,Minghong Cai,Mingming Yin,Peijie Jiang,Peilong Zhao,Pingping Liu,Qian Zhao,Qing Cui,Qingxiang Huang,Qingyuan Yang,Quankun Yu,Shaowei Wei,Shijie Lian,Shoujian Zheng,Shun Song,Shungen Zhang,Shuo Zhang,Siyuan Li,Song Liu,Ting Guo,Tong Zhao,Wanli Gu
机构: Inclusion AI
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Ling 2.0 Technical Report
Abstract:We introduce Ling 2.0, a series reasoning-oriented language foundation built upon the principle that every activation boosts reasoning capability. Designed to scale from tens of billions to one trillion parameters under a unified Mixture-of-Experts (MoE) paradigm, Ling 2.0 emphasizes high sparsity, cross-scale consistency, and efficiency guided by empirical scaling laws. The series includes three non-thinking (instruct) models - Ling-mini-2.0, Ling-flash-2.0, and Ling-1T - ranging from 16B to 1T total parameters and achieving up to 7-fold active-compute efficiency compared with dense counterparts. Ling 2.0 integrates coordinated innovations across model architecture, pre-training, post-training, and infrastructure: a high-sparsity MoE with MTP for efficient reasoning, reasoning-oriented data and mid-training CoT activation, reinforcement-based fine-tuning (DFT, Evo-CoT), and full-scale FP8 training with fine-grained heterogeneous pipelines. At the trillion scale, Ling-1T establishes a new Pareto frontier of reasoning accuracy versus computational efficiency, demonstrating that sparse activation, when properly aligned with reasoning objectives, enables scalable and efficient intelligence. Collectively, Ling 2.0 provides a coherent, open, and efficient foundation for advancing future reasoning and thinking models, including the Ring series built upon the same base.
[NLP-87] Quantifying Logical Consistency in Transformers via Query-Key Alignment
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在多步逻辑推理任务中缺乏有效评估机制的问题，尤其针对现有链式思维提示（Chain-of-Thought prompting）方法虽能生成中间推理步骤但无法衡量其逻辑连贯性这一关键缺陷。其解决方案的核心在于提出一种轻量级的逻辑推理评估策略，通过分析Transformer模型注意力头内部的查询-键（Query-Key, QK）对齐模式，仅需一次前向传播即可提取“QK得分”（QK-score），从而揭示模型隐含表示中区分有效与无效推理的能力。该方法不依赖复杂的消融实验，具备良好的可扩展性，并在多个逻辑推理基准测试上验证了其对干扰项更强的鲁棒性及支持更深层次推理的能力，适用于从1.5B到70B参数规模的多种模型。
链接: https://arxiv.org/abs/2502.17017
作者: Eduard Tulchinskii,Anastasia Voznyuk,Laida Kushnareva,Andrei Andriiainen,Irina Piontkovskaya,Evgeny Burnaev,Serguei Barannikov
机构: Skolkovo Institute of Science and Technology (斯科尔科沃科学与技术学院); AI Foundation and Algorithm Lab; Moscow Institute of Physics and Technology (莫斯科物理技术学院); Artificial Intelligence Research Institute (AIRI) (人工智能研究院); CNRS, Université Paris Cité, France (法国国家科学研究中心，巴黎城市大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Theory (cs.IT); Machine Learning (cs.LG); Logic (math.LO)
备注:
Abstract:Large language models (LLMs) have demonstrated impressive performance in various natural language processing tasks, yet their ability to perform multi-step logical reasoning remains an open challenge. Although Chain-of-Thought prompting has improved logical reasoning by enabling models to generate intermediate steps, it lacks mechanisms to assess the coherence of these logical transitions. In this paper, we propose a novel, lightweight evaluation strategy for logical reasoning that uses query-key alignments inside transformer attention heads. By computing a single forward pass and extracting a “QK-score” from carefully chosen heads, our method reveals latent representations that reliably separate valid from invalid inferences, offering a scalable alternative to traditional ablation-based techniques. We also provide an empirical validation on multiple logical reasoning benchmarks, demonstrating improved robustness of our evaluation method against distractors and increased reasoning depth. The experiments were conducted on a diverse set of models, ranging from 1.5B to 70B parameters.
[NLP-88] What Does MMLU Actually Measure? A Psychometric Audit of Difficulty Structure in Aggregate Benchmark Scores
【速读】： 该论文旨在解决当前通用人工智能评估基准MMLU（Multi-Subject Multi-Task Learning Understanding）在衡量模型能力时存在的根本性缺陷：其综合得分主要反映模型的事实检索能力，而非推理能力，导致两者被错误地混为一谈。解决方案的关键在于采用项目反应理论（Item Response Theory, IRT）对1,000个开放权重语言模型在14,042个MMLU测试项上的表现进行难度校准，并进一步将题目难度与可确定性提取的结构复杂度框架进行回归分析。通过引入受主题聚类协方差调整的联合Wald检验，研究发现MMLU在STEM（科学、技术、工程、数学）与非STEM领域间对结构复杂度与难度的关系并非不变，表明其内部构念存在实质性分离。这一发现揭示了综合排行榜更敏感于非STEM领域的准确率，而非STEM推理表现，从而导致在推理密集型部署中选择前50名模型时，约有22%原本适合的STEM模型被排除。此外，在控制模型内生的多选猜测基线后，高能力模型在推理深度增加时仍表现出更陡峭的性能退化，说明MMLU对事实检索能力赋予了过高的权重，无意中偏向于经过检索优化的模型。为此，作者提出应采用可复现的确定性审计框架，并倡导对不同能力维度进行拆分报告，以实现更准确的能力评估。
链接: https://arxiv.org/abs/2609.09372
作者: Dana Paquin,Riddhiman Jain
机构: Stanford University (斯坦福大学)
类目: Number Theory (math.NT); Computation and Language (cs.CL)
备注: 14 pages
Abstract:Although MMLU is widely adopted as a benchmark for calibrating general AI capabilities, we psychometrically demonstrate that its aggregate score primarily evaluates a model’s factual retrieval capacity rather than its reasoning ability. By calibrating item difficulty for 1,000 open-weights language models over 14,042 MMLU test items using Item Response Theory, we show that evaluating both abilities via a single test is inherently flawed. Difficulty is then regressed on a deterministic, text-extractable framework of structural complexity. Applying a joint Wald test with subject-clustered covariances demonstrates that the MMLU conflates fundamentally separable constructs. The mapping from structural complexity to difficulty is not invariant across the benchmark’s STEM and non-STEM partitions. This finding has practical consequences. Aggregate leaderboard ranks track non-STEM accuracy more closely than STEM accuracy, so selecting a Top-50 model on the aggregate for a reasoning-intensive deployment displaces roughly 22% of the STEM-appropriate choices. Furthermore, when controlling for the multiple-choice guessing floor natively inside the response model, we find that higher-ability models continue to degrade more steeply under increased reasoning depth. The MMLU aggregate therefore weights retrieval capacity and reasoning stability unequally, inadvertently favoring models optimized for retrieval. We release our deterministic framework as a reproducible auditing instrument and recommend disaggregated reporting.
信息检索
[IR-0] GANDR: Claim Auditing for Verifiable Legal Answer Generation
链接: https://arxiv.org/abs/2609.10293
作者: Chen Qian,Yimeng Wang,Yu Chen,Lingfei Wu,Andreas Stathopoulos
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:In high-stakes domains such as legal practice, a language-model answer is only useful to the extent that a reader can verify each claim against the source the system cites. Current grounded-generation pipelines score the answer as a whole, so a correct conclusion can rest on fabricated or loosely matched citations and still score well. Closing this gap requires both a system built for per-claim verification and an evaluation that measures it. We introduce GANDR (Grounded ANswer DRafter), a two-agent system in which a Drafter writes an answer in a structured legal-reasoning format and a separate Critic, with the same view as a human verifier, audits each claim against its cited source and emits a per-claim audit trace on every round. We pair it with a strict correctness criterion requiring every citation to resolve to a passage the retriever returned. On a 185-item legal benchmark where all six systems share one backbone, one retrieval surface, and one citation instruction, GANDR ranks first on every primary metric, reaching 70.8% strict accuracy and leading the strongest baseline by 11.3 points (p0.01). Reverting the protocol-anchored commit rule lowers strict accuracy by 22.7 points, and the strict lead stays positive on three further backbones, at +3.2 to +6.5 points. This lead traces to the Drafter configuration and the protocol-anchored commit, not to rewriting. Against two law-trained annotators the audit flags under-supported claims at F1 0.84 as a binary detector, while its four-way verdict labels agree only weakly and are advisory. Code is available upon request.
[IR-1] LiteRAG : Cost-Efficient Graph-Based Retrieval-Augmented Generation
链接: https://arxiv.org/abs/2609.10239
作者: Daniel Alejandro Coll Tejeda,Pedro García López,Daniel Barcelona-Pons
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 16 pages, 2 figures
Abstract:Graph-based retrieval can improve multi-hop question answering, but existing approaches often incur high query-time costs and produce diffuse, oversized contexts that reduce generation efficiency. We present LiteRAG, a graph-based retrieval method that replaces expensive retrieval-time LLM control with query-conditioned algorithmic exploration and reasoning-chain context construction. On DistComp, a benchmark for multi-hop retrieval over distributed-systems papers, LiteRAG attains the highest overall quality among the evaluated methods (0.798) while reducing per-query latency by over 100 \times and cost by over 99% relative to GraphRAG Global and DRIFT. On UltraDomain, it matches LinearRAG on overall quality while using about 14 \times fewer tokens. An ablation study indicates that LiteRAG’s query-adaptive thresholding and community-aware hub penalization are the main drivers of its token-efficiency gains.
[IR-2] he Answer Path and the Grounding Instruction in LLM Question Answering over Knowledge Graphs
链接: https://arxiv.org/abs/2609.10237
作者: Arquimedes Canedo
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:A graph retrieval-augmented generation pipeline chooses which triples to put in the prompt, a syntax to write them in, an order to write them in, and a sentence telling the model what to do with them. We vary all four over six large language models and two knowledge-graph question answering benchmarks. Two of the four choices move the answer and the other two are flat. The first is whether the answer path, the triples needed to reach the answer, is in the prompt at all. Holding the number of triples fixed and replacing every triple that is not on the chain with material from an unrelated entity changes answer accuracy by +0.003 F1, while removing the chain costs most of what the graph was worth. Retrieval budget belongs on recall, and precision in the range we can test buys nothing. There is no retriever here: subgraphs come from gold SPARQL, so precision describes the context we build, not a system setting. The second is the grounding instruction. With no facts in the prompt, telling a model to answer using only the provided facts drops F1 from 0.299 to 0.035, a factor of 8.63. That figure describes an evaluation with an empty context arm rather than a working pipeline, and an experiment that applies the instruction to its context arm but not to its no-context baseline manufactures a spurious finding that graph context hurts at depth. We found one in our own results and retract it. Syntax, triple order and subgraph size produce no effect we can measure at multi-hop depth. The comparison that would price the grounding instruction against correct context is not measurable with a format-sensitive scorer, because the instruction determines the response format; we report it as an open contrast rather than a number.
[IR-3] From Retrieval to Weights: Parametric Individualization of Small Language Models with Individual Text Corpora
链接: https://arxiv.org/abs/2609.10155
作者: Christoph Wigbels,Ali Abusaleh,Markus T. Jansen,Alexander Mehler,Markus J. Hofmann
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:We approach a cognitive simulation perspective on episodic and semantic memory in multiple-choice question answering by incorporating text from individual text corpora (ITC) into retrieval-augmented generation and DoRA fine-tuning. We web-crawl the search histories of 515 participants who answered 36 multiple-choice knowledge items and analyze a stratified subsample of 150 participants. For each participant, one DoRA adapter consolidates their ITC into a small language model (SLM) whose baseline correctness falls below the participants’ lowest quartile. The adapter measurably writes the ITC into the weights: it fits its own participant’s held-out text better than other participants’ texts (dz =1.27), an individuality effect that increases with ITC size in rank order. On the generalized knowledge test, however, the adapter adds knowledge rather than alignment with the individual: log-loss match improves, whereas match accuracy under a bias-corrected PMI readout does not, and retrieval adds nothing on top. Our results demonstrate that ITCs can be consolidated into the weights of SLMs, an encouraging basis for individualized tutoring agents, and we discuss how to move from there toward a realistic simulation of episodic and semantic memory at the individual level.
[IR-4] Guaranteeing Faithful Evidence Extraction in Speculative Retrieval-Augmented Generation
链接: https://arxiv.org/abs/2609.10046
作者: Quentin Signé,Mohand Boughanem,Jose Moreno,Thiziri Belkacem
类目: Information Retrieval (cs.IR)
备注:
Abstract:Large Language Models (LLMs) are increasingly used as interfaces for information retrieval, but they remain prone to hallucinations and faithfulness errors, in which the generated answers diverge from the retrieved evidence. While Retrieval-Augmented Generation (RAG) and recent hybrid or semi-extractive approaches mitigate this issue, they do not guarantee that quoted or extracted spans are verbatim from the retrieved context. This limitation can have severe consequences in safety-critical domains, where answers must exactly match certified documentation. We introduce Constrained Hybrid Decoding (CHyD), a novel faithfulness-first paradigm for speculative RAG. While traditional speculative decoding is optimized for inference speed, CHyD repurposes this architecture to ensure faithful verbatim evidence extraction when the extraction mode is correctly triggered. Our approach enforces hard decoding constraints that restrict generation to continuous spans present in the retrieved documents. This design provides a robust but straightforward guarantee: any explicitly quoted span in the output appears verbatim in the provided context. We evaluate our method across state-of-the-art LLMs on diverse abstractive, extractive, and semi-extractive QA benchmarks, including technical datasets motivated by aircraft maintenance. Results show that existing hybrid methods frequently hallucinate quoted spans, with exact extraction accuracy dropping below 40% in technical domains. In contrast, our approach achieves near-perfect extraction faithfulness regardless of the model used. Although enforcing hard constraints introduces a trade-off with fluency-oriented metrics, our method improves exact answer correctness and remains competitive overall, highlighting its suitability for safety-critical information retrieval applications. Subjects: Information Retrieval (cs.IR) Cite as: arXiv:2609.10046 [cs.IR] (or arXiv:2609.10046v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2609.10046 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-5] Purchase Advice and Observable Buyer Responses in Real AI Conversations
链接: https://arxiv.org/abs/2609.09878
作者: Benjamin Tannenbaum
类目: Information Retrieval (cs.IR)
备注: 16 pages, 6 figures, 4 tables. Exploratory observational audit using Aiso’s proprietary research database; text-free annotations and reproducibility code included as ancillary files
Abstract:How often does a generative assistant persuade someone to buy, or persuade them not to buy? Conversation logs contain recommendations, but they do not necessarily record subsequent decisions. We audit 317 historical interactions from Aiso’s proprietary research database of licensed, consent-based, de-identified conversations with commercially available AI assistants. Single-agent AI-assisted screening identifies 68 purchase-directed records; collapsing one shared-prefix copy yields 67 retained episodes, dated April 2023 to July 2025. Assistant responses provide candidate options, acquisition channels, or conditional preferences in 52 episodes (77.6%). One episode contains conditional redirection away from a named accommodation candidate. No episode is coded as advice to abandon or defer the purchase category. Only 18 episodes (26.9%) contain a subsequent user turn within the same purchase-related mission, compared with 23 (34.3%) that contain any later user turn. Using conversation depth alone therefore overstates this follow-up availability by 27.8%. Across 47 retained user follow-up messages, no explicit post-advice purchase commitment, completed-purchase report, or purchase-category abandonment statement is observed. These zeros describe recorded statements, not conversion or persuasion rates. The paper supplies operational definitions, text-free annotations, and reproducible descriptive results. Its central finding is a measurement limitation: recommendation content is observable much more often than a buyer’s subsequent decision. The selected historical sample, unvalidated AI annotations, and missing transaction outcomes do not support a population-level or causal estimate of persuasion.
[IR-6] When Does Low-Bit Quantization Preserve the Decisions of Vector Search?
链接: https://arxiv.org/abs/2609.09854
作者: Wenxuan Xiao,Xu Cao
类目: Databases (cs.DB); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: JMLR-style preprint with theoretical and experimental appendices
Abstract:Low-bit quantization can achieve high recall on some vector representations and fail sharply on others, while average distortion and global rank correlation do not explain the difference. We study quantized vector search at the level of the comparisons consumed by ranking and graph-pruning algorithms. Our first result is a distribution-free decomposition: the probability that a comparison flips is bounded by the probability mass of exact margins near zero plus the tail probability of the calibrated residual. We then account for dependence between residuals that share a query or graph node, and derive covariance-aware second-moment identities and tail bounds under a joint MGF proxy. For a frozen candidate permutation, we prove a deterministic coupling theorem for Vamana neighbour selection: the approximate replay returns the exact neighbour list exactly when all candidate-level pruning actions agree on the frozen exact states. We connect these results to representation geometry through an exact Gaussian oracle, establish a strict correlation gain from a deterministic magnitude bit in an aligned bilinear model, and give a rare-contamination construction showing why marginal Gaussian diagnostics do not imply the required residual tails. When analytical assumptions are unavailable, a held-out block certificate bounds the selective failure risk of a frozen quantized rule. Across learned, classical, and synthetic embeddings, standardized exact margins predict held-out ranking and pruning flip rates substantially better than global rank correlation. The framework applies to coordinate binary codes, RaBitQ, Lucene BBQ, and product quantizers through a common decision interface.
[IR-7] Should I Be Polite to My LLM Relevance Judge? Tone as a Severity Operating-Point Shift RECSYS’26
链接: https://arxiv.org/abs/2609.09703
作者: Tian Zhang,Meng Li
类目: Information Retrieval (cs.IR)
备注: 3 pages, 2 figures, 2 tables. Accepted at the 20th ACM Conference on Recommender Systems (RecSys '26), Reproducibility and Practice Notes track. Code, prompt variants, and collection pipeline: this https URL
Abstract:Large language models are increasingly used as relevance judges, yet their labels can shift with prompt surface form. We study one such feature – tone – on 3,498 TREC DL19/DL20 query-passage pairs, across eight judge models, five classifier-calibrated politeness levels, and three paraphrases per level. Effects are strongly model-dependent: one judge shows a structured U-shaped response, whereas most show only small changes. Where tone changes agreement, the results are more consistent with a shift in the judge’s severity operating point – its overall scoring leniency – than with improved judgment. Agreement rises or falls as this shift moves the judge toward or away from human annotators’ strictness. A query-disjoint cross-fit retains the expected association (Spearman \rho = -0.683 ; exact model-block permutation p = 0.019 ). Tone affects calibration-based agreement more than ranking outcomes: across 32 model-tone contrasts, the largest absolute mean change in NDCG@10 is 0.011, although Kendall’s \tau as low as 0.743 shows that reordering is reduced, not absent. The account reconciles prior contradictory findings and identifies prompt tone as a potential validity threat when absolute relevance labels matter.
[IR-8] Extracting Semantics from Cattle Reporting Categories for Data Interoperability and Findability
链接: https://arxiv.org/abs/2609.09381
作者: Kassy Raymond,Andrew Hamilton-Wright,Deborah Stacey
类目: Information Retrieval (cs.IR)
备注:
Abstract:Livestock population data disaggregated by age, sex, and production are important inputs to calculations and models that inform our understanding of global health, yet these data are fragmented across disparate sources. Bridging data siloes to improve the findability of data requires interoperability. Conventional approaches to improving the findability and interoperability of data include indexing standardized metadata. However, creating metadata is time and resource-intensive and is often difficult in domains such as livestock, which lack standards that address the needs of broad user groups. When metadata exist, they typically need to be standardized against a pre-existing vocabulary, ontology, or thesaurus, requiring a technique known as `crosswalking’. To overcome issues in the absence of metadata, the lack of standards, and the resource-intensive solutions that currently exist, this study uses a bottom-up approach. By leveraging real-world reporting categories in datasets, the composition and semantics of terms already present in the data were extracted and analyzed. Using cattle data as a pilot, we find the age, sex, and production modifiers present across cattle terms from five datasets from four data sources capture granularity not present in AGROVOC, the largest agricultural vocabulary in the world. We discuss how the composition and semantics of these terms can be used to improve the interoperability and findability of data without first requiring metadata to be generated or standards to be created. Rather than forcing datasets to conform to an existing vocabulary, this approach uses the semantics embedded in terms already present in datasets, allowing systems to make data more discoverable and interoperable while maintaining culturally and dataset-specific terminology.
[IR-9] High-probability guarantees for linear accessibility in feature superposition
链接: https://arxiv.org/abs/2609.09556
作者: Enrico Vompa
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Machine Learning (cs.LG); Probability (math.PR)
备注: preprint
Abstract:Neural networks can leverage feature superposition to encode more concepts than dimensions, but cross-feature interference constrains the linear accessibility of simultaneously active features. By framing linear accessibility as a compressed sensing problem, we derive high-probability bounds for fixed supports under subgaussian noise, proving the sufficient dimension scales linearly ( d=O_\varepsilon(k \log m) ) rather than prior worst-case quadratic limits. We then validate these bounds across system parameters through Gaussian-tail approximations. These results quantify the geometric constraints of the linear representation hypothesis, providing a framework for evaluating sparse autoencoders, compositional generalization, and neural interpretability.
人机交互
[HC-0] A Confidence-Aware Multimodal Fusion Framework for Industrial Human-Robot Collaboration
链接: https://arxiv.org/abs/2609.10339
作者: Xinyu Liu,Qiqi Dong,Boya Jia,Yi Zhang,Binbin Lian
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC)
备注:
Abstract:A confidence-aware multimodal fusion framework (CAMF) is proposed to realize reliable human intention prediction for industrial human-robot collaboration. This framework fuses four heterogeneous modalities including object 6D pose, gaze, skeletal motion and IMU-based hand motion. It embeds a confidence-trend-driven dynamic fusion mechanism into BiLSTM to adaptively balance bidirectional temporal features according to real-time modality reliability. A confidence-guided balanced learning strategy combined with a confidence freezing mechanism is further adopted to adjust network gradients dynamically, suppress noise from low-quality modalities and mitigate cross-modal learning bias. A physical platform based on the UR3 collaborative robot is built for experimental validation. Comparative results show that the proposed method reaches an intention recognition accuracy of 91.86% and outperforms existing multimodal fusion approaches in overall performance and stability. It also maintains satisfactory accuracy under low light and partial occlusion interference. In practical assembly tasks, the framework enables proactive and stable human-robot cooperation with strong environmental adaptability.
[HC-1] meCues Studio: A Workspace for Music Annotation and Algorithm Prototyping
链接: https://arxiv.org/abs/2609.10338
作者: Sapir Caduri,Yoav Goldberg
类目: ound (cs.SD); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG); Multimedia (cs.MM)
备注: 8 pages, 2 figures, to appear in Proceedings of the 34th ACM International Conference on Multimedia (MM '26)
Abstract:Multimedia applications require precise music annotation-labeled positions, segments, or loops-placed by hand or algorithmically. Machine-learning algorithms are scalable and effective but need annotated training data, scarce for many tasks. TimeCues Studio is an open-source workspace where algorithm-development teams annotate a music corpus, compare detection algorithms against those annotations, and prototype new ones. Unlike existing tools built for a single track at a time, TimeCues targets teams annotating whole collections, tightly integrated with algorithm development. Annotators place several marker types-each supporting ambiguity-aware labeling-on a grid-locked timeline that visualizes many music features, including separated audio stems. The same timeline drives an algorithm-comparison engine with bundled baselines, a Python sandbox for prototyping new models, and an ambiguity-aware evaluator that honors the structured fields. The same visualization suits solo annotators on music-sync projects. TimeCues is MIT-licensed and deploys via one Docker Compose command.
[HC-2] Senseful Consense: Towards Simplified Cookie Banners using Plain Language
链接: https://arxiv.org/abs/2609.10271
作者: Minela Bećirović,Ha Dao,Mannat Kaur,Martin Johns,Alexandra Dirksen
类目: Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注: 14 pages, 3 figures
Abstract:While the GDPR and ePrivacy Directive mandate that consent information must be clear and accessible, most modern cookie banners remain obscured by technical jargon, vague phrasing, and frequent content overload or underload. This feasibility study investigates the impact of applying plain language (Einfache Sprache) to cookie banners within the IAB Transparency Consent Framework (TCF). In our study, we analysed cookie banner texts from 200 websites, using AI-based mapping to categorise extracted content into standardised processing purposes. By substituting complex legal terms with simplified descriptions, we successfully demonstrated that the comprehension barrier can be lowered from a college-graduate level to a 7th-grade level. However, the effectiveness of plain language is inherently constrained by the informativeness of the original content; it cannot compensate for banners that omit legally required details. We conclude that while plain language is a vital tool for digital accessibility, it must be paired with standardised implementation guidelines to ensure that cookie banners are both readable and informative.
[HC-3] Seeing the Voice Preserving the Self: A Participatory Design Approach to Deaf-Centric Text-to-Speech
链接: https://arxiv.org/abs/2609.10199
作者: Shela Atemnkeng,Patrick Boudreault,Paige DeVries,Lloyd May,Christian Vogler
类目: Human-Computer Interaction (cs.HC)
备注: Accepted for publication at ACM ASSETS 2026
Abstract:We describe a participatory design approach toward developing Deaf-centric text-to-speech (TTS) technologies. While TTS is growing rapidly in the mainstream, it has received little attention to date in the deaf and hard of hearing (DHH) technology space. Critical problems have remained unaddressed for DHH users, including the ability to manipulate tone, emotions and delivery via non-auditory means. Verifying that the generated speech matches intent and is appropriate for a given situation without having to listen to it is another challenge. Respecting cultural and identity factors in the generated speech is also important. This work explores the design space with DHH participants through two focus groups, three co-design sessions, and four one-on-one early-stage design evaluation sessions. Participants included people both familiar and unfamiliar with TTS, as well as DHH content creators. We describe key findings, design ideas, results, and implications for future Deaf-centric TTS development. We also identify unmet technology requirements that pose barriers to adoption of Deaf-centric TTS technology.
[HC-4] Streaming P300 Acquisition and Statistical Signal Validation Across Five EEG Platforms: A Hardware-Agnostic BrainFlow/LSL Pipeline
链接: https://arxiv.org/abs/2609.10047
作者: Isabella Guan,Rui Liu,Fusheng Wang
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:P300 spellers offer people with severe motor impairment, such as ALS, an effective communication channel and remain one of the most established surgery-free alternatives to intracortical interfaces. Advanced language models have made spellers faster and more robust, yet the hardware beneath them is under-studied. We present a hardware-agnostic, real-time P300 acquisition pipeline built on BrainFlow and Lab Streaming Layer (LSL) that runs unchanged across consumer- and research-grade EEG headsets, with permutation tests of signal separability. Using a standard 6 x 6 row/column paradigm, we piloted five configurations: a custom dry system, a custom wet/gel system, Emotiv Flex, Emotiv EPOC X, and Muse 2. The custom systems and EPOC X showed weak or inconsistent signal separability, Muse 2 had the highest acquisition reliability despite limited centro-parietal coverage, and Flex showed the most promising signal. In 20 further Flex sessions varying subject, timing, and phrase length (131 target characters), a peak-amplitude permutation test and a cross-validated xDAWN decoder both detected a significant target response under two channel-exclusion policies, with decoder AUC reaching about 0.72 after 15 repetitions. Character accuracy depended heavily on evaluation methodology: in-sample majority voting reached 94.7%, whereas character-held-out accuracy was 31.3% with evidence accumulated across repetitions, about three times that of held-out majority voting. These analyses indicate that Flex captured a detectable, if still weak, P300 under the studied conditions, while broader participant-level validation and improved decoding remain necessary.
[HC-5] Somatosensory Activation and Attentional States in Creative Making
链接: https://arxiv.org/abs/2609.09960
作者: Katherine Rees
类目: Human-Computer Interaction (cs.HC)
备注: In Proceedings of The First Reflection in Creative Experience (RiCE) Workshop (RiCE W1) arXiv:2607.24558
Abstract:The methods for capturing the creative process come with associated tensions around memory recall, articulation, and communication during the act of making, as well as how to record these considerations. This paper has a twofold purpose: first, to offer an example of a mixed methodology, drawn from dance anthropology, sensory ethnography, and design, that applies embodied methods as an alternative for documenting creative making. Specifically, this incorporates the researcher-as-participant and the collation of fieldnotes, embodied knowledge/movement recall, with notation forms, and participant interviews. These are existing methods in dance anthropology; however, using them alongside exploratory prototyping and workshop approaches broadened this work into transdisciplinary practice. Second, it discusses the activation of somatosensory systems through wearable technology and the facilitation of heightened sensory awareness for the practitioner, leading to a subsequent ability to focus on creative decisions linked to reflection and metacognition.
[HC-6] Pairit: A Platform for Live Experiments on Human-AI Collaboration
链接: https://arxiv.org/abs/2609.09789
作者: Harang Ju,Sinan Aral
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: 15 pages, 3 figures
Abstract:Organizational design in the era of artificial intelligence requires experimental methods that can test how human-AI groups coordinate, delegate, and make decisions. Programmable platforms coordinate live human-to-human sessions or real-time human-AI chat, but researchers cannot easily declare experiment protocols in which AI participants both communicate and act on shared work within one auditable configuration. Here we introduce Pairit, an online platform that facilitates the design, testing, and deployment of experiments that test human-AI organizational designs and interventions. Through a single YAML configuration file, researchers declare an executable experiment graph (pages, routing, randomization, matchmaking, chat, shared workspaces, server-hosted agents, surveys, timers, and custom HTML components) and combine any number of humans and AI agents in live sessions. We have validated the feasibility of the platform through multiple live deployments, including peer-reviewed published studies, capturing high-resolution process traces of communication, negotiation, and collaborative work in live human-AI dyads. By representing complex interactive protocols as standardized, auditable configuration files, Pairit provides reusable infrastructure for specifying, deploying, and sharing live human-AI organizational experiments.
[HC-7] How Far Do Capability Cues Travel? Anthropomorphism and Differentiated Trust in a Platform-Embedded AI Assistant
链接: https://arxiv.org/abs/2609.09713
作者: Chenchen Mao,Hanjing Shi,Haiyan Jia,Dominic DiFranzo
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Visible AI capabilities need not translate into broader judgments of trustworthiness. In a randomized 2 x 2 experiment with 270 U.S.-based Reddit users, an embedded assistant displayed one or three functions, with or without a brief rationale. Displaying three functions increased perceived multifunctionality; no other randomized main effect survived correction across the six outcomes. Rationale availability did not reliably increase perceived intelligence. Exploratory analysis indicated stronger uptake of the functional display at higher objective AI literacy. Among concurrently measured judgments, perceived multifunctionality was associated with perceived intelligence, which was associated with anthropomorphism and all three trust dimensions. After accounting for perceived intelligence, anthropomorphism was positively associated with benevolence, but not reliably with integrity or ability. These findings separate interface effects from relationships among users’ perceptions and show why ability, integrity, and benevolence should be evaluated separately.
[HC-8] AppetiteCheck: Feasibility of Momentary Vagus Nerve Stimulation as an Implicit Intervention for Eating Behavior
链接: https://arxiv.org/abs/2609.09700
作者: Tan Gemicioglu,Jas Brooks,Pedro Lopes,Tanzeem Choudhury
类目: Human-Computer Interaction (cs.HC); Emerging Technologies (cs.ET)
备注: 28 pages, 9 figures. Published in Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, Volume 10, Issue 3, 2026
Abstract:Overeating and emotional eating are common health issues that affect people even without an eating disorder. The vagus nerve plays a critical role in the gut-brain axis, and implanted vagus nerve stimulators have been associated with reduced appetite. In this paper, we propose transcutaneous cervical vagus nerve stimulation (tcVNS) as a ubiquitous system to provide immediate, low-effort intervention during an eating episode. In a study with 24 participants, we evaluated a mobile, handheld tcVNS device during a single episode of distracted snacking. We found that participants ate 9.6% less and 23.6% more slowly during vagus nerve stimulation than during sham stimulation. Post-snacking satiety was the same in both conditions, while heart rate was lower during vagus nerve stimulation. The stimulation was described as subtle and barely noticeable. Overall, these results provide evidence for the feasibility of non-invasive vagus nerve stimulation as a low-attention intervention for managing eating behavior – one that can be packaged inside ubiquitous interactive systems. As such, we extend the design space of implicit interfaces toward physiological intervention, motivating future ubiquitous systems that pair eating-related sensing with low-attention interventions.
[HC-9] Beyond Top Words: MonoTM for Topic Modeling with Interpretable Monosemantic Features AACL
链接: https://arxiv.org/abs/2609.09575
作者: Una Joh,Bei Yu
类目: Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注: Accepted to appear in the Proceedings of AACL-IJCNLP 2026
Abstract:Topic models summarize large text corpora, but top-ranked words often provide only a limited representation of topic semantics. Sparse autoencoders (SAEs) offer a way to move beyond word-level descriptors by extracting interpretable features from dense representations, yet how feature interpretability relates to topic-inference quality remains unclear. We introduce \textbfMonoTM, an interpretable topic modeling framework that decouples these roles. Across three benchmark corpora, we show that document–topic mixture estimation and semantic interpretation favor different SAE configurations and feature subsets. MonoTM estimates mixtures from the full SAE bag-of-features representation and, with them fixed, learns topic descriptors over a separate vocabulary of corpus-grounded semantic features. This design preserves global topic structure while representing topics with semantic units more meaningful than individual words, making them more useful for downstream corpus analysis.
[HC-10] he Mutations of Machine Speech
链接: https://arxiv.org/abs/2609.09496
作者: Mauricio Figueroa
类目: Computation and Language (cs.CL); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注:
Abstract:Algorithmic outputs now populate the digital environments through which contemporary life is organized. The role of law in facilitating and constituting (rather than merely responding to) these processes is gaining increasing traction across scholarly accounts. This inquiry traces the evolution of algorithmic outputs attending to their legal underpinnings and social implications, surfacing the mutations of machine speech. The first mutation redefined speech as data to be queried: search engines transformed the web from a space of information retrieval into an economic regime of algorithmic visibility. The second mutation reframed speech as engagement: social media platforms fused moderation with amplification, turning expression into a metric of attention, governed by corporate architectures. The third mutation emerges in conversational systems and interfaces, where generative text displaces information retrieval, bringing with it dense technolegal entanglements and profound epistemic consequences. Scholars of freedom of expression, informational privacy, and communication studies have long grappled with these dynamics, yet their implications for broader legal thought have also become urgent. This piece seeks to organize and clarify the evolving debate around algorithmic speech, making this critical but often fragmented discourse more accessible to wider legal and interdisciplinary audiences. In doing so, it bridges the gap between observing technological transformation and critically assessing the constitutive role of law within it, offering a conceptual resource for researchers, students, policymakers, and practitioners navigating and contesting this evolving landscape. Subjects: Computation and Language (cs.CL); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC) ACMclasses: K.4 Cite as: arXiv:2609.09496 [cs.CL] (or arXiv:2609.09496v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.09496 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.1007/978-3-031-87993-7_188-1 Focus to learn more DOI(s) linking to related resources
[HC-11] Integrating Multi-Source Feedback in Computational Design
链接: https://arxiv.org/abs/2609.09483
作者: Francisco Erivaldo Fernandes Junior,Thomas Langerak,Mira Keränen,Danqing Shi,Ardak Alipova,Antti Oulasvirta
类目: Human-Computer Interaction (cs.HC)
备注: Accepted for publication in ACM Transactions on Interactive Intelligent Systems (TiiS)
Abstract:In real-world design practice, evaluations rarely rely on a single source of judgment. Designers routinely combine expert opinions, empirical studies, and computational models, each with distinct strengths and limitations. While machine learning offers methods to integrate multiple feedback sources, these approaches remain largely inaccessible to designers without technical expertise. In this paper, we explore how to integrate multiple feedback sources, primarily through: (1) a practical approach for multi-source integration, and (2) its implementation in MUSE, a no-code tool that allows designers to combine and balance diverse sources. Our technical findings show that independent modeling of multiple evaluation sources enables exploration across heterogeneous feedback, accommodates different evaluation speeds, surfaces disagreements between sources, and supports an adaptable evaluation setup that designers can reconfigure during their process. In a visualization design study, participants navigated their own judgments alongside simulator feedback, reporting a perception of enhanced confidence and flexibility. Our results highlight the viability of multi-source integration to support computational design, offering a step toward bridging the gap between advanced optimization methods and design practice.
[HC-12] Exploring 3D Glyph Physicalizations for Public Engagement through River Health
链接: https://arxiv.org/abs/2609.09472
作者: Maria Teresa Ortoleva,Min Chen,Rita Borgo,Alfie Abdul-Rahman
类目: Human-Computer Interaction (cs.HC)
备注: 6 pages, 2 figures, 1 table. Accepted for publication in the Visualising Climate 2026 Proceedings. Conference: 4-6 Nov 2026, Bologna, Italy
Abstract:Introduction: In this paper, we present the preliminary design of a toolkit for making glyph-based physicalizations for public engagement. We use London river health data as a case study: a data set of significance to urban issues related to climate change and of interest to draw public attention, as part of the Greater London Authority’s strategies. Design: We present the components of a 3D glyph-making toolkit, its encodings, and a step-by-step process for crafting a physicalization of a river’s water quality using recycled materials. We reason about how users can use the template to learn about a data set while reflecting on the data’s significance to their personal experience and self-mapping onto the physicalization. Reflection: We reflect on the opportunities that extending the design space of glyphs to 3D physicalization offers for supporting public engagement with complex, multi-dimensional data sets, scaffolding cognitive processes, and self-reflection, thereby bringing crucial environmental data to life. Conclusion: Future implementation of the 3D glyph template will enable the public of all abilities to explore river health data, physicalize complexity, and realize its relevance. We hope that its use in public engagement workshops will help raise awareness, invite care, and foster a sense of belonging.
[HC-13] “Its Like Drinking from a Fire Hose”: Understanding and Characterizing Video Learning Experiences for Individuals with ADHD
链接: https://arxiv.org/abs/2609.09443
作者: Hanxiu ‘Hazel’ Zhu,Weiyu Zhang,Ru Wang,Yuhang Zhao
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Video lectures have become increasingly prevalent for education and professional development, yet their static visuals, dense information, and long duration pose attentional challenges for individuals with ADHD. While adaptive learning offers opportunities towards ADHD-accessible video learning, little is known about how to suitably adapt such videos: What components in multimodal video lectures are challenging for ADHD viewers? How do these experiences surface in behavioral signals to trigger an adaptation? What presentations do they prefer? To answer these questions, we conducted an eye-tracking-based retrospective think-aloud study with 16 participants with ADHD, who watched and reflected on a curated set of video lecture segments. Our study uncovered video design elements that hindered learning and revealed participants’ coping strategies along with their limitations. By jointly analyzing behavioral signals and retrospective reflections, we characterized how these experiences manifested in behavioral patterns. We further surfaced participants’ practices for addressing learning needs beyond the video watching process, and derived design implications for future ADHD-friendly adaptive video learning systems.
[HC-14] Agent ic Web Accessibility Auditing: Authoring and Evaluating Per-Criterion Worker Agents for WCAG
链接: https://arxiv.org/abs/2609.09379
作者: Arjun Mishra,Pranav Karthik,Byungjun Bae,Dongwook Yoon
类目: Human-Computer Interaction (cs.HC)
备注: Preprint. 26 pages, 7 figures. Supplementary material and supporting research records included as ancillary files
Abstract:Automated accessibility assessments differ in the evidence they collect and the requirements they address. We present a framework that combines shared browser tools with criterion-specific worker agents, implementing 39 WCAG 2.1 Level A and AA criteria and one additional WCAG 2.2 criterion. We analyze archived predictions on 250 page-criterion records derived from professional audits of scholarly platforms. Workers recover 0.86 of positive reference labels, compared with 0.36 for axe-core and 0.67 for an uncued vision-language model, with lower precision. Criterion-level results, abstentions, development-exposure sensitivities, and separately instrumented runs qualify these comparisons. Inferred negative labels and differences between evaluated configurations limit conclusions about true accuracy and causal effects. We contribute the framework, its criterion-specific implementation, and an evaluation account that distinguishes detection, evidence availability, and resource use, motivating further study of inspectable automated assessments within professional auditing.
[HC-15] Echoes in the Algorithm: Analyzing the Fidelity of User Preferences Against Realized Platform Reach
链接: https://arxiv.org/abs/2609.09365
作者: Emelia Hughes,Tim Weninger
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:What does popular content look like when platforms withhold the usual cues? On TikTok, users still form impressions about which videos are taking off even when likes and view counts are hidden, delayed, or pushed to the margins of the interface. We study this problem through TokOrNot, a web-based game in which participants compared pairs of TikTok videos and reported (i) which one they preferred and (ii) which one they believed had reached a larger audience. We benchmark these judgments against verified public view counts, which we use as a bounded proxy for realized platform reach. Across 3,513 judgments from 363 participants, participants identified the higher-reach video only modestly above chance (56.75%, 95% CI: 56.01-58.55). Preference aligned with the higher-view video at a similar rate, while preference and prediction matched in 83.48% of trials (95% CI: 83.12-85.95). Performance also varied across content categories. Taken together, these results do not suggest that users can reliably read platform success from content alone. Instead, they point to a looser and more uncertain interpretive process in which reach judgments often track personal taste or other weak heuristics when explicit popularity cues are absent. We discuss the implications for algorithmic literacy and for interface designs that reduce visible metrics without leaving users to infer reach from uneven or idiosyncratic cues alone.
[HC-16] Where Does the Human End? Creative Agency with Generative AI across Five Years of Chinese Digital Painting
链接: https://arxiv.org/abs/2609.09333
作者: Yibo Meng,Ruiqi Chen,Shuheng Cao,Weijia Zhang,Chengxi Zang
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:As generative AI enters creative work, practitioners must decide where AI assistance ends and human authorship begins. Human-agent interaction (HAI) research has examined AI as a tool, collaborator, consultant, and competitor. The longitudinal problem is how these roles are revised as systems become more capable, public, and economically embedded. We report a five-year interview study with 17 Chinese digital painters, based on annual semi-structured interviews from 2021 to 2025. Participants described recurring but non-uniform patterns of protective resistance, pragmatic task delegation, and, for some, reflective agency repartitioning. Early resistance protected observation, originality, signature, and ownership from AI. Later delegation placed AI in bounded tasks such as references, backgrounds, rough sketches, and client-facing drafts. By 2025, some participants built hybrid workflows around human-only zones, while others described fatigue, precarity, or difficulty locating a remaining human role. Peer norms, emotional climates, and production pressures shaped which delegations felt useful, acceptable, or exhausting. Copyright, authorship, and creative labor remained recurring limits on what participants were willing to delegate. We frame these accounts as longitudinal agency partitioning, the situated work of deciding which stages, responsibilities, values, and claims remain human in creative human-agent interaction. We discuss design implications for revisable agency-boundary controls, provenance scaffolds, and community-facing authorship norms.
[HC-17] Early Epistemic Settlement in AI-Assisted Writing
链接: https://arxiv.org/abs/2609.09332
作者: Han-yu Wang
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:A language model can resolve a writer’s current organizing problem while the construction needed for her own resolution remains unfinished. I call this early epistemic settlement. The supplied organization meets every demand then governing the passage, yet proceeding from it can displace work through which the writer would have changed those demands or become able to form further organizations. I distinguish the coordination needed to complete an already formable organization from construction that changes which organizations are formable in the first place. Settlement in the first case changes the relative work still required to bring available organizations to sufficiency. In the second, it can remove the need for the work through which another organization would become formable. Even when supplied resolution and continued construction leave the same visible qualification, different dependencies in the writer’s inquiry may support different later organizations. Model suggestions can also contribute to this development when writers work through them while the problem remains unresolved. In theoretical and exploratory writing, the relations developed in reaching local adequacy help determine what the writer can later defend or develop. A sound judgment that the present passage is sufficient can therefore make further inquiry dispensable before that generative work has occurred.
[HC-18] Ephemeral Feeds and Enduring Rituals: RushTok and the Formation of Event-Based Algorithmic Communities
链接: https://arxiv.org/abs/2609.09331
作者: Emelia Hughes,Tim Weninger
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY); Social and Information Networks (cs.SI)
备注:
Abstract:Each August, TikTok’s For You page turns the University of Alabama’s sorority recruitment into RushTok. We examine RushTok as an event-based algorithmic community: a collective assembled around a bounded offline ritual and sustained by recommendation. Using a mixed-methods survey (n=71) and a reflexive account of creator outreach, we ask who participates, how, and with what stakes. Findings show an ambiguous and entertainment based throughline; many called it a community (51/71) but few claimed membership (11/71). Affiliation centered on creators rather than shared practices, with parasocial attention clustering around a small set of potential new members (PNMs) and returning figures. Higher content exposure tracked with self-identification as a community member; those members commented, followed creators, and engaged across videos. Attempts to interview creators were met with silence or refusals, reflecting community boundary-work despite viral visibility. We outline implications for platform governance, including time-bounded context, graduated visibility, and aftercare.
[HC-19] Endorsement Without New Evidence: How Sequential Voting Inflates Mandates in Online Community Governance
链接: https://arxiv.org/abs/2609.09321
作者: Zihan Chen,Lei Nico Zheng,Di Zhu
类目: ocial and Information Networks (cs.SI); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注: 11 pages, 2 figures, 5 tables. Under review
Abstract:Online communities often treat large support margins in public elections as strong mandates. We argue that such margins can overstate the independent scrutiny behind a decision. Using 198,275 free-text rationales from Wikipedia admin elections, we introduce vote-text divergence, a measure that flags a decisive vote paired with a thin, deferential rationale. Divergence rises as voters arrive later, even after controlling for voter and election fixed effects. The pattern is consistent with information saturation: once prior text is accounted for, arrival order no longer predicts divergence, while accumulated prior evidence does. The effect is strongest among peripheral voters in the co-voting network. Yet divergence does not predict worse post-promotion outcomes, such as administrative activity or survival. Public tallies can therefore weaken the scrutiny signal even while selecting capable administrators: a margin may appear to reflect more consensus and support than it actually contains.
计算机视觉
[CV-0] Programmable World Model
链接: https://arxiv.org/abs/2609.10540
作者: Zheng-Hui Huang,Guixu Lin,Jiacheng Lin,Yi-Chuan Huang,Ruihan Yu,Muyao Niu,Siqi Yang,Yu-Lun Liu,Yung-Yu Chuang,Kaipeng Zhang,Zhixiang Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Homepage: this https URL GitHub: this https URL
Abstract:Recent video world models generate increasingly realistic and interactive visual experiences, yet lack reliable mechanisms for maintaining persistent world state and enforcing programmable rules over extended interactions. We introduce Programmable World Model, a framework that decouples world-state evolution from visual observation generation. An agent translates natural-language instructions into executable programs that specify entity states and state-transition rules, enabling direct control over individual entities and their interactions. A lightweight engine executes these programs to update and maintain an explicit, persistent global world state, including off-screen entities and non-visual attributes. To connect world state with visual generation, we introduce state-augmented 3D oriented bounding boxes (OBBs) as an intermediate representation. This representation, together with the target camera trajectory, is deterministically compiled into pixel-aligned spatiotemporal conditioning signals for a pretrained video model serving as the generative renderer. This design allows users to create playable games with predefined mechanics, direct control over individual entities, and persistent world state throughout gameplay. We further introduce CombatStateBench, a benchmark for evaluating programmable world models. On CombatStateBench, our method achieves 94% Count Accuracy and 98% State Accuracy, substantially outperforming existing interactive video world models while supporting coherent long-horizon generation. These results demonstrate the effectiveness of separating explicit state evolution from generative rendering for building persistent, programmable worlds.
[CV-1] Guiding Image-to-3D Generation with Test-Time Partial Observations
链接: https://arxiv.org/abs/2609.10531
作者: Jerred Chen,Simon Weber,Ronald Clark
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Image-to-3D models can generate visually compelling 3D assets from a single RGB image, but their geometry is often only loosely constrained by the available observations, limiting their use in applications that require geometric fidelity. In many real-world settings, however, partial geometric observations of the object may be available at test time. We introduce a training-free framework for incorporating such evidence into pretrained image-to-3D generative models without retraining or finetuning. To do this, we guide generation using a ray-consistent observation likelihood defined over the model’s occupancy representation, combining surface occupancy and free-space evidence. Applied to SAM 3D and its multi-view extension, our approach substantially improves geometric fidelity across different levels of observability, as well as visual quality. Our results demonstrate that pretrained image-to-3D models can effectively integrate partial geometric observations through explicit test-time guidance, complementing their learned generative priors without modifying the underlying model.
[CV-2] Precision in Rice Variety Classification using Stacking-Based Ensemble Learning
链接: https://arxiv.org/abs/2609.10524
作者: Md. Masudul Islam,Galib Muhammad Shahriar Himel,Md. Golam Moazzam,Mohammad Shorif Uddin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Rice, a staple food for a significant portion of the global population, exhibits remarkable diversity in its varieties, presenting substantial challenges for accurate identification by consumers, traders, and farmers. This complexity often facilitates fraudulent practices, such as the unauthorized mixing of rice types, which undermines quality and trust in the supply chain. Despite its critical importance, existing research falls short of providing robust and efficient methods for precise rice variety classification based on external characteristics like color, size, and texture. To address this gap, our study introduces a comprehensive rice variety identification framework designed to enhance transparency and quality assurance. We developed a stacked ensemble model tailored for rice variety classification and curated a comprehensive dataset comprising 20 rice varieties, each distinguished by unique visual attributes. The proposed approach achieved an unprecedented classification accuracy of 100%. Furthermore, we integrated our model into a mobile application, enabling even novice users to effortlessly identify rice varieties using grain images from a smartphone camera. These findings underscore the transformative potential of advanced machine learning techniques in mitigating fraudulent practices and ensuring stringent rice quality control. Our work holds significant implications for agricultural stakeholders, paving the way for automated crop identification systems and advancing precision agriculture practices.
[CV-3] Show-Harness: Just a VLM Agent Can Play Robots
链接: https://arxiv.org/abs/2609.10522
作者: Yanzhe Chen,Zechen Bai,Zhijun Cao,Wenzheng Zeng,Kevin Qinghong Lin,Yiqi Lin,Guoqiang Liang,Kevin Yuchen Ma,Qiming Huang,Mike Zheng Shou
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Project website: this https URL
Abstract:Foundation vision-language models (VLMs) exhibit broad intelligence about the world, yet translating this intelligence into robot control remains challenging. We present Show-Harness, an Embodied Harness that enables VLMs to “play” robots through a compact semantic interface linking intent to action. Show-Harness exposes discrete semantic action units that VLMs can naturally reason over, while embodiment-specific interpreters deterministically ground them into local robot actions, keeping the VLM directly responsible for fine-grained physical decisions. Through the same interface, Show-Harness demonstrates the feasibility of (1) directly unlocking closed-source frontier VLMs for zero-shot robot control, and (2) adapting small-scale open-source VLMs for low-cost deployment with just a few GPU-hours of fine-tuning. We further develop GUMI (GUI Manipulation Interface), which extends the same semantic action space to GUI-based demonstration collection, allowing humans and agents to “play” robots across embodiments without specialized teleoperation hardware. Extensive experiments show that Show-Harness-equipped VLM agents generalize robustly across tasks, embodiments, and environments, outperforming representative agentic and VLA paradigms. These results suggest that the right interface can unlock substantial embodied capability from foundation VLMs, without requiring additional model capacity or costly embodiment-specific pretraining.
[CV-4] BrainTaskonomy: Learning How to Pretrain and What to Transfer in fMRI Foundation Models
链接: https://arxiv.org/abs/2609.10518
作者: Junfeng Xia,Wenhao Ye,Junxiang Zhang,Jiayu Zuo,Mo Wang,Quanying Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Neurons and Cognition (q-bio.NC)
备注:
Abstract:fMRI foundation models increasingly aggregate heterogeneous data across brain states, cohorts, and acquisition settings, yet pretraining domains are commonly treated as a flat mixture and downstream tasks are adapted independently. We study whether measured learning relations can organize both stages without modifying the backbone. During pretraining, a lightweight Brain-DiT proxy estimates difficulty and directed facilitation across ten fMRI domains, yielding a priority-guided cumulative domain curriculum combined with high-to-low-noise timestep scheduling and joint consolidation. During adaptation, controlled first- and higher-order transfer across fifteen tasks constructs a directed taskonomy, from which budgeted integer programming (BIP) selects directly supervised source tasks and target-specific routes. The joint priority-domain and high-to-low-timestep curriculum reduces v-NMSE, PSD-NMSE, and FC-MSE by 6.5%, 16.3%, and 10.5%, respectively, relative to uniform sampling over both dimensions, and shows strong downstream performance across six in- and out-of-domain tasks. The taskonomy reveals asymmetric, target-dependent transfer, while exploratory sealed-test evaluation shows larger descriptive gains for BIP policies when higher-order route spaces are available than for matched random controls. Together, these findings support organizing fMRI pretraining and adaptation by measured learning relations rather than treating domains and tasks as independent flat sets.
[CV-5] DUET-DINO: Simultaneous Cross-View World Modeling for Latent Planning in Robot Manipulation
链接: https://arxiv.org/abs/2609.10506
作者: Nisarga Nilavadi,Ralf Römer,Moritz Reuss,Michael Krawez,Tobias Jülg,Angela P. Schoellig,Rudolf Lioutikov,Wolfram Burgard
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: Preprint, Project Page: this https URL
Abstract:Action-conditioned latent world models predict future visual representations, enabling zero-shot goal-conditioned robot planning and control. However, their predictions for fine-grained spatial and rotational actions are unreliable for full 7-DoF end-effector control. To address this gap, we introduce DUET-DINO, a simultaneous cross-view latent world model that jointly learns action-conditioned predictions from static side- and wrist-camera observations through cross-view conditioning. By exploiting complementary global scene and gripper-centric information, DUET-DINO enables latent planning over the full 7-DoF action space. Across spatially diverse reach, orientation-intensive angled-reach, and multi-goal grasp-and-lift tasks, DUET-DINO consistently outperforms single-view and independent dual-view baselines, achieving 92% success on reach, 72.5% on angled-reach, and 60.0% on lift tasks. DUET-DINO is trained from scratch on DROID and RoboArena datasets and generalizes robustly under visual distribution shifts. We further show that while V-JEPA 2 wrist-view predictions underestimate visual dynamics induced by fine-grained actions, DINOv3 predictions better capture action-conditioned scene changes, leading to stronger downstream planning. The code and model checkpoints will be open-sourced. Project page: this https URL
[CV-6] Field Converter: Geometry-Initialized Temporal Residual Refinement for World-Grounded Player Pose Estimation from Soccer Broadcasts
链接: https://arxiv.org/abs/2609.10498
作者: Simon Khan,Laurent Gajny,Jennyfer Lecompte,Sébastien Laporte
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 11 pages, 5 figures. Code available at this https URL
Abstract:Recovering 3D human pose from monocular sports broadcasts remains challenging when players must be localized in a shared metric world coordinate system rather than only reconstructed relative to their own body. We introduce Field Converter, a geometry-initialized temporal residual framework for world-grounded 3D player pose estimation from calibrated soccer broadcasts. Our method first uses camera and pitch geometry to initialize the player root through ray-ground intersection, then predicts a temporal residual correction from pose, image, camera, and geometric cues. On match-disjoint evaluation sequences, residual refinement reduces root error from 49cm with geometry alone to 14cm with a frame-wise MLP and 10cm with a TCN, while a Transformer achieves a comparable 11cm. The resulting world-space MPJPE reaches 13.2cm, and ablations show that residual prediction clearly outperforms direct global-root regression while temporal context matters more than the specific temporal backbone. Failure analysis further identifies airborne motion as the main limitation of the ground-based geometric initialization.
[CV-7] Cross-Model Agreement as a Deployment-Time Reliability Signal for Automatic Polyp Segmentation
链接: https://arxiv.org/abs/2609.10495
作者: Siddharth Gupta,Jitin Singla
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:In real-time colonoscopy, ground-truth annotations are unavailable at inference, so polyp segmentation models can fail silently. We propose Referee-Based Quality Estimation (RBQE), a reference-free framework measuring agreement between a primary segmentation model and an independently trained referee on the same image. RBQE is evaluated on a standardized 1,223-image external benchmark drawn from four public datasets, using four referee configurations chosen to separate two design axes: referee independence and architectural diversity. Using a common Agreement Dice descriptor, a same-architecture referee differing from the primary model only in random initialization already yields a useful reliability signal (ROC-AUC = 0.923), showing that independent training alone is sufficient. Cross-architecture referees improve further: SegFormer-B0 achieves the strongest performance (ROC-AUC = 0.960), significantly outperforming the same-architecture control and UNet++, and exceeding a representative Test-Time Augmentation baseline by 0.055 ROC-AUC under an identical protocol, whereas a prompt-coupled MedSAM referee underperforms despite maximal architectural diversity. Because empty-mask agreement is trivially separable, we also report a restricted evaluation excluding such cases: ROC-AUC falls to 0.876 (SegFormer-B0, 1,046 images) and 0.783 (same-architecture control, 975 images), yet RBQE’s margin over both baselines widens on this identical subset. RBQE additionally increases the mean Dice of retained predictions as low-agreement cases are progressively rejected, supporting selective prediction, and requires only one additional deterministic referee forward pass at inference. Our study therefore supports cross-model agreement as a practical, interpretable reliability framework for automated polyp segmentation.
[CV-8] Artificial Intelligence Literacy and Sustainable Development: An Ethical Governance and Development Goals Framework
链接: https://arxiv.org/abs/2609.10489
作者: Md. Masudul Islam,Mirza Niaz Morshed,Md. Shafiqul Islam
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:AI literacy provides foundational competencies that support ethical, transparent, and sustainable technological development, although higher-order capabilities such as governance, critical evaluation, and strategic decision-making extend beyond basic literacy into advanced levels of AI competency. This study positions AI literacy as a governance capacity that complements and strengthens all 17 SDGs. It introduces a six-level taxonomy of artificial intelligence reasoning and ethics that extends traditional learning models by incorporating ethical judgement and strategic foresight. This taxonomy forms the foundation of an integrated framework linking education, governance, and sustainable development. A survey of 300 participants from diverse professional backgrounds within a national context which reveals strong technical awareness but limited ethical and governance readiness, highlighting critical gaps in public capacity to manage artificial intelligence responsibly. Findings show that ethical reasoning and reflective thinking are the strongest predictors of sustainable and trustworthy artificial intelligence use. The study proposed to embed literacy-based competencies into curricula, institutional policies, and governance mechanisms to accelerate equitable and responsible progress toward sustainable development goals
[CV-9] AgroVisNet: A lightweight Convolutional Network and the BD-PlantDX Expert-Validated Benchmark for Radish Potato and Pointed Gourd Disease Classification
链接: https://arxiv.org/abs/2609.10469
作者: Md. Abdullah Mandal,Saad Ahmed,Md. Khalid Syfullah
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Automated plant disease diagnosis is increasingly deployed on farmer-held devices in regions where agronomic expertise is scarce and network connectivity is unreliable. Three obstacles limit its practical value: public benchmarks are dominated by a small set of non-native crops, region-specific datasets are rarely validated by domain experts, and the architectures that reach competitive accuracy carry parameter budgets that are unsuited to low-cost hardware. We propose AgroVisNet, a compact convolutional network trained from scratch, together with BD-PlantDX, an expert-validated benchmark of 12,432 field images spanning 12 classes of radish, potato and pointed gourd in healthy and diseased states, collected across the Bogura and Nilphamari districts of Bangladesh. AgroVisNet couples grouped bottleneck residual blocks carrying sequential channel and spatial attention with multi-scale depthwise blocks and a dual-pooling classification head, reaching 290,572 trainable parameters. On BD-PlantDX the model attains 99.52% test accuracy and 99.52% weighted F1, exceeding all six ImageNet-pretrained lightweight backbones evaluated under an identical protocol while using 8.7 to 16.8 times fewer parameters and 1.3 to 8.5 times fewer multiply-accumulate operations. Exported for deployment, the model quantises to a 0.46 MB full-integer network at a 0.22 percentage-point accuracy cost and classifies an image in 8.40 ms on a single CPU. Across five random seeds accuracy remains at 99.57 ± 0.10%, a ten-variant ablation isolates the contribution of each component, and the same architecture transfers without redesign to two independently collected datasets at 98.71% and 99.05% accuracy. Grad-CAM evidence indicates that predictions rest on lesion-bearing leaf regions rather than on background cues.
[CV-10] Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization
链接: https://arxiv.org/abs/2609.10464
作者: Andy Zeyi Liu,Haoran Sun,Lucas Baker,Randall Balestriero,John Sous
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Joint-Embedding Predictive Architecture (JEPA) world models learn a compact latent representation of the world that supports prediction and planning, but their capability to learn physics and generate physically realistic dynamics remains hitherto untested. In this work, we introduce SemiGroup-JEPA (SG-JEPA), which extends the LeWorldModel framework by supplying the parameter governing the physics to the temporal model via action-conditioning and jointly training an encoder and predictor through an autoregressive latent rollout. To evaluate the model’s ability to generalize out of distribution, we design dynamical tasks under different gravitational fields that, despite obeying the same physical law, exhibit qualitatively different dynamics, ranging from floating motion in weak gravitational fields to rapid bouncing in strong ones. In contrast to DINO-WM, SG-JEPA reduces open-loop prediction error by up to 2 times on two-dimensional datasets, and increases control success rate up to 2.5 times for three-dimensional robotic datasets, for which we train independent diffusion policies. To explain this advantage, we develop a linear feature model that separates local law-conditioned error from its recursive amplification under rollout. Guided by this model, we find that back-propagating the multi-step rollout loss into the representation trains the encoder to keep the features that the predictor can carry forward, and that those are the features the dynamics depend on, so most of the gain comes from the encoder learning better features rather than from the predictor learning better dynamics. See project page at this https URL.
[CV-11] Advanced Brain Tissue Imaging with Data-Consistent Diffusion Priors in Laminographic X-Ray Nanoimaging
链接: https://arxiv.org/abs/2609.10456
作者: Wenxuan Fang,Abraham L. Levitan,Ana Diaz,Carles Bosch,Adrian Wanner,Andreas T. Schaefer,Mirko Holler,Tomas Aidukas,Nicholas W. Phillips,Yuxin Zhang,Alexandra Pacureanu,Manuel Guizar-Sicairos,Luis Barba
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Nanoscale imaging of mammalian brains is critical for connectomics. X-ray laminography enables high-throughput imaging of extended, plate-like biological specimens. However, the tilted acquisition geometry leads to incomplete Fourier-space coverage, giving rise to a missing-cone of information. Conventional reconstruction methods cannot recover unmeasured information within the cone, resulting in artifacts that distort fine brain structures. While resolving these requires modeling 3D structure, direct 3D deep learning approaches are limited by data scarcity and computational cost. Here we introduce LUCID (Laminography with Unified Consistent Diffusion), a framework that combines multi-view diffusion priors with projection-domain data consistency. LUCID integrates complementary 3D structural information while enforcing strict alignment with the laminography forward model. On simulated datasets, LUCID substantially improves spatial fidelity and restores missing Fourier components, outperforming baseline methods. Applied to experimental laminography data, LUCID generalizes robustly despite being trained exclusively on fully sampled tomographic volumes, and effectively recovers unmeasured Fourier information.
[CV-12] Enhanced Deformable Convolution with Center-invariant Offset and Edge-aware Mask
链接: https://arxiv.org/abs/2609.10387
作者: Yixiao Li,Xiaoyuan Yang,Jin Jiang,Minghao Zou,Guanghui Yue,Baoquan Zhao,Jun Liu,Wei Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Deformable convolution networks have recently become popular for many computer vision tasks, especially for semantic segmentation, because of their exceptional capabilities in dynamic spatial modeling. However, due to the dense deformable offsets and the lack of longer-range dependencies, they can not fully adopt proper and precise deformations for feature representations. To tackle the issues, in this paper, we propose Enhanced Deformable ConvNets (EDCN) for semantic segmentation. Specifically, a novel Enhanced Deformable Convolution (EDC) is exploited in the decoder, which integrates the Center-invariant Offset Module (COM) and Edge-aware Mask Module (EMM). The COM employs larger kernels and eliminates deformations at the kernel center, obtaining offsets that are more in line with the target from richer spatial information. Concurrently, the EMM obtains the significance of image content via Sobel edge detection, then selectively applies deformations based on the content significance, minimizing unnecessary deformations associated with relatively less important information, thereby avoiding impact from less informative regions. Experiments show that EDC outperforms state-of-the-art deformable convolution variants, including Deformable ConvNets V1-V4 and Entire Deformable ConvNets, across mainstream segmentation datasets with various decoder settings. Moreover, ablation studies confirm the effectiveness of each component. In addition, visualizations illustrate that EDC enhances spatial adaptation and target focus. We further analyze the extendibility of EDC to larger kernels on the image classification benchmark. Code will be publicly released.
[CV-13] Data-Driven Risk Fields for Safer End-to-End Autonomous Driving
链接: https://arxiv.org/abs/2609.10377
作者: Yuanxin Tian,Zhiyuan Liu,Jinhao Li,Zhenhua Xu,Wenhao Yu,Jianqiang Wang
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Safety is a fundamental requirement for autonomous driving, yet existing end-to-end driving models still lack explicit risk-aware learning capacities. Existing rule-based risk models provide interpretable safety priors, yet their absolute risk scores depend on handcrafted functions, coefficients, and thresholds. Learning-based risk representations reduce part of this manual design, but their supervision often relies on occupancy-derived labels or heuristic cost values, which may not capture ego-conditioned planning risk. In this paper, we propose DRiF, a data-driven risk-field framework for safer end-to-end autonomous driving. DRiF learns a shared BEV feature with static map segmentation, dynamic risk prediction, and vehicle planning. For dynamic risk learning, DRiF converts rule-based safety priors into pairwise risk labels, and trains the risk field to preserve relative risk ordering instead of regressing handcrafted absolute scores. Experiments on Bench2Drive show that DRiF achieves competitive overall performance, with consistent improvements in driving score, success rate, and collision-related metrics. These results establish relative risk supervision as an effective way to connect explicit safety structure with end-to-end planning. The data and code will be publicly available.
[CV-14] Shape-guided Gaussian Splatting for Sparse-View X-ray 3D Reconstruction MICCAI2026
链接: https://arxiv.org/abs/2609.10376
作者: Pranav Poudel,Florence Dell’Aniello Picard,Nairouz Shehata,Frédéric Lavoie,Herve Lombaert
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at: Off-Grid: 1st Workshop on Continuous Representations and Grid-Free Methods in Medical Imaging, MICCAI 2026
Abstract:Sparse-view X-ray 3D reconstruction is essential for reducing radiation exposure, but recovering a density field from a handful of X-ray projections is severely ill-posed. Recently, 3D Gaussian Splatting has achieved state-of-the-art performance in sparse-view reconstruction by representing the volume using explicit, optimized primitives, but it requires dozens of projected views. With fewer views, reconstruction quality degrades severely since the explicit primitives are optimized freely without any anatomical information. Anatomical structures, in contrast, share similar geometry and density across a population. Their variations are bounded within a limited range that statistical shape models can capture. This paper proposes a shape-guided Gaussian splatting framework for sparse-view X-ray 3D reconstructions. Our contribution lies in driving Gaussian positions toward anatomically valid configurations, alongside atlas-based density regularization. Our method ensures anatomically consistent reconstruction and improves PSNR by 2.83 dB over a state-of-the-art Gaussian splatting baseline with as few as 5 views. Code Available: this https URL
[CV-15] PACE: Perceived-Latency-Aware Cascading Service Routing and Filler Control for QoE-Efficient Retrieval-Augmented Dialogue Serving
链接: https://arxiv.org/abs/2609.10372
作者: Lin Huang,Yujuan Tan,Weisheng Li,Lixiang Zeng,Kun Yang,Suihan Xiao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注:
Abstract:We present the PACE, a framework for retrieval-augmented dialogue serving that formalizes Perceived Time-to-First-Response (PTFR) as a QoE objective and minimizes it under quality/cost constraints. Unlike prior work on cascaded routing, semantic caching, or adaptive retrieval, PACE jointly controls which answer source composes the response and what fills the waiting window. Deployed on a humanoid-robot sales service, it combines three mechanisms: a load-adaptive cascading router, a joint path-filler controller, and volatility-aware cache admission. On 75k CarQA requests, the cascade halves pure-LLM PTFR at P95 (0.29 vs 0.53s at c16). The adaptive controller reaches 0.41s P95, outperforming RAG by 2.4 times at high load with equal quality. The filler controller cuts calls by 94% with zero conflict. Volatility-aware admission reduces stale answers from 86% to 0%. A gating rule ensures the controller never worse than the baseline, with exposure bounded by one hold period. This is the first quantification of filler-answer conflict risk in deployed services.
[CV-16] Beyond Weak Labels: Prompt-Guided Local Refinement for Weakly Supervised Water Segmentation in High-Resolution Multispectral Imagery ICIP
链接: https://arxiv.org/abs/2609.10371
作者: Muhammad Farhan Humayun,Mohammad Imangholiloo,Afifah Shah,Tomi Westerlund,Jukka Heikkonen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted for presentation at ICIP Workshop 2026 and to be published as part of the conference proceedings
Abstract:High-resolution water mapping supports environmental monitoring and related applications, but accurate pixel-level labels are difficult and costly to produce. Official hydrographic vectors provide scalable weak supervision, but they contain artifacts like boundary noise, temporal mismatch, and omissions of small water structures. We propose a two-stage framework for weakly supervised water segmentation in high resolution multispectral imagery. Stage 1 learns initial masks from rasterized vector pseudo-labels, and Stage 2 converts these masks into structured component-wise prompts for localized refinement. On a manually corrected validation set, refinement improves SegFormer-B0 from 0.9509 to 0.9535 IoU and U-Net from 0.9408 to 0.9486 IoU, with corresponding F1 gains from 0.9749 to 0.9762 and 0.9695 to 0.9736. It leads to sharper shorelines, reduced boundary spillover, and better thin-structure delineation. The results indicate that prompt-guided refinement can improve pseudo-label-based water segmentation by targeting local errors that are poorly captured by global training supervision.
[CV-17] SceneHI: High-Resolution 3D-Consistent Scene Texturing with Controllable Illumination ECCV2026
链接: https://arxiv.org/abs/2609.10363
作者: Athanasios Tragakis,Marco Aversa,Daniela Ivanova,Chaitanya Kaul,Roderick Murray-Smith,Daniele Faccio,Paul Henderson
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: ECCV 2026
Abstract:SceneHI is a framework that lifts high-resolution, illumination-aware priors from 2D diffusion models to perform 3D texture synthesis. It is the first to demonstrate that high-resolution textures, previously limited to 2D synthesis, can be generated directly on 3D objects without model fine-tuning or optimization. Designed for complex, multi-object environments, SceneHI uniquely combines 3D-consistency, high-resolution fidelity, and physically plausible baked shadows within a single generative pipeline. To enforce strict geometric coherence, we introduce an exact analytical pixel-to-texel mapping that aligns diffusion trajectories across multiple viewpoints. We utilize High-Resolution Latent Textures (HRLTs) as a persistent canvas for gradually denoised textures, while camera views perform the denoising steps in latent pixel space. This ensures a shared base texture that can be subsequently refined to high resolution without compromising multi-view consistency. Finally, a light-aware generative pass embeds realistic geometry-consistent shadows directly into the atlases, bridging the gap to production workflows. SceneHI achieves high visual fidelity while reducing generation time by 80% compared to existing scene-level methods.
[CV-18] Spot-the-shift: Evaluating Grounded Image Difference Captioning of Long-term Changes
链接: https://arxiv.org/abs/2609.10356
作者: Benedetta Liberatori,Nermin Samet,Paolo Rota,Matthieu Cord,Elisa Ricci,Andrei Bursuc,Monika Wysoczańska
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Preprint
Abstract:Long-term change understanding from images of the same place revisited over time is a challenging task with applications in map maintenance and urban infrastructure monitoring. Prior work addresses it either through pixel-level prediction or difference captioning, neither of which is sufficient to reliably measure how well models detect and describe such changes. We introduce SPOT-THE-SHIFT, a human-verified benchmark for grounded image difference captioning of long-term changes in real-world driving scenes. Our benchmark provides natural language captions and spatial masks for structural changes across each image pair. We further propose an evaluation protocol that reliably assesses models’ captioning ability, validated through human studies. Benchmarking state-of-the-art MLLMs, we find that models struggle with the fine-grained multi-image spatial capability required for this task. Finally, we develop a synthetic data generation pipeline that improves an off-the-shelf MLLM without sacrificing general capabilities.
[CV-19] Beyond One-Size-Fits-All: Sample-Adaptive Strategy Routing for Vision Token Pruning in MLLM s
链接: https://arxiv.org/abs/2609.10346
作者: Haiji Liang,Pengfei Zhou,Zhenglin Wan,Wei Wang,Yang You,Wangbo Zhao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 26 pages, 6 figures. Code will be released soon
Abstract:Multimodal large language models (MLLMs) process hundreds or thousands of visual tokens per image, incurring prohibitive inference costs. While existing vision token pruning methods mitigate this overhead, they implicitly assume that a single fixed pruning strategy can be applied uniformly across all inputs. Our analysis further reveals that ranking pruning methods by average benchmark accuracy conceals substantial sample-wise complementarity: although the average-best strategy excels overall, alternative strategies prove superior on a significant fraction of individual samples. To harness this diversity, we propose VIP-Router, a lightweight VIsion Pruning Router that adaptively selects the pruning strategy predicted to be best suited to each input at a specified pruning level. Conditioned on low-cost visual and textual features, VIP-Router identifies the most suitable candidate strategy while retaining full-token inference as an option when pruning is predicted to be unfavorable. Evaluated on a curated suite of pruning-sensitive visual perception benchmarks, VTC-Bench Group A, VIP-Router consistently outperforms the best fixed strategy baseline across all reduction ratios, achieving a 26.9% relative improvement in average accuracy, and a 22.0% relative increase in average utility after accounting for realized token cost. Crucially, VIP-Router operates in a plug-and-play manner without modifying underlying pruning algorithms or model weights, introducing trainable parameters equivalent to merely 0.017% of the backbone. Furthermore, VIP-Router proves effective across various MLLM backbones and yields consistent gains on unseen benchmarks, highlighting the potential of sample adaptive routing for visual token pruning.
[CV-20] Dimensionality Reduction for Hyperspectral Image Classification KR
链接: https://arxiv.org/abs/2609.10334
作者: Mohamed Cherifi,Ammar Mesloub,Mohammed Nabil El Korso,Tayeb Touhami,Abdennour Hacine Gharbi
类目: Computer Vision and Pattern Recognition (cs.CV); Signal Processing (eess.SP)
备注: Published in 2024 8th International Conference on Image and Signal Processing and their Applications (ISPA), Biskra, Algeria
Abstract:This paper addresses the issue of supervised classification in the context of hyperspectral satellite images. It deals with two fundamental aspects: dimensionality reduction of data and the selection of appropriate supervised classification techniques. Firstly, we delve into dimensionality reduction, a critical step in simplifying the management of hyperspectral data. The reduction aims to decrease complexity in terms of memory and computing time. We examine two commonly used methods: Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). Subsequently, we explore the selection of the most suitable supervised classification algorithms for hyperspectral images. We compare the performance of three methods: K-Nearest Neighbors (KNN), Support Vector Machines (SVM), and Random Forest (RF) using real hyperspectral data. The results highlight that the combination of PCA and RF yields the highest overall accuracy and Kappa coefficient. Comments: Published in 2024 8th International Conference on Image and Signal Processing and their Applications (ISPA), Biskra, Algeria Subjects: Computer Vision and Pattern Recognition (cs.CV); Signal Processing (eess.SP) MSC classes: 68T10, 68U10 ACMclasses: I.4.6; I.5.1; I.5.4 Cite as: arXiv:2609.10334 [cs.CV] (or arXiv:2609.10334v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.10334 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Journalreference: 2024 8th International Conference on Image and Signal Processing and their Applications (ISPA), Biskra, Algeria, 2024, pp. 1-6 Related DOI: https://doi.org/10.1109/ISPA59904.2024.10536775 Focus to learn more DOI(s) linking to related resources
[CV-21] Learning to Adapt and Calibrate: Score Distribution Alignment for Few-Shot Uncertainty Prediction in Medical VLMs
链接: https://arxiv.org/abs/2609.10333
作者: Xuan Cuong Ngo,Ngan Le
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Uncertainty estimation for medical vision–language models (VLMs) using conformal prediction has gained increasing attention due to its distribution-free coverage guarantees. However, standard conformal prediction relies on exchangeability between calibration and test data and typically requires a sufficiently large calibration set to obtain reliable coverage. These assumptions are difficult to satisfy in few-shot transfer settings, where only a small labeled support set is available to adapt a pretrained VLM to a new medical task, while an unlabeled query set is used for evaluation. Supervised fine-tuning on the support set changes the model parameters and consequently shifts the nonconformity score distribution, breaking exchangeability between calibration and query samples and leading to unreliable coverage under distribution shift. Existing transductive conformal adaptation methods often preserve validity by avoiding supervised updates. While this helps maintain conformal assumptions, it underutilizes the scarce labeled support data and limits task adaptation, which is the primary objective in few-shot learning. In this setting, conformal prediction should serve as an uncertainty estimation layer that supports the adapted model, rather than preventing adaptation itself. To this end, we propose AlignCP, a framework that reconciles supervised few-shot adaptation with conformal uncertainty estimation under non-exchangeability. AlignCP learns a reweighted calibration distribution that reduces the score-level discrepancy between the labeled support set and the unlabeled query set. By aligning the one-dimensional nonconformity score distributions, AlignCP aims to close the coverage gap induced by adaptation without requiring query labels.
[CV-22] Geometry Without Coordinates: LiDAR Diffusion as a 3D Feature Bridge
链接: https://arxiv.org/abs/2609.10322
作者: Samed Doğan,Nico Leuze,Alfred Schöttl
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Transferring the rich priors of large 2D foundation models to sparse 3D LiDAR remains challenging, as training native 3D foundation models at comparable scale is limited by data and annotation scarcity. We introduce a LiDAR-conditioned diffusion model trained on pseudo-labels from off-the-shelf 2D foundation models. The model supports multiple output modalities, including depth, semantic segmentation and instance prediction, selectable via a textual task prompt. Because the model is conditioned on LiDAR, both its outputs and its intermediate UNet features can be projected back onto the input point cloud, enabling analysis of a 3D representation learned entirely under 2D supervision. We study this representation directly in point-cloud space, explicitly excluding raw spatial coordinates to isolate feature content from projection geometry. Linear probes recover up to ~23% Mean Intersection over Union (MIoU) on 3D semantic classes, compared to ~3.5% for a matched Gaussian-noise control, indicating substantial non-trivial structure. Pairwise cosine similarity across modality-specific feature streams reveals a layered organization. Early encoder layers remain weakly aligned across modalities while individually decodable, intermediate layers converge toward a shared representation, and decoder layers re-specialize toward task-specific outputs. These findings indicate that LiDAR-conditioned diffusion models can induce structured 3D representations from 2D supervision alone, with a modality-dependent manifold that locally unifies near a shared bottleneck. This positions diffusion as a viable mechanism for transferring large-scale 2D priors into sparse 3D domains.
[CV-23] Decoupled Self-Forcing Distillation for Streaming Talking Head Generation
链接: https://arxiv.org/abs/2609.10317
作者: Yanru An,Ruiyan Wang,Wenwu Wei,Rui Bu,Qi Wang,Hongwei Hu,Zhengxue Cheng,Rong Xie,Li Song,Wenjun Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Streaming talking-head generation produces each frame as its driving audio arrives, yet fidelity and efficiency have so far pulled in opposite directions: end-to-end methods condition a video diffusion model on audio directly and achieve high quality but only at large scale, while cheaper two-stage methods generate an intermediate motion representation and trail in fidelity. We argue the cost of the former lies in the target of fusion: the video latent is dominated by identity, appearance and background, none of which audio bears on, so coupling audio to every pixel blurs detail and wastes capacity. We instead fuse conditions in a low-dimensional identity-disentangled motion space, routing audio and motion captions by their temporal granularity, and generate motion latents with a small causal autoregressive transformer that a pretrained diffusion renderer turns into video. Conditions thus control video transitively, and high fidelity no longer requires a large backbone. Streaming this decomposition needs both models to be causal, and the exposure-bias problem could be solved by self-forcing given a bidirectional teacher. But there is no such teacher in motion space. Our decoupled self-forcing distillation resolves both models under one frozen teacher: conditioned on motion, it distills the renderer into a block-causal student; unconditionally, it scores rendered rollouts against real videos, supervising motion by the video it produces. This lifts the fidelity ceiling from the motion generator onto the stronger renderer. The two models run as parallel causal streams, reaching 15.4 FPS at 1.3 s latency with no quality degradation.
[CV-24] One Loop Two Gains: Can Active Learning win the Lottery for Free?
链接: https://arxiv.org/abs/2609.10311
作者: Benedikt Tscheschner,Eduardo Veas,Marc Masana
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:The lottery ticket hypothesis posits the existence of winning tickets: sparse subnetworks that, when trained in isolation from their original initialization, match the accuracy of the full dense network. The predominant method for discovering such tickets, iterative magnitude pruning, alternates pruning with full retraining from scratch until convergence over many cycles. Similarly, deep active learning also retrains a model from scratch after each acquisition round as new labels become available. Despite this shared reliance on iterative retraining with a substantial computational overhead, the two paradigms have been studied separately. We observe that the iterative training loop inherent to pool-based active learning already provides the exact computational structure that iterative magnitude pruning exploits, and propose Improve Prune (IP), a method that integrates magnitude pruning into each active learning retraining cycle at practically no additional cost. This raises a key empirical question: can iterative magnitude pruning produce winning tickets under the non-stationary data regime of active learning? We investigate this question across multiple acquisition functions, architecture families, and image classification datasets, including an active fine-tuning scenario. Our results demonstrate that IP yields sparse, deployable models at each active learning iteration. Those match the accuracy of their dense counterparts at sparsities up to 95%, effectively obtaining winning tickets as a byproduct of the active learning pipeline. These per-iteration sparse models can address two computational bottlenecks - per-round model retraining and acquisition scoring over the unlabeled pool - that currently prevent the practical adoption of DAL on large architectures and large unlabeled pools.
[CV-25] View-Structured Conformal Prediction for 3D Gaussian Splatting
链接: https://arxiv.org/abs/2609.10307
作者: Junzheng Chu,Bin Pan,Zhenwei Shi
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:3D Gaussian Splatting (3DGS) renders novel views in real time, but an uncertainty heatmap does not certify that a rendered view meets a certain prediction coverage. We treat novel-view synthesis as structured regression and ask that, with probability at least 1-\alpha , RGB prediction boxes cover at least a 1-\beta fraction of pixels in a new view. We propose View-Structured Conformal Prediction (VSCP). It splits the pre-calibration scale into a spatial shape from the renderer and a transferable view-difficulty factor, which predicts the smallest view-wise multiplier that shape needs. A held-out quantile over views (View-CP) then gives finite-sample validity even when transferring to new scenes. The same factorization makes the analysis exact: a conformity score is the ratio of oracle to predicted view difficulty, and excess width separates into a test-side and a calibration-side term. Across 13 real scenes, pixel-pooled calibration reaches 89.9% marginal pixel coverage but only 61.4% view-event coverage at a 90% target, while View-CP reaches 91.7–92.0%. At matched coverage VSCP cuts width by 22.1% against a constant scale, and matches a ten-model ensemble’s 21.0% reduction using only one model per scene and four rather than ten rasterization passes per query. VSCP also improves on the closest single-model baseline, the 3DGS-U field, by 4.7 points ( p=0.0225 ). The view predictor transfers from bounded source families to all nine unbounded Mip-NeRF~360 scenes. There the full scale beats the constant scale with 20.7% width saving on all nine scenes. It also keeps an 18.3% saving under a different densification backbone and runs at 216–280 FPS on an RTX~4090.
[CV-26] SynThermFace: Amplifying Limited Paired Data for Visible-Thermal Face Recognition via Synthetic Data Generation BMVC
链接: https://arxiv.org/abs/2609.10303
作者: Anjith George,Adam Unal,Sebastien Marcel
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted in BMVC Workshops 2026
Abstract:Face recognition (FR) is a widely used modality for biometric authentication, but conventional models rely on visible-spectrum imagery and degrade when high-quality RGB images cannot be captured. Cross-spectral face recognition addresses this limitation by matching visible images with other modalities such as thermal imagery, enabling more reliable performance in low-light, nighttime, and unconstrained conditions. However, progress is limited by the scarcity of paired visible-thermal data, which is difficult and costly to collect at scale. We propose SynThermFace, a framework that amplifies limited real visible-thermal supervision into larger paired adaptation datasets for cross-spectral face recognition. A diffusion model is first adapted using a limited set of paired visible–thermal images and then used to generate large-scale paired visible–synthetic thermal data from existing real or synthetic visible face datasets. The generated pairs are used to adapt a pretrained visible-spectrum face recognition model into a CFR model. Unlike synthesis-based approaches that require image translation at test time, the proposed method shifts generation to the training stage and performs inference with a single forward pass through the adapted recognition model. Under the same MCXFace real-pair protocol, PACT improves over the evaluated CFR adaptation baselines, isolating the effect of the proposed adaptation objective. Training PACT on larger generated paired datasets provides additional improvements over both the unadapted model and the real-pair PACT configuration. Cross-database evaluation on the Tufts dataset provides evidence that the learned representation transfers to an unseen database. The source code and trained models will be made publicly available.
[CV-27] RACE: Trajectory-robust Admission with Evidence Ordering for Efficient GUI Agents
链接: https://arxiv.org/abs/2609.10297
作者: Yuhao Wang,Mu Qiao,Xindong Zhang,Yunzhi Zhuge,Lei Zhang,Huchuan Lu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:GUI agents accumulate high-resolution screenshots as the trajectory unfolds, increasing inference latency and memory usage. Training-free visual token pruning can reduce this cost, but cache reuse introduces a fundamental constraint. Once tokens are discarded, the corresponding visual evidence cannot be recovered without re-encoding. Pruning therefore becomes an \textitirreversible admission decision that must remain useful for unknown future targets while preserving coverage of operable regions under tight budgets. To address these challenges, we propose \textbf\method, a training-free framework for \emph\textbfTrajectory-\textbfrobust \textbfAdmission and \textbfCoverage-aware \textbfEvidence ordering. Specifically, we combine a query-independent layout-derived interaction prior with instruction relevance and feature novelty to rank visual evidence according to both potential future utility and diversity. Then, we reserve part of the budget for native visual tokens distributed across the screen, repairing missing spatial coverage without breaking the ordering. Together, these mechanisms produce a nested token order, allowing retained visual evidence to shrink monotonically across budgets while remaining reusable throughout the trajectory. Finally, our monotone KV contraction incrementally contracts retired frames into compact session state, avoiding repeated visual encoding or pruning. Extensive experiments across six GUI benchmarks and diverse models verify the effectiveness of our proposed \method under tight budgets. The source code will be released.
[CV-28] Isotropic Embedding Perturbations for Robust Vision Language Encoders ECCV2026
链接: https://arxiv.org/abs/2609.10292
作者: Hyesong Choi,Daeun Kim,Song Park,Taekyung Kim,Byeongho Heo,Sangdoo Yun,Dongbo Min,Dongyoon Han
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ECCV 2026
Abstract:Data augmentation is fundamental to training modern deep vision and multimodal models. While individual methods, such as RandAug, CutMix, Mixup, RandErase, and DropPath, offer strong regularization effects, their combined use has saturated in performance due to overlapping functionalities, and aggressive pixel-level manipulations may disrupt delicate cross-modal alignment. This saturation motivates the search for a new augmentation axis within the embedding space rather than the input space. We introduce Aether, a simple plug-in method that applies diffusion-style random perturbations in the embedding space via controlled alpha-mixing, specifically designed to provide isotropic regularization that remains semantically consistent. Inspired by feature-space perturbations in language models and image degradation in generative pretraining, Aether induces mild yet effective perturbations that smooth the representations without compromising the fine-grained structural information required for strong vision-language encoders. Across diverse architectures and across multiple recognition tasks, Aether delivers consistent gains over the advanced recipe combining CutMix, Mixup, DropPath, and RandAug—a level of improvement rarely observed with modern augmentation alternatives. Notably, Aether demonstrates superior effectiveness in multi-modal alignment, succeeding where traditional pixel-space augmentations fail by providing a stable, isotropic regularization signal that respects the integrity of the high-dimensional feature space.
[CV-29] FreqFLD: Towards All-in-One Facial Landmark Detection via Frequency Modulation
链接: https://arxiv.org/abs/2609.10278
作者: Shun Ren,Kaijie Jin,Shengkai Hu,Beihang Song,Hang Sun,Wenwen Min,Youfa Liu,Jun Wan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Recent progress in deep learning has significantly advanced facial landmark detection. However, most existing methods process features in a spatial-domain manner under a dataset-specific training paradigm, which overlooks the fact that facial landmark detection is inherently geometry-driven and sensitive to frequency variations, thereby limiting cross-dataset generalization under complex scenarios and hindering the development of a facial landmark detection model. To address this issue, we propose \textbfFreqFLD, a \textbffrequency-modulated framework towards All-in-One \textbffacial \textbflandmark \textbfdetection. Specifically, FreqFLD introduces a Frequency Modulation Module (FreqMoM) to explicitly induce the frequency prior by decoupling and modulating low- and high-frequency components, which is then injected into subsequent feature modeling to enable balanced modeling of global facial structure and local landmark details. Furthermore, FreqFLD employs a Frequency-Modulated Mixture-of-Experts (FreqMoE), with expert selection adaptively conditioned on frequency-modulated priors, enabling flexible modeling of heterogeneous facial landmark patterns under diverse and challenging scenarios. To regularize frequency-consistent modeling under the All-in-One paradigm, we further introduce a Frequency-Consistent Routing (FreqCR) loss, which constrains the routing and assignment of frequency-aware experts to promote balanced expert utilization across diverse facial scenarios, thereby enabling stable expert specialization and achieving robust facial landmark detection. Extensive experiments demonstrate that the proposed FreqFLD achieves comparable performance on popular datasets. The code is available at: this https URL.
[CV-30] When Fusion Fails: Corruption-Aware Rebalanced Fusion for Multi-Modal Medical Image Segmentation ACM-MM2026
链接: https://arxiv.org/abs/2609.10261
作者: Yuchen Pei,Xiaoyu Hu,Yixiong Zou,Dingwen Hu,Hui Chu,Yutao Ma,Shijun Qiu,Gang Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by ACM Multimedia (ACM MM 2026)
Abstract:Multi-modal medical image segmentation leverages complementary diagnostic information, yet fusion can underperform single-modality baselines when spatially aligned inputs differ in quality. Here, “corruption” primarily denotes resolution-induced degradation rather than misalignment or complete modality absence, while synthetic noise is evaluated only as an auxiliary setting. We identify a critical optimization-inference inconsistency: degraded modalities can receive weak training updates yet substantially affect predictions, indicating active interference with fusion. We attribute this failure to resampling-induced feature corruption and optimization bias, where noisy features propagate through skip connections and encourage unreliable modality selection. We therefore propose CoReFuse-Med, a Corruption-aware Rebalanced Fusion framework that suppresses corruption during feature transmission and rebalances modality contributions during high-level fusion. Experiments on EPVS, BraTS, and WMH, including multiple Z-axis slice-retention ratios and an auxiliary noise test, demonstrate improved accuracy and robustness under modality-quality discrepancies. Our code is available at this https URL.
[CV-31] UOT-Gap: A Variational Principle for the Modality Gap in Vision-Language Models via Unbalanced Optimal Transport
链接: https://arxiv.org/abs/2609.10224
作者: Zonglin Yang,Huilan Ma,Xudan Zheng,Yuejun Xie
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at PRCV 2026. 14 pages, 6 figures
Abstract:Vision-language models such as CLIP embed images and text in a shared space, where modality-specific distributions often remain separated. Existing accounts connect this modality gap to initialization, contrastive dynamics, and information imbalance, while its distributional and pairwise contributions to retrieval remain unresolved. We introduce UOT-Gap, a training-free variational diagnostic that models frozen image and text embeddings with unbalanced entropic optimal transport (UOT). The UOT optimum separates transport, coupling complexity, and marginal mass variation; a complementary pair-aware residual compares observed image-caption pairs with the UOT soft matching. On Flickr8K and COCO-1K with frozen CLIP, OpenCLIP, and SigLIP encoders, caption degradation reduces Flickr8K Recall@1 from 0.559 to 0.003. Across six dataset-model conditions, the pair-aware residual tracks retrieval degradation with mean absolute Spearman 0.973, compared with 0.392 for the mean gap. The association remains stable across five random COCO-1K subsets at 0.954\pm0.026 , with a minimum of 0.943. UOT barycentric updates reduce the transport objective while degrading retrieval, distinguishing geometric objective descent from task improvement. These results establish UOT-Gap as a diagnostic for caption quality, modality alignment, and retrieval robustness.
[CV-32] 3rd Place Solution to Human Motion Challenges in Real-World and Clinical Settings (MoCha) @ECCV2026: Language-Aligned Motion Representations for Domain-Generalizable UPDRS-Gait Severity Estimation ECCV2026
链接: https://arxiv.org/abs/2609.10187
作者: Soojie Kim,Muhammad Munsif,Minkyung Kim,Seungryul Baek
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 3rd Place Solution to the MoCha 2026 Challenge at ECCV 2026
Abstract:In this work, we introduce language-aligned motion representations for domain-generalizable UPDRS-Gait severity estimation, aiming to learn semantically structured motion features that generalize across heterogeneous clinical domains. We first learn motion representations using a Bi-GRU backbone that captures the temporal dynamics of SMPL sequences. Prior to model training, motion captions are generated offline using Qwen2.5-7B-Instruct. The backbone is then trained with both classification and text-alignment objectives to learn discriminative and semantically structured motion representations while accounting for the class imbalance present in the training data. We subsequently adapt the learned backbone independently to each source domain so that the model can capture domain-specific motion characteristics. The resulting source-specific models are then merged at the parameter level to consolidate complementary knowledge across source domains into a single domain-generalized model. To further mitigate class imbalance, we perform GPT-5.5-based pseudo labeling, and our final merged models for each site do not use any class-prior correction during inference. The resulting model is evaluated under the unseen-site setting of the MoCha Challenge, using Macro F1 as the primary evaluation metric. Our method achieves a macro-F1 of 0.57 on the hidden test set with only 637K active parameters at inference, ranking 3rd among 58 leaderboard entries in the MoCha 2026 Challenge. The challenge attracted 1,669 submissions from 112 participants and offered monetary prizes sponsored by Machine Medicine Technologies.
[CV-33] ScopeMamba-YOLO: Widening the Perceptual Scope Inward and Outward for Small Object Detection in Remote Sensing Imagery
链接: https://arxiv.org/abs/2609.10156
作者: Junjie Fan,Yijun Mai,Linduo Wei,Jiayu Rao,Junmin Bao,Qiushi Jin,Guijia Li,Yong Qi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Small object detection in unmanned aerial vehicle (UAV) and remote sensing imagery requires preserving high-resolution detail while modeling long-range context. Adding a stride-4 detection level and removing the stride-32 stage benefits tiny targets but weakens peripheral spatial support, whereas directly inserting selective scanning into the main feature path can interfere with weak local cues. We propose ScopeMamba-YOLO, built around an off-path, zero-gated selective-scanning principle that decouples contextual modeling from the convolutional stream. The principle is instantiated by a Cascaded Global-Context Module (CGCM) in the backbone and a Selective-Scan PAN (SS-PAN) in the neck. An Adaptive Multi-scale Strip (AMS) Block reduces the cost of high-resolution feature extraction, while a Scale-Adaptive DFL (SA-DFL) head reallocates distributional support and regression capacity across scales with only 0.008M additional parameters. Controlled experiments show that matched main-path selective scanning reduces mAP50 by 0.98 pp, whereas off-path CGCM improves the final configuration by 0.67 pp over the three-seed no-CGCM mean; operator controls indicate that this gain is not explained by auxiliary branch capacity alone. ERF analysis further shows that the complete context pathway increases the peripheral energy ratio from 0.008 to 0.090 at stride 8. On VisDrone-2019, ScopeMamba-S achieves 50.8% mAP50 with 3.57M parameters, exceeding YOLOv8s by 10.8 pp while using 32% of its parameters; ScopeMamba-M reaches 52.6% mAP50 with 6.48M parameters. Consistent improvements are also observed on AI-TOD, especially for very-tiny and tiny objects.
[CV-34] ransGaze-Object: Transformer Based Driver Gaze Object Prediction Framework in Real Driving
链接: https://arxiv.org/abs/2609.10139
作者: Pavan Kumar Sharma,Ayush Pande,Pranamesh Chakraborty
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 32 pages, 17 figures
Abstract:Driver gaze provides information regarding driver visual attention and situational awareness to the surrounding traffic. Existing driver gaze estimation studies represent gaze in terms of gaze zone or gaze vector/point-of-gaze (PoG). However, object-level gaze information provides a more semantically meaningful representation of visual attention by identifying attended objects, such as vehicles, pedestrians, or traffic signals. In this study, we propose an end-to-end driver gaze object prediction framework, TransGaze-Object, Transformer-based Gaze Object prediction model. The proposed framework first extracts facial features, including face and iris-weighted eye features, along with trafficobject spatial features. A transformer based cross-attention mechanism is then used to compute similarity scores and attention weights for predicting the drivers gaze object. To train this model, we propose a benchmark driver gaze dataset, Urban Driving-Face Scene Gaze (UD-FSG), comprising synchronized driver-face and traffic-scene images, scene objects bounding boxes, and gaze labels in terms of 2D gaze coordinate and gaze object. The TransGaze-Object model achieves an overall accuracy of 60% for gaze-object prediction, compared to 51% accuracy obtained from associating the estimated Point-of-Gaze to traffic objects. The error analysis reveals that TransGaze-Object reduces confusion between traffic objects (predicted) and the background (ground-truth), achieving an error rate of 11.68%, a 49.7% relative reduction compared with 23.21% error obtained from PoG-based gaze-object association. Overall, the results demonstrate the effectiveness of directly predicting gaze objects from driver-face and traffic-scene information, rather than estimating an intermediate Point-of-Gaze and subsequently associating it with traffic objects.
[CV-35] SA-Profile: Automated Sulcus Angle Profiling from Super-Resolution MRI MICCAI
链接: https://arxiv.org/abs/2609.10125
作者: Michael Wehrli,Leo Widmer,Edwin Li,Noel Fiechter,Lorenzo Pettinari,Sidaty El Hadramy,Carol C. Hasler,Philippe C. Cattin
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at MICCAI endorsed Event MICAD 2026
Abstract:Trochlear dysplasia (TD) is an abnormality of the femoral trochlea associated with anterior knee pain and patellar instability. The sulcus angle (SA) is used to assess trochlear morphology, but it is typically measured on a single axial MR slice with no clear guidance on which to select, making it sensitive to slice selection and landmark placement. We propose an automatic framework for continuous SA profiling from super-resolved MR volumes. Clinically acquired axial, coronal, and sagittal MR scans are combined using implicit neural representations to reconstruct a high-resolution volume. SA measurements are computed across the trochlear region using two landmark detection U-Net models. The approach was evaluated on the public fastMRI dataset and a small in-house cohort of patients with TD. Compared with conventional manual single-slice SA measurements, the proposed automated method yielded a mean absolute error of 11.6 ^\circ while providing continuous characterization of trochlear morphology. Population-level analysis demonstrated distinct mean SA profiles between the public cohort and the in-house TD cohort, highlighting the potential of profile-based assessment to characterize TD. By reducing reliance on a single manually selected axial slice, the proposed framework extends conventional SA assessment to a continuous profile-based description of trochlear morphology without additional imaging, while remaining conceptually linked to current clinical assessment. Further validation is required. The code is available: this https URL.
[CV-36] LinearMask-GS: Stable-Mask Importance Pruning for Compact 3D Gaussian Splatting BMVC2026
链接: https://arxiv.org/abs/2609.10095
作者: Donghun Ryu,Minhyeok Lee
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to BMVC 2026. 17 pages main paper + 17 pages supplementary material, 3 figures, 4 tables in the main paper
Abstract:3D Gaussian Splatting (3DGS) enables real-time novel view synthesis but produces millions of primitives through adaptive densification, leading to significant storage overhead. Learned-mask pruning methods such as LP-3DGS address this by assigning each Gaussian a learnable mask to identify and prune redundant primitives. However, we identify a limitation of this paradigm: the steep slope of the Gumbel-Sigmoid activation drives mask values to the extremes within the short mask-training window, before the importance ranking has stabilized, producing a sharply bimodal distribution from which that ranking can no longer be reliably recovered. We propose LinearMask-GS, which replaces Gumbel-Sigmoid with a linear increment activation that keeps mask values in a mid-confidence regime throughout mask training, producing a stable, unimodal mask distribution whose ranking tracks importance. On Mip-NeRF 360, our method achieves 3.6x and 1.6x Gaussian reductions over 3DGS and LP-3DGS, respectively, while maintaining or improving rendering quality. For outdoor scenes, it yields a 1.6x reduction (from 2.18M to 1.36M) with notable gains in PSNR (+0.38 dB), SSIM (+0.025), and LPIPS (-0.029).
[CV-37] Automatic Reproducible Camera Intrinsic Calibration
链接: https://arxiv.org/abs/2609.10082
作者: Xiangcheng Hu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 6 pages, 7 figures
Abstract:Accurate camera intrinsic calibration is fundamental to robot perception, and the accuracy depends on the quality of the collected images. However, existing target-based calibration methods often require the practitioner to manually filter out high-quality images and to specify an appropriate radial distortion order. This paper presents a fully automatic intrinsic calibration pipeline that determines both from the collected data. We adopt an iterative rejection scheme that estimates parameters on a candidate image set and removes views whose mean residual exceeds a multiple of the median. Crucially, this process runs independently under each candidate distortion order, so that the retained image set is consistent with the residual scale of that order. Further, the distortion order is selected on held-out images, with the intrinsics and distortion fixed and only the board pose re-estimated, ensuring that an added coefficient is supported by independent observations. Finally, we integrate both steps into an interactive calibration tool that supports full-pipeline data inspection and parameter estimation. Experiments on our own camera data and five public real-world datasets show that image filtering reduces the held-out reprojection error by 25%, the order selection further by 5%, achieving the lowest held-out mean among four compared configurations without manual image selection. We will release the code and data to facilitate future research.
[CV-38] Elastoformer: Enabling Dynamic Adaptivity via Elastic Model Transformation
链接: https://arxiv.org/abs/2609.10018
作者: Sudaksh Kalra,Dolly Sapra
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Performance (cs.PF); Systems and Control (eess.SY)
备注: Published at SEC’25
Abstract:EdgeAI systems are increasingly employing computer vision applications to enable intelligent, on-device decision-making in real-time. However, these deployments face highly dynamic operational conditions, with fluctuating constraints on latency, power availability, and memory resources. Deep Neural Networks (DNN), which follow fixed computational execution flows, lack the flexibility to adapt to such variability, resulting in inefficient and suboptimal performance in edge scenarios. This underscores the need for architectures that are not only efficient but also dynamically scalable at runtime. In this paper, we propose Elastoformer: A framework that transforms conventional neural networks (NN) into Elastic NN capable of real-time elastic inference. Unlike the conventional bag-of-models approach, which requires maintaining multiple independent models for different operating conditions, Elastoformer offers a single, modular solution that dynamically switches between multiple modes of operation at runtime, adapting efficiently to the changing computational budgets of edge devices without the overhead of managing separate models. Experiments reveal that our framework achieves up to 85% reduction in computation FLOPs, 50% reduction in latency and 76% reduction in memory overhead, while showcasing the architecture agnostic nature of the framework across both Vision Transformers and CNNs. Our code is available at this https URL.
[CV-39] Beyond Similarity: Foundation Models as an Efficient Backbone for Training-Free Composed Video Retrieval
链接: https://arxiv.org/abs/2609.10008
作者: Dmitry Demidov,Muhammad Zaigham Zaheer,Omkar Thawakar,Abdelrahman Mohamed Shaker,Rao Anwer
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Composed video retrieval (CoVR) searches a gallery for the target video that realizes a natural-language modification of a source clip. However, at gallery scale, this creates a fundamental tension: compact embeddings enable efficient, reusable search but can miss the transient actions, state changes, and subtle constraints that demand fine-grained video reasoning, whereas applying large multimodal models uniformly sacrifices scalability. To address these limitations, we propose that frozen foundation models should instead occupy complementary roles, with inference depth adapted to query difficulty. Based on this premise, we introduce \methodname, a framework for training-free \methodexpansion. Specifically, a composed-query embedding first searches reusable video-only gallery representations; uncertain queries undergo bounded reranking and candidate expansion; ambiguous edits trigger target-description generation; and only close leading candidates reach multimodal verification. To support these roles, frame selection, spatial resolution, and time cues are adapted to each stage. Across complete target-gallery evaluations, our method reaches state-of-the-art performance among training-free approaches, with 89.55 and 93.43 R@1 on Dense-WebVid-CoVR and CoVR-R, respectively (with more than +35% and +25% absolute margins to the closest counterpart). These results show that adaptively orchestrating foundation-model capabilities can combine scalable retrieval with fine-grained reasoning without task-specific training. The source code and all relevant guidelines are available on this https URL.
[CV-40] What Makes Adversarial Examples Transfer Across Deepfake Detectors?
链接: https://arxiv.org/abs/2609.10002
作者: Rafael M. Mamede,Pedro C. Neto,Ana F. Sequeira
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
备注:
Abstract:Deepfake detectors remain vulnerable to transfer-based black-box attacks, in which adversarial examples are generated on a source surrogate model and transferred to a target model, unknown to the attacker. Yet how source–target compatibility shapes attack success remains poorly understood. Prior studies evaluate limited detector pools and rarely disentangle architectural from training factors. We conduct a controlled evaluation of adversarial transferability across 60 detectors spanning six backbones, two pretraining regimes, and five training-data configurations, using two attack procedures: AutoAttack (AA) and the Carlini–Wagner attack with Expectation over Transformation (CW–EOT). Matched comparisons reveal significantly higher transfer when source and target share an exact backbone, architecture family, pretraining regime, or training data. This compatibility structure is attack-dependent: exact backbone compatibility has the largest effect under AA, whereas shared pretraining and training data have the largest effects under CW–EOT. When transfer is averaged across non-target sources, mean attack success rate (ASR) is 7.21% under AA and 19.52% under CW–EOT. By contrast, a multi-source oracle combining both attacks attains a (64.48%) mean ASR after excluding exact backbone and training-data matches, showing that source averaging can substantially understate target vulnerability. We release 240,000 adversarially perturbed images, complete pairwise transfer results, detector configurations, and evaluation code. These findings establish source–target compatibility and source-model selection as central dimensions of credible transfer-based black-box robustness evaluation.
[CV-41] From Few-Shot Segmentation to Clinician-in-the-Loop Medical Image Analysis
链接: https://arxiv.org/abs/2609.10001
作者: Yazhou Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 19 pages, 2 figures, 4 tables. Perspective article
Abstract:Few-shot medical image segmentation (FSMIS) seeks to delineate unseen structures from a small support set, but its standard formulation fixes task-defining evidence before inference. This assumption is fragile when query cases exhibit acquisition shift, atypical pathology, ambiguous boundaries, or poor image quality. Prototype learning, cross-domain matching, interactive segmentation, uncertainty estimation, test-time adaptation, and promptable foundation models address parts of this problem, yet have not been jointly evaluated under a common model of expert attention and clinical risk. This Perspective reframes FSMIS as a sequential clinician-model decision problem with a static support budget K and a distinct interaction budget B . At each step, a system accepts the current segmentation, requests feedback, or defers to full expert review. Queries vary in location and modality and are selected by response-conditioned net expected value of information; clinician-provided feedback informs bounded adaptation only after prespecified provenance, consistency, and safety gates. The framework separates distributional atypicality from predicted clinical failure and treats clinician responses as informative but fallible observations. We synthesize the transition from few-shot and cross-domain segmentation to interactive and selective adaptation, delineate the integration gap, and define four research directions with falsifiable hypotheses. Evaluation spans external-domain calibration, quality-effort trade-offs, reader studies, and prospective workflow assessment. The central claim is not that interaction alone resolves domain shift, but that scarce expert attention should be allocated only when it is expected to reduce clinically relevant risk.
[CV-42] Putting Captions to the Test: Evaluating Video Caption Quality through Multiple-Choice Question Answering ACL2026
链接: https://arxiv.org/abs/2609.09973
作者: Zizhen Wang,Bo Feng,Zhengfeng Lai,Shiyu Li,Yang Lu,Meng Cao,Ping Huang,Xiaoming Simon Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by ACL 2026 main conference
Abstract:Evaluating video captioning remains a critical challenge for Visual Large Language Models (VLLMs). Existing metrics primarily rely on matching generated text against ground-truth references. This paradigm suffers from the ``one-to-many’’ nature of video description, where high-quality captions are often penalized for lexical mismatches or valid shifts in visual focus. Furthermore, such assessments are typically one-dimensional, failing to provide a fine-grained analysis of caption quality. To address this, we redefine caption quality through the lens of information fidelity: A caption must maximize the coverage of salient visual information while ensuring strict factuality. We introduce CapQuiz, a novel reference-free benchmark that assesses captions based on their utility in answering human-verified, fine-grained, multiple-choice questions derived from the video. CapQuiz features a hierarchical taxonomy of 10 question types (spanning Descriptive and Inferential categories) across 24 diverse video domains. Extensive experiments demonstrate that CapQuiz correlates significantly better with human judgments than existing metrics and offers interpretable insights into model performance.
[CV-43] Multimodal Emotion Recognition in Conversations via Class-Wise Adaptive Modality Fusion and Affective Geometry ECCV2026
链接: https://arxiv.org/abs/2609.09924
作者: Oriol Marín,Roger Marí,Gloria Haro,Rafael Redondo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the 11th Workshop and Competition on Affective Behavior Analysis in-the-Wild (ABAW) at ECCV 2026
Abstract:Emotion Recognition in Conversations (ERC) requires integrating heterogeneous textual, audio, and visual cues while accounting for conversational context and emotional dynamics. We extend the Self-Distillation Transformer architecture for ERC with appearance+geometry visual representations, class-wise adaptive modality fusion, and a valence-arousal prior for affective transitions. On the MELD and IEMOCAP datasets, geometry-enhanced visual representations improve weighted F1 by 0.27 and 4.36 points over appearance-only features, respectively, while class-wise adaptive fusion provides further gains of 0.17 and 0.25 points over the original softmax gate. The valence-arousal prior yields targeted improvements of 0.30 and 0.74 accuracy points on emotionally shifted utterances while preserving performance on stable turns. These results indicate that structured facial cues, emotion-dependent modality weighting, and affective geometry provide complementary benefits for multimodal ERC.
[CV-44] Interpreting Object-Dependent Concept Brittleness in Text-to-Image Diffusion Models ACM-MM2026
链接: https://arxiv.org/abs/2609.09909
作者: Yifan Yuan,Xiangyu Liu,Hongming Shan,Yu Han,Yu Jiang,Hao Tan,Junping Zhang,Linlin Shen
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Accepted at ACM MM 2026. 27 pages, 17 figures, including appendices
Abstract:Although text-to-image diffusion models generally exhibit strong prompt-following ability, we identify a persistent and previously underexplored failure pattern in which a small subset of prompts differing only in the object consistently fails to realize the same target concept under identical generation settings. We term this phenomenon object-dependent concept brittleness. Such cases suggest systematic internal blind spots rather than random sampling noise. In this paper, we present an interpretability-oriented framework to audit and minimally correct these failures. Our key idea is to analyze denoising trajectories in a step-wise sparse autoencoder (SAE) space, where abstract style and attribute concepts become more separable than in the raw denoising representation. This sparse space enables us to compare successful and failed generations, identify concept dimensions whose evidence is missing, weakened, or temporally delayed, and construct class-level concept prototypes from reliable class-consistent samples. Based on this audit process, we introduce a lightweight inference-time correction strategy that interpolates denoising features toward the corresponding prototype in SAE space. Rather than serving as a task-specific retraining method, this intervention acts as a validation of the diagnosed concept deficiency. We evaluate the proposed framework on style and attribute failure cases across multiple diffusion backbones, with significant improvements in concept consistency, text fidelity, and repair success. Further analyses show that deeper denoising representations provide clearer concept structure, while early-stage intervention offers the strongest correction leverage. Code is available at this https URL.
[CV-45] Can We Trust Video Hallucination Detectors? VidHalLoc for Evaluating the Evaluators
链接: https://arxiv.org/abs/2609.09895
作者: Xinyu Chen,Adnan Mahmood,Mark Dras
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 29 pages, including appendices
Abstract:Video-language models and video agents can produce hallucinations that conflict with spatiotemporal evidence. Existing benchmarks mainly evaluate model hallucinations, and heterogeneous mechanisms make detector reliability difficult to compare. We introduce VidHalLoc, a benchmark that evaluates hallucination detection methods under a unified diagnostic evaluation protocol using 2,000 adversarial hallucination samples across Video Question Answering and Video Captioning tasks, spanning Ontology and Dynamic hallucination categories. To construct VidHalLoc efficiently, we introduce VideoHALO, a Harness Engineering-informed multi-agent workflow that decomposes data construction into four executable stages supported by a memory system and a communication protocol. Evaluation of fifteen methods reveals that the four dedicated detectors peak at an Overall accuracy of only 34.63%, indicating limited reliability across video hallucination types [Dataset Repository: this https URL].
[CV-46] StreetDiff: Multi-view Street Scenes Generation via Cross-view Consistent Multi-view Stable Diffusion with Structure Prompts
链接: https://arxiv.org/abs/2609.09890
作者: Qi Zhang,Yanyifan Wang,Weiyuan Zhang,Hui Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 4 figures
Abstract:Multi-view diffusion models have shown strong performance in scenes with strong geometric priors and sparse semantics, such as indoor rooms or simple outdoor environments (e.g., fields, courtyards). However, they often fail to maintain cross-view consistency under camera rotation, especially in structurally complex urban environments. Without explicit modeling of spherical correspondence across views, existing approaches tend to produce object duplication, structural distortion, and layout inconsistency. To address this limitation, we propose StreetDiff, a multi-view diffusion framework that explicitly enforces cross-view alignment during denoising. StreetDiff introduces a Panorama–Perspective Synergy design to decouple global layout reasoning from local detail synthesis, and incorporates a Panorama Alignment Module (PAM) that establishes spherical-projection-based attention constraints across views. By injecting structured alignment constraints without modifying the diffusion backbone, our framework achieves robust cross-view coherence in challenging urban street scene generation tasks. In addition, we construct Street360, a large-scale HDR multi-view urban panorama dataset. Extensive experiments demonstrate that StreetDiff significantly improves structural consistency and visual fidelity compared to prior multi-view diffusion generation methods.
[CV-47] Albedo Estimation via Latent Bridge Matching
链接: https://arxiv.org/abs/2609.09884
作者: Carme Corbi,David Serrano-Lozano,Javier Vazquez-Corral,Maria Vanrell
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accpeted at the Color and Imaging Conference (CIC 2026), hosted by the Society for Imaging Science and Technology (IST)
Abstract:Recent advances in Intrinsic Image Decomposition (IID) have increasingly relied on generative models. However, progress remains limited by three key challenges: (a) insufficient physical consistency, (b) high computational cost at inference time, and © limited generalization capabilities. In this work, we show that latent bridge matching (LBM) effectively addresses these limitations for albedo estimation. We introduce a novel LBM-based architecture that enforces physical consistency through a pixel reconstruction loss, benefits from the inherent efficiency of LBM low-cost inference, and improves generalization across diverse datasets by incorporating a shading conditioning. In this extended version, we additionally show that conditioning the shading estimator itself on the predicted albedo further improves reconstruction fidelity, and we benchmark our best model against stateof-the-art IID methods across five real and synthetic datasets.
[CV-48] CLFTv2: Efficient Camera-LiDAR Fusion for Semantic Segmentation via Hierarchical Feature Pyramids
链接: https://arxiv.org/abs/2609.09881
作者: Toomas Tahves,Mauro Bellone,Raivo Sell
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Semantic segmentation for autonomous driving requires reliable detection of vulnerable road users (VRUs) despite heavy class imbalance. We introduce CLFTv2, a hierarchical camera-LiDAR fusion framework replacing global ViT attention with a Swin-based multi-scale encoder and a lightweight FPN-style residual decoder. Operating in the 2D perspective domain, CLFTv2 integrates multi-scale geometric cues through shifted-window attention and per-scale residual fusion, avoiding the computational overhead of query-matching decoders. Across three driving datasets, CLFTv2 consistently improves VRU recall. On ZOD, CLFTv2-Large achieves 53.5% mIoU, improving pedestrian IoU from 35.5% to 44.9% over the prior CLFT model. On Waymo, CLFTv2 reaches 61.7% mIoU. Additionally, a modality-isolation study suggests ViT’s global receptive field yields stronger fusion gains only under dense LiDAR returns. Compared to a Swin-based Mask2Former adaptation, CLFTv2 requires 1.4 \times fewer GFLOPs and delivers 2.2 \times higher throughput, while achieving comparable overall accuracy. These results demonstrate that hierarchical local-attention fusion offers an efficient, scalable alternative to global-attention and query-based decoders for real-time on-vehicle perception in intelligent transportation systems. Source code is publicly available.
[CV-49] From Pixels to Hierarchical Sequences: Quadtree Mask Encoding for Vision-Language Binary Change Detection
链接: https://arxiv.org/abs/2609.09876
作者: Xiao An,Ruikang Zhang,Chen Zhong,Xuli Shen,Jiaxing Sun,Jiang Wu,Wei He
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages, 16 figures
Abstract:Dense change detection in remote sensing requires vision-language models (VLMs) to compare bi-temporal images and generate accurate pixel-level masks. Existing VLMs are largely confined to change captioning outputs, and the few that produce pixel-level masks still rely on external decoders or flat text-as-mask serialization, which are less effective for small and fragmented changes. We introduce QUAKE-CD, a framework that recasts dense change prediction as syntax-verifiable structured generation. QUAKE-CD represents binary change masks as grammar-constrained quadtree token sequences, making the masks compact, syntactically checkable, and deterministically decodable within an autoregressive generation space. We further construct QUAKE-CoT, which pairs these sequences with chain-of-thought traces grounded in visual evidence, and jointly optimizes textual reasoning and spatial dense prediction through a progressive curriculum followed by grammar-gated dual-reward RL. On QUAKE-CoT, QUAKE-CD achieves 78.31% accumulated F1, outperforming decoder-based and flat text-as-mask VLMs while producing more faithful bi-temporal reasoning.
[CV-50] Pretraining and Distillation Matter More Than Architecture Family for Label-Free Single-Cell Classification
链接: https://arxiv.org/abs/2609.09863
作者: Philip Graemer,Giuseppe Di Caprio
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 25 pages, 3 figures
Abstract:Choosing a deep learning architecture for label-free single-cell classification remains an open question, with microscopy benchmarks reporting conflicting conclusions about CNNs versus transformers. We present a controlled benchmark on LIVECell phase-contrast microscopy data using source-image-disjoint train/validation/test splits to prevent parent-image leakage and matched optimisation, augmentation, and evaluation protocols across EfficientNet, Vision Transformer (ViT), and EVA-02 models. This allows the effects of architecture, pretraining, fine-tuning, tokenisation, and distillation to be disentangled. We find that the previously reported CNN advantage is largely explained by pretraining rather than architecture: the smallest pretrained model outperforms the strongest model trained from scratch despite far fewer parameters. Pretraining improves macro-F1 by 3-4 points, while the gap between the best pretrained CNN and transformer is below 0.5 points. Architectural choices nevertheless matter: ViT-S/8 outperforms ViT-S/16 and matches the four-times-larger ViT-B/16 at a quarter of the parameters, showing that finer tokenisation benefits small cell crops. Conversely, layer-wise learning-rate decay, central to the EVA-02 fine-tuning recipe, degrades performance, highlighting that transfer heuristics from natural-image recognition may not generalise to microscopy. Finally, knowledge distillation substantially improves the deployment frontier: compact EfficientNet-B0 students distilled from teacher councils outperform every individually trained backbone, including the EfficientNet-B5 and EVA-02 teachers. Overall, our results show that rigorous control of pretraining and evaluation is essential for interpreting biomedical architecture benchmarks, while distillation may be a more effective route to practical single-cell classification than architecture choice alone.
[CV-51] SkNeXt enables topology-guided neuronal reconstruction from petabyte-scale microscopy data
链接: https://arxiv.org/abs/2609.09832
作者: Jiayi Ding,Hu Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Recent advances in high-resolution fluorescence and electron microscopy have enabled nanoscale imaging across increasingly large brain volumes, but the resulting terabyte- to petabyte-scale datasets make complete neuronal reconstruction prohibitively expensive in computation, data movement, and manual proofreading. Here, we present SkNeXt, a topology-first framework for scalable neuronal reconstruction from large volumetric microscopy datasets. Instead of densely processing entire image volumes, SkNeXt first converts neuronal morphology into compact SWC skeletons that preserve long-range connectivity. Proofreading is therefore focused on sparse neuronal trees, allowing branch, continuity, and connectivity errors to be corrected before high-resolution reconstruction. The corrected skeletons then serve as persistent structural priors for recovering detailed morphology while preserving neuronal identity and topology. Crucially, SkNeXt also uses neuronal skeletons as spatial indices for selective data access, retrieving high-resolution image regions only along reconstructed trajectories and bypassing most background and signal-free volumes. This substantially reduces I/O and computational overhead, allowing reconstruction cost to scale with neuronal morphology rather than total dataset size. Using SkNeXt, we reconstructed neurons from a petabyte-scale super-resolution fluorescence dataset of the mouse brain on a single GPU within one week, without requiring exhaustive dense inference across the complete imaging volume.
[CV-52] RealSimLoop: Online Real-to-Sim Adaptation via Differentiable Reduced-Order Simulation with Vision Feedback
链接: https://arxiv.org/abs/2609.09828
作者: Zhihao Cen,Chuhua Xian,Hailin Sun,Yuliang Liufu,Zhen Zhang,Xiangyu Chu,Hongmin Cai,Yunbo Zhang,Guoxin Fang
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Real-world observations of deformable objects are often sparse or surface-level, while downstream tasks require hidden physical quantities such as internal deformation, stress fields, and interaction forces. Physics-based simulation can recover these quantities, but online real-to-sim adaptation remains challenging due to costly full-space optimization, limited feedback, and time-varying material properties. To address these challenges, we propose RealSimLoop, a differentiable framework for online real-to-sim adaptation using vision data as physical feedback. Our approach achieves quasi-real-time performance by executing differentiable simulation within a reduced-order neural subspace, drastically accelerating the optimization loop. We couple this efficient dynamics model with differentiable rendering, enabling direct gradient backpropagation that leverages high-fidelity pixel data to refine physical parameters such as material stiffness. Furthermore, by employing a sliding-window objective function, RealSimLoop enables robust online adaptation, allowing the system to track time-varying material properties and effectively bridge the real-to-sim gap arising from model reduction or unmodeled dynamics. Extensive experiments demonstrate that our method outperforms conventional offline methods, and we validate the framework’s versatility in downstream applications, including external force prediction and 3D stress field reconstruction with novel view synthesis.
[CV-53] Layerwise Tunable Lifting Scheme for the Convolutional Neural Network
链接: https://arxiv.org/abs/2609.09827
作者: Abdumannon Yovkochov,An Le,Sungbal Seo,You-Suk Bae,Truong Nguyen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:This work introduces a family of tunable lifting schemes for biorthogonal wavelet filter banks. We propose three lifting strategies: low-pass tuning (LS-LayLatt-LP), high-pass tuning (LS-LayLatt-HP), and a sequential lifting scheme that jointly adapts low- and high-frequency branches (LS-LayLatt-Sequential). All proposed designs are formulated using a lattice-based lifting structure, which guarantees invertibility and stability for arbitrary parameter values within the lifting functions. We evaluated the proposed methods by integrating them into a ResNet-18 backbone for image classification on the Describable Textures Dataset (DTD), as well as for anomaly detection on hazelnut images from the MVTec-AD dataset and private KRC102S dataset. Experimental results demonstrate consistent performance improvements across all evaluated tasks.
[CV-54] Freezing of Gait Prediction Under Spatial Occlusion: An IMU-Supervised Cross-Modal Distillation Approach
链接: https://arxiv.org/abs/2609.09826
作者: Chandan Biswas,Aryan Singh,Anabik Pal
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 9 pages , 2 figures
Abstract:Parkinson’s disease is a progressive neurodegenerative disorder characterised by gradual deterioration of movement control. Automated freezing-of-gait (FOG) detection supports the objective assessment of gait-related motor impairment. Two common approaches are used for FOG prediction: (i) analysing video recordings of the patient’s movements and (ii) analysing data collected using inertial measurement unit (IMU) wearable sensors attached to the patient’s lower limbs. Video-based approaches may suffer detection errors during continuous turning-in-place tasks because the lower limbs undergo substantial geometric self-occlusion, degrading pose-estimation accuracy. IMU-based approaches are generally less affected by visual occlusion; however, they are difficult to deploy outside clinical or laboratory settings, as the sensors must be attached securely and remain in place throughout the assessment. Motivated by this, we propose a cross-modal subspace distillation framework to mitigate the limitations of unimodal FOG detection by combining IMU accuracy with video-based practicality. We extract invariant latent topologies from a pre-trained kinematic oracle to structurally supervise a non-encoded visual architecture during training. To resolve periods of severe spatial occlusion, a dual-stream visual model probabilistically fuses skeletal graph nodes and continuous spatial pixels, dynamically shifting reliance to uninterrupted pixel boundaries as joint tracking confidence drops. Evaluated against a public, multi-modal sequence dataset of Parkinson’s individuals executing continuous 360^\circ turns, empirical results demonstrate that applying sensory boundary topologies strictly mitigates tracking evaluation entropy. Our constrained optimisation confirms that highly precise FOG prediction bounds can be achieved over zero-wearable inference environments.
[CV-55] MethaneFuse: Learning from Multi-Sensor Satellite Observations for Methane Plume Detection
链接: https://arxiv.org/abs/2609.09762
作者: Yuyao Wang,Juliana Y. Leung,Di Niu
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Methane plume detection from satellite imagery is constrained by incomplete observations: public satellites provide complementary spatial, spectral, and atmospheric evidence, but real plume cases rarely contain fully paired multi-sensor measurements because of revisit schedules, cloud coverage, acquisition quality, and the transient nature of emissions. Most learning-based detectors rely on single-sensor inputs, especially Sentinel-2 (S2), leaving many reported plume cases unusable. We construct MethaneUnion, a temporal multi-sensor dataset built from Carbon Mapper plume reports and matched S2, Landsat 8/9 (L8/9), EMIT, and Sentinel-5P (S5P) observations. Built on MethaneUnion, MethaneFuse learns from heterogeneous satellite observations under partial sensor availability without requiring complete four-sensor measurements. MethaneUnion expands usable coverage from 3,211 valid S2-matched plume cases to 8,981 reported plume cases with multi-sensor observations. At the representative 480 m setting, MethaneFuse achieves 84.87 F1 and 93.62 AUROC, improving over the strongest baseline by 5.65 F1 and 8.30 AUROC points while reducing false positives by 8.19 points. Sensor-availability experiments show that MethaneFuse improves detection when S2 is available and transfers plume knowledge to L8/9, EMIT, and S5P when S2 is unavailable. These results demonstrate the value of learning from incomplete heterogeneous sensor observations for practical methane plume detection. Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG) Cite as: arXiv:2609.09762 [cs.CV] (or arXiv:2609.09762v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.09762 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-56] Arti-JEPA: Adapting Video World Model to Real-Time MRI of the Vocal Tract for Speech-Production Analysis
链接: https://arxiv.org/abs/2609.09757
作者: Hong Nguyen,Sean Foley,Christina Hagedorn,Yijing Lu,Sudarsana Reddy Kadiri,Dani Byrd,Shrikanth Narayanan
类目: ound (cs.SD); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Real-time MRI (rtMRI) captures the dynamics of the entire vocal tract during speech, but labeled data are scarce and the modality - single-slice, grayscale, low-resolution - differs substantially from the natural videos that video foundation models are trained on. We introduce Arti-JEPA, a joint embedding predictive architecture to model vocal tract rtMRI by continuing its self-supervised objective on about 62h of unlabelled vocal-tract videos, and evaluate the frozen representation on three tasks: cross-domain phoneme prediction (on typical speakers), fluent-vs-disfluent classification (a corpus containing stuttered speech), and characterizing pre/post-operative transfer (after partial glossectomy). Three key findings emerge. (1) A temporal video prior decisively outperforms per-frame image encoders, and latent prediction (V-JEPA) is at least as strong as pixel reconstruction (VideoMAE), with the edge on fine-grained phonemes. (2) Domain adaptation is \emphtask-dependent: it roughly doubles cross-domain phoneme prediction \kappa (to 0.352) but does not help binary stuttering classification. (3) Arti-JEPA was able to recover phoneme signal from pre/post glossectomy speech — an in-domain probe decodes patients at least as well as a typical speaker, indicating that the residual transfer gap is cross-speaker/domain misalignment, not surgical signal loss, and post-operative decoding does not fall below performance on pre-operative speech. Together, these position a frozen, domain-adapted rtMRI encoder as a reusable measurement tool for articulatory and clinical speech science.
[CV-57] Distilling Image Prototypes for Guided Test-Time Adaptation
链接: https://arxiv.org/abs/2609.09737
作者: Liwen Wang,Xingbo Dong,Iman Yi Liao,Deyin Liu,Massimo Tistarelli,Lin Yuanbo Wu,Zhe Jin
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Test-Time Adaptation (TTA) enhances the robustness of models against distribution shifts but faces two critical challenges: error accumulation from noisy pseudo-labels and catastrophic forgetting of source knowledge. Uncertainty-based approaches designed to mitigate error accumulation often yield overconfident or computationally expensive estimates, while strategies intended to prevent forgetting via prototype replay rely on static representations that easily become misaligned as the model adapts. To address these issues, this paper proposes a novel framework, Distilling Image Prototype for Guided Test-Time Adaptation (DIPTTA). The core of the proposed approach is the introduction of a Distill Image Prototype (DIP), a compact set of synthetic images that serves as a dynamic and regenerative anchor of source knowledge. This prototype enables a dynamic feature replay mechanism that continuously generates feature prototypes aligned with the current state of the model, thus effectively preventing catastrophic forgetting. Furthermore, the DIP anchors a source-calibrated uncertainty estimation method, which provides a less biased measure of sample reliability by leveraging stable source knowledge, thereby robustly suppressing error accumulation. Extensive experiments on multiple benchmarks demonstrate that DIPTTA significantly outperforms state-of-the-art methods, particularly under severe domain shifts. The source code is available at this https URL.
[CV-58] IAE-VTG: Interaction-Aligned Action-Entity Video Temporal Grounding
链接: https://arxiv.org/abs/2609.09736
作者: Shiwen Zhao,Qi Zhang,Sezer Karaoglu,Theo Gevers,Martin R. Oswald
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注:
Abstract:Video Temporal Grounding (VTG) localizes the video segment that matches a natural-language query. Many queries describe an action performed by a particular entity. Existing methods often encode the query as a whole or use general video-text interactions, without explicitly checking whether the action and entity occur together. They may therefore select a segment that contains both concepts but not the event described by the query. We propose Interaction Aligned Action-Entity Video Temporal Grounding (IAE-VTG), which models this rela?tionship at both the representation and training assignment levels. First, the Fine-grained Disentangled Interaction Module (FDIM) separates action and entity related query information and aligns it with complementary motion and appearance features. It then combines token-level interactions to build representations that capture the relationship between the action and entity. Second, Interaction-Sensitive Assignment (ISA) adds this interaction evidence to bipartite matching, so training targets are selected using both temporal overlap and semantic compatibility. This reduces supervision from temporally plausible but semantically incorrect proposals. Experiments on QVHighlights, Charades?STA, and TACoS show that IAE-VTG consistently improves strong baselines and achieves competitive or state-of-the-art performance on standard grounding metrics. Additional analyses show that the method is especially effective when similar actions or entities appear at multiple times and produces more reliable assignments for complex events.
[CV-59] VFNet: Multi-View Spatio-Temporal Model for Void Fraction Estimation in Gas-Liquid Two-Phase Flow
链接: https://arxiv.org/abs/2609.09711
作者: Md Adnan Faisal Hossain,Raghav Rajeev,Kumar Nishant,Justin A Weibel,Satish Kumar,Fengqing Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Void fraction, which quantifies the proportion of the fluid flow volume occupied by the gas phase, is a key parameter in the characterization of gas-liquid two-phase flow. Existing estimation methods either rely on flow assumptions that do not generalize across different fluids or on intrusive sensing that disturbs the flow behavior. We propose VFNet, a dual-branch spatio-temporal neural network for void-fraction prediction from synchronized multi-view videos of two-phase flow. A local branch extracts features from confined spatial regions and fuses the synchronized dual views, while a spatio-temporal branch captures the global evolution of the flow across space and time to refine a coarse geometric estimate. Trained on simulated computational fluid dynamics (CFD) data with known ground-truth void fractions and evaluated against both learning-based and traditional baselines, VFNet achieves the best performance across a broad range of metrics and also improves downstream flow-pattern classification on real two-phase flow data.
[CV-60] Cross-Species Animal Re-Identification with Semantic Consistency Learning ECCV2026
链接: https://arxiv.org/abs/2609.09705
作者: Shuoyi Chen,Yuejia Li,Mang Ye
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV 2026. 18 pages, 5 figures
Abstract:Generalizable animal Re-Identification (ReID) aims to recognize individual animals across species with diverse morphologies and ecological contexts. Unlike person ReID, where different domains share similar body structures, animal species often exhibit drastically different anatomical structures and visual patterns, making it difficult to establish shared visual correspondences. As a result, representations learned across species tend to form fragmented embedding spaces, which severely limits cross-species generalization. To address this challenge, we propose Semantic Consistency Learning (SCL), a framework designed to learn representations that remain stable across appearance variations while preserving semantic structures shared across species. SCL consists of two complementary components. Foreground-Background Decoupled Spectral Normalization (FDSNorm) stabilizes feature statistics by suppressing environment-induced style variations in a region-aware manner, while Cross-species Neighborhood Modeling (CNM) captures transferable relational structures across species through dynamic feature neighborhoods. Extensive experiments on 11 public animal ReID datasets demonstrate that SCL consistently outperforms state-of-the-art methods under multiple cross-species evaluation protocols and generalizes effectively to previously unseen species and ecological domains. Code is available at this https URL.
[CV-61] Recovering Biomechanical Signals from Missing Keypoints Using Temporal Interpolation in Monocular Gait Analysis
链接: https://arxiv.org/abs/2609.09670
作者: Shubham Jariwala
类目: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注: 5 pages, 1 figure
Abstract:Monocular pose estimation enables low-cost gait analysis but is sensitive to missing keypoints caused by occlusion, detection errors, or efficiency-driven model reduction. While prior work on recovering missing joints focuses on complex learned models, the effectiveness of simple temporal methods remains underexplored. We evaluate knee-angle estimation under a missing-ankle-keypoint condition and test a first-order temporal interpolation scheme as a recovery mechanism. Across 527 frames of monocular walking video (428 with valid baseline detections), removing the ankle keypoint increased mean angular error to 23.4° +/- 46.7° and collapsed signal variance to near zero. Temporal interpolation reduced error to 1.1° +/- 6.7° and restored variance and smoothness to within a few percent of baseline. These results indicate that gait signals possess sufficient temporal redundancy for a simple, computationally trivial interpolation scheme to recover a critical missing joint, without resorting to learned reconstruction models. The findings support low-complexity, real-time-compatible designs for gait analysis in resource-constrained or occlusion-prone monocular settings.
[CV-62] LightMedSeg-ISLES: Stroke Lesion Segmentation with 81x Fewer Parameters than nnU-Net
链接: https://arxiv.org/abs/2609.09634
作者: Giorgi Nikvashvili,Hanxue Gu,Jie Bao,Kang Wang,Yang Yang
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 8 pages, 3 figures. Submitted to ISLES 2026 challenge. To be published in Nature Lecture Notes in Computer Science (LNCS)
Abstract:Large networks and ensembles often lead medical image segmentation challenges, but their storage and inference demands complicate deployment. We present LightMedSeg-ISLES, a 1.26-million-parameter pipeline for T1-weighted stroke lesion segmentation in ISLES’26. On a 146-case held-out cohort, flip test-time augmentation produces 0.618 mean Dice and 0.599 lesion-wise F1. A 102.35-million-parameter nnU-Net ResEnc-L produces 0.634 Dice and 0.544 lesion-wise F1 after size filtering. LightMedSeg therefore retains 97.5% of nnU-Net’s Dice with 81.4 \times fewer parameters while improving lesion-wise F1 by 0.055. Its four-pass TTA operating point requires 4.7 \times fewer FLOPs per standardized patch than nnU-Net. It also slightly exceeds filtered UNETR++ and nnFormer. Longer training and stronger augmentation add 0.0358 Dice without increasing capacity, establishing a strong single-checkpoint alternative to much larger models.
[CV-63] Hyperbolic Geometry for Open-World Object Detection in Remote Sensing Imagery
链接: https://arxiv.org/abs/2609.09626
作者: Wuzhou Li,Jiawei Zhou,Shenghang Wang,Xiang Li
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Open-world object detection (OWOD) extends closed-set detection by requiring models to identify unknown objects and incrementally learn them once annotations become available. In remote sensing imagery, object categories often exhibit latent hierarchical relationships that may be inadequately represented in the Euclidean spaces commonly adopted by existing methods, limiting unknown-object recall and incremental-learning performance. To address this issue, we investigate hyperbolic geometry for OWOD in remote sensing imagery and propose HyRS-OWOD. To improve unknown object recall, we design a two-step unknown-object discovery mechanism: a Decoupled Objectness Learning (DOL) module that disentangles foreground perception from semantic information to separate foreground proposals from background regions, followed by a Hyperbolic Uncertainty Learning (HUL) component that leverages the radius of hyperbolic embeddings as an uncertainty-aware cue for known-unknown discrimination. For incremental learning, we develop a Hyperbolic Metric Learning (HML) strategy that enhances inter-class separability, facilitating the incorporation of novel categories while mitigating catastrophic forgetting. Experiments on three remote sensing benchmarks demonstrate consistent improvements in unknown recall and incremental learning over state-of-the-art OWOD methods.
[CV-64] Marker-free eye-gaze estimation using a single image and depth from defocus
链接: https://arxiv.org/abs/2609.09610
作者: David Hurtubise-Martin,Feriel Fass,Djemel Ziou,Marie-Flavie Auclair-Fortier
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:This paper presents a marker-free eye-gaze estimation approach using a single 2D camera, such as an integrated laptop webcam. The gaze-related features are estimated from iris localization and head pose estimated by using depth from defocus. A variational Bayesian multinomial logistic regression framework is used as mapping from the estimated features to the position of regard, based on an 8-dimensional feature vector of head-pose and iris-displacement parameters. No external marker is needed. Experiments were conducted by estimating the gaze of people watching a computer screen at different distances and compared against five existing methods. The obtained scores demonstrate the effectiveness of the proposed approach.
[CV-65] RouteBridge: Reliability-Routed Bidirectional Distillation Between Neural Radiance Fields and 3D Gaussian Splatting
链接: https://arxiv.org/abs/2609.09606
作者: YuanHang Wang,Xin Cao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Neural radiance fields (NeRFs) and 3D Gaussian Splatting (3DGS) encode a scene with complementary inductive biases, but existing cross-representation distillation typically fixes one representation as teacher for the entire scene. A globally fixed teacher can propagate local reconstruction errors. We present RouteBridge, a bidirectional framework that selects the teaching direction for each ray. Its reliability estimator combines photometric residuals with representation-specific geometric evidence and routes supervision from NeRF to 3DGS, from 3DGS to NeRF, or abstains. A renderer-independent interface transfers color, opacity, and normalized depth without shared features or point correspondence. On mip-NeRF 360, the NeRF and 3DGS exports reach 28.56 and 28.77 dB, respectively. The 3DGS export improves over 3DGS by 1.56 dB and over NeRF-GS by 0.45 dB while reducing LPIPS to 0.207. On static three-view DTU, RouteBridge obtains 21.12 dB. Ablations show that both adaptive routing and geometric ray targets contribute to the improvement.
[CV-66] MotionBlind: Probing the Illusion of Motion Understanding in Video-LLM s
链接: https://arxiv.org/abs/2609.09528
作者: Dhairya Bhatia,Bishoy Galoaa,Oliver Fritsche,Shahid Kamal,Muhammad Obaidullah Abdul Salam,Umer Saleem,Om Rastogi,Frania Felix Chettiar,Nesli Erdogmus,Sarah Ostadabbas
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Video large language models (Video-LLMs) are increasingly used as the perceptual front end of world models, a role that assumes they can read motion: how fast something moves, which way it travels, how hard it is pushed. We show they cannot. A Video-LLM can watch two clips of the same person in the same room, name every object in both, and still fail to say which clip moves faster. We introduce MotionBlind, a contrastive benchmark of self-recorded video for physically grounded motion(speed, magnitude, and direction), the variables a world model must predict. Each instance is a pair of near-identical clips that differ only in motion. Each clip carries two complementary yes/no questions, giving four items per instance, and a model earns credit only if all four are correct. We report Instance Accuracy(IAcc), which has a 6.25% chance floor. Single-frame, appearance, and language-only shortcuts all collapse to it. MotionBlind complements the recent TimeBlind benchmark. We run a controlled study of six open and two frontier Video-LLMs, varying whether the video is present, whether frames are shown in the correct temporal order, and how frames are sampled (1 to 24 frames, four selection strategies). Open models sit near the 6.25% floor, and scale does not help. Removing the video drops every model to zero IAcc, and shuffling frames collapses IAcc to chance, so the task genuinely needs video in order. Neither more frames nor smarter frame selection closes the gap, because these change which frames are seen, not whether motion is read. Only Gemini3.1 Pro clears the benchmark overall, and even it fails on speed. A frontend that cannot tell two speeds of the same action apart is not yet a trustworthy source of supervision, reward, or evaluation for a world model.
[CV-67] AnimalLift: Reconstructing Animatable 3D Animals from a Single Image by Learning Canonical Shape Texture and Fur Maps
链接: https://arxiv.org/abs/2609.09513
作者: Chunyi Sun,Ruyi Zha,Weijian Deng,Junlin Han,Dylan Campbell,Stephen Gould
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Reconstructing a fully animatable 3D animal from a single image remains challenging because animation-ready assets require not only plausible geometry, but also a unified topology, editable appearance, and fur representations compatible with deformation and simulation. Existing image-to-3D approaches often rely on implicit or loosely structured representations that are difficult to rig or edit, while parametric animal models support animation but cannot capture detailed texture and fur appearance. We present AnimalLift, a framework for reconstructing structured, animation-compatible 3D animal assets with explicit fur from a single image. Our method lifts an input image into a shared canonical space with a consistent topology and UV parameterization across the dataset, enabling joint prediction of canonical geometry, texture, and fur in a unified feed-forward architecture. A key component of our representation is a UV-aligned fur map that encodes strand geometry in a surface-aligned canonical domain, allowing explicit fur reconstruction compatible with mesh deformation and fur simulation. To train the model, we introduce a procedural data generation pipeline that provides large-scale supervision with aligned geometry, texture, and fur across diverse animal species and appearances. Experiments on synthetic and real-world datasets demonstrate strong reconstruction quality and generalization across animal categories. Beyond reconstruction, our structured representation directly supports downstream applications including animation, pose transfer, fur editing, and simulation-compatible rendering.
[CV-68] RoMa-Ω: What Feed-Forward 3D Models Know About Image Matching
链接: https://arxiv.org/abs/2609.09507
作者: David Nordström,Xinyue Zhang,Thibaut Loiseau,Vincent Lepetit,Fredrik Kahl
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Learned image matching has experienced significant progress in recent years, culminating in robust and accurate matchers such as RoMa, whose robustness is often attributed to its use of frozen DINO features. In a parallel development, feed-forward reconstruction models, such as VGGT, have been trained on ever-growing datasets to accurately regress dense 3D point maps and camera poses. The distinction between matchers and feed-forward reconstruction models has become increasingly blurred with the introduction of matching losses in models such as MASt3R and VGGT- \Omega . This raises a natural question: what do feed-forward 3D models know about image matching? In this work, we answer this question by analyzing three scenarios: (i) zero-shot matching of patch features, (ii) direct matching of 3D point predictions, and (iii) training a full matcher on top of the learned representations. We find that, despite performing poorly in zero-shot matching, especially in later layers, feed-forward reconstruction models provide strong representations for linear probing and full matching pipelines. We further show that, even without any training, their raw predictions alone enable competitive matching, albeit only under moderate viewpoint changes and modality gaps. Based on these insights, we retrain RoMa v2 by replacing its DINO backbone with VGGT- \Omega . Our resulting model, \ours, outperforms state-of-the-art matchers on a wide range of benchmarks, e.g. +8.1 mAA compared to RoMa v2 on WxBS.
[CV-69] Learning Global Camera Poses from Noisy View-Graphs for Structure from Motion ECCV2026
链接: https://arxiv.org/abs/2609.09491
作者: Fadi Khatib,Meirav Galun,Ronen Basri
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV 2026. Project page: this https URL
Abstract:Camera pose estimation is a key step in 3D reconstruction and view-synthesis pipelines. We present a deep, global Structure-from-Motion framework based on learned view-graph aggregation. Our method employs a permutation-equivariant, edge-conditioned graph neural network that takes noisy pairwise relative poses as input and outputs globally consistent camera extrinsics. The network is trained without ground-truth supervision, relying solely on a relative-pose consistency objective. This is followed by 3D point triangulation and robust bundle adjustment. Our approach is efficient, scalable to more than a thousand images, and robust to graph density. We evaluate our method on MegaDepth, 1DSfM, Strecha, and BlendedMVS. These experiments demonstrate that our method achieves superior rotation and translation accuracy compared to deep track-centric methods while registering more images across many scenes, and competitive results compared to state-of-the-art classical pipelines, while being much faster.
[CV-70] Efficient Fairness Auditing Across Guidance Scales in Text-to-Image Diffusion Models via Causal Abstraction
链接: https://arxiv.org/abs/2609.09486
作者: Nabila Tasfiha Rahman,Rajatsubhra Chakraborty,Depeng Xu,Lu Zhang
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Fairness auditing of text-to-image diffusion models often requires generating large numbers of images across sampling configurations, making comprehensive evaluation computationally expensive. We propose a causal-abstraction-based audit instrument for efficiently evaluating fairness under interventions on the classifier-free guidance scale. Given a fixed prompt and a target feature function, we represent the diffusion process as a low-level structural causal model and construct a corresponding high-level model over abstract denoising states. We characterize the projected causal structure, establish identifiability of the fairness-relevant interventional query, and provide sufficient conditions under which the high-level model preserves this query. A probabilistic transformer implements the high-level model as an amortized predictor of target-feature distributions across guidance scales. Experiments evaluate distributional fidelity, fairness-query accuracy, and computational efficiency. We present two auditing demonstrations: one using standard Stable Diffusion 1.5 and another using StayFair, a fairness-enhanced Stable Diffusion model, to examine their behavior across guidance scales.
[CV-71] Infra-Bench CLS: A Global Open-Source Benchmark for Critical Infrastructure Classification with Earth Observation Foundation Models
链接: https://arxiv.org/abs/2609.09482
作者: Justin Guthrie,Edward Oughton,Konrad Wessels,Matthew Rice,Isaac Corley
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 9 figures. Supporting information with 10 figures and 13 tables. Submitted to Big Earth Data
Abstract:Critical infrastructure location data is often incomplete and unevenly distributed globally, especially in developing regions. Earth observation foundation models are proposed as a new step in enabling us to more efficiently understand the natural and built environment, raising questions as to their effectiveness in performing challenging downstream tasks. Yet, foundation models remain largely untested for detecting and classifying the facility-scale critical infrastructure that underpins a range of important societal and economic functions. Subsequently, Infra-Bench CLS is introduced as a benchmark to test foundation models on 18,756 Sentinel-1 SAR and Sentinel-2 multispectral facility-scale critical infrastructure asset images covering seven continents and 13 infrastructure classes, with results reported for the 10 retained classes. Using linear probing and fine-tuning for two training dataset levels (1.0x and 0.3x), seven foundation models are evaluated (SatlasPretrain S2, SatlasPretrain S1, CROMA, Prithvi-EO-2.0, AlphaEarth Foundations, OlmoEarth v1.1-Base, and DINOv3 ViT-L/16). When comparing macro F1 scores to a ResNet-18 supervised baseline of 39.2 percent, the best foundation model achieved 57.9 percent, a 48 percent improvement. Top performing classes were airports (F1 85.3 percent), train stations (F1 82.1 percent), and data centers (F1 77.6 percent). By contrast, many of the power sector classes perform poorly (F1 27.5-46.2 percent). These findings suggest foundation models can enable superior critical infrastructure classification, but future work should evaluate performance on higher-resolution imagery, particularly for poorly performing sectors, such as power.
[CV-72] LeCor: Learning to Be Corrected by Meta-Learned Test-Time Training for Interactive 3D Lung-Tumour Segmentation
链接: https://arxiv.org/abs/2609.09477
作者: Yi Luo,Yike Guo,Wenxuan Li,Zongwei Zhou,Rui Zhang,Kai Ding
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 18 pages, 4 figures
Abstract:Delineating lung tumours on computed tomography (CT) takes a considerable share of the time spent on radiotherapy planning, and a contour proposed by a model can be refined interactively by the clinician. Promptable foundation models such as SAM 3 support this workflow by writing each correction into a session memory that conditions the remaining slices, while the model weights stay fixed. On 690 test cases from five public CT cohorts, fine-tuning SAM 3 on lung tumours raises the Dice obtained from a single point prompt from 0.298 to 0.757, and seven rounds of corrections raise it further to 0.765, but under memory conditioning alone the accuracy on slices the annotator has not touched stops improving after six rounds. We therefore treat each correction as a training signal and propose LeCor, which performs test-time training on a small set of case adapters that are reset for every case and meta-learned such that a single gradient step driven by a click improves the slices that were not clicked. On the 133 test cases that span at least eight slices, LeCor raises the Dice reached after seven correction rounds from 0.787 with the fine-tuned model to 0.827, reduces the number of cases that never reach a Dice of 0.80 from 47 to 27, and reaches in three correction rounds the accuracy that the fine-tuned model attains in seven.
[CV-73] Low-Rank Prompt Learning for Vision-Language Models with Fixed-Token Bases
链接: https://arxiv.org/abs/2609.09462
作者: Tanvir Muntakim Tonoy,Sajjad Ghiasvand,Mahnoosh Alizadeh,Ramtin Pedarsani
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Prompt learning adapts CLIP to downstream recognition by replacing hand-written templates with learned continuous context vectors, which in Context Optimization (CoOp) form a dense prompt matrix \mathbfP\in\mathbbR^m\times d trained from only a few examples per class. We study whether this matrix is over-parameterized by factorizing it as \mathbfP=\mathbfB\mathbfA , which cuts the trainable prompt parameters from md to r(m+d) , and to rd once the token-side factor \mathbfB is fixed. Across seven few-shot benchmarks and two CLIP backbones, low-rank prompts match or improve dense CoOp at far fewer parameters, with the clearest gains on low-shot base-to-new generalization. We then find that the token-side factor need not be learned at all: fixing \mathbfB to a Gaussian, orthogonal, SVD-derived, or even random basis and training only the embedding-side factor \mathbfA stays on par with the fully trainable factorization, and a source-trained \mathbfB offers no advantage over a random one. A prompt-factor asymmetry and a local update-space dimension gap show why fixing \mathbfB is far less restrictive than fixing \mathbfA , and a smoothness-only guarantee certifies that optimizing \mathbfA over a fixed \mathbfB converges. In the CLIP prompt setting, the embedding-side coefficients carry the adaptation while the token basis can simply be fixed.
[CV-74] Longitudinal tracking of multiple sclerosis lesions in the spinal cord: A validation study
链接: https://arxiv.org/abs/2609.09424
作者: Pierre-Louis Benveniste,Julian McGinnis,Shannon Kolind,Larry D. Lynd,Sarah A. Morrow,Jiwon Oh,Alexandre Prat,Alice Schabas,Penelope Smyth,Roger Tam,Anthony Traboulsee,Mark Mühlau,Herve Lombaert,Julien Cohen-Adad
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 5 figures
Abstract:Longitudinal characterization of multiple sclerosis (MS) lesions remains constrained by the lack of frameworks capable of establishing consistent instance-level correspondences across time. Conventional segmentation approaches produce semantic lesion masks at each visit and therefore fail to capture the complex instance temporal patterns associated with lesion appearance, disappearance, splitting, or merging. This study presents a comparative evaluation of five strategies for automated tracking of spinal cord MS lesions in longitudinal MRI data from a multi-site cohort. The investigated strategies rely either on deformable registration or on a spinal anatomical reference system, and encompass overlap-based matching, coordinate-based Hungarian algorithm, gradient-boosted classification, and Siamese model classification. Tracking accuracy is quantified using instance-level true positives, false positives, and false negatives, allowing to assess the presence of one-to-many and many-to-one associations. Results show best performance for the registration-based overlap method. This study provides the first systematic analysis of lesion-instance correspondence in the spinal cord and outlines the strengths and limitations of registration-based and registration-free paradigms for longitudinal MS assessment. The code is available at this http URL .
[CV-75] Vision-language models know more about agriculture than they show and rubric-grounded verifications close the gap NEURIPS
链接: https://arxiv.org/abs/2609.09417
作者: Earl Ranario,Jared Smith,Lars Lundqvist,Urmil Jatin Chandarana
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to the AI for Science Workshop (NeurIPS Workshops 2026)
Abstract:Vision-language models (VLMs) show promise for agricultural classification, but zero-shot performance on disease, pest, damage, quality, and species identification remains poor, and it is unclear whether this reflects weak visual features or a failure to connect them to domain knowledge. We build a benchmark of 116 datasets, 834 classes, and 8,324 images spanning these tasks to isolate where the gap arises. Linear probing shows VLM vision encoders already encode agricultural features nearly as separable as a self-supervised DINOv3 baseline, ruling out weak visual representations as the primary bottleneck. Conditioning each model on an oracle reference description (an upper bound on its parametric knowledge) closes most of the gap left by an unaided lower bound, showing VLMs already know more about agriculture than they show. To close this gap without an oracle description at inference time, we structure test-time reasoning around a fixed, per-task diagnostic rubric: the model generates K candidate responses and a Probabilistic Pivot Tournament (PPT) verifier, scored pairwise against the rubric, selects the best one. This nearly doubles judged F1 over the lower bound and matches or exceeds the upper bound on several tasks, notably pushing Gemma 4 E4B-it’s disease F1 to 0.71, above its own upper bound of 0.60. However, the verifier’s letter-scale confidence score has the opposite of its intended effect: filtering to its most confident predictions does not improve accuracy and correlates negatively with correctness across every model and pool size tested, so the score cannot serve as a measure of predictive uncertainty, and most of the observed gain likely comes from rubric-grounded generation rather than pairwise verification.
[CV-76] VANTAGE-Bench: Evaluating the Infrastructure AI Gap in Vision-Language Models
链接: https://arxiv.org/abs/2609.09396
作者: Zaid Pervaiz Bhat,Nimra Nayyar,Arihant Jain,Lap Fung Chan,John Suchanek,Yu Wang,Varun Praveen,Tomasz Kornuta,Vidya Nariyambut Murali
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 23 pages, 2 figures, 14 tables. Project page: this https URL dataset: this https URL
Abstract:As Vision-Language Models (VLMs) advance toward physical deployment, the focus has remained on action-oriented Embodied AI evaluated on subject-centric consumer video. This overlooks a pervasive class of Physical AI: Infrastructure AI, which relies on fixed cameras for open-loop insights like safety monitoring and operational logging. We introduce VANTAGE-Bench, a benchmark measuring this “Infrastructure AI Gap.” It spans three operational domains (Logistics, Transportation, and Smart Spaces), unifies image and video evaluation across semantic, spatial, temporal, and spatio-temporal capabilities, and moves beyond multiple-choice to eight task formulations including dense captioning and spatio-temporal grounding. It adds a single-pass trajectory protocol for Single Object Tracking and, to our knowledge, the first such evaluation on fixed-camera infrastructure video, scored against specialist trackers. Annotation spans three regimes over 3,346 media assets: 3,342 video-task annotations, 4,281 image-grounding annotations, and 27,404 detection boxes. Evaluating 17 models zero-shot, we find the shortfall relative to consumer-centric benchmarks is concentrated, not general. Event verification, referring expressions, and temporal localization fall roughly 9 to 24 points at every model scale, while video question answering stays within 5.3 points of VideoMME and 2D spatial pointing shows no shortfall against BLINK. The temporal pillar is weakest in absolute terms: no system exceeds 55.7 mIoU on temporal localization or 37.3 SODA_c on dense video captioning. On tracking, frontier models come within roughly 5 points of specialist trackers over short horizons but separate as the horizon extends. Open-weight models lead 2D object localization outright, so neither scale nor proprietary access explains the pattern. Data, evaluation harness, and leaderboard: this https URL Comments: 23 pages, 2 figures, 14 tables. Project page: this https URL dataset: this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.09396 [cs.CV] (or arXiv:2609.09396v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.09396 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-77] OmniPoint: Universal Monocular Metric Pointcloud from Any Camera ECCV20026
链接: https://arxiv.org/abs/2609.09394
作者: Botao Ye,Marc Pollefeys,Ming-Hsuan Yang,Abhijit Kundu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ECCV 20026. Project Page: this https URL
Abstract:Recovering metric 3D geometry from monocular images is a fundamental computer vision task, yet current methods remain heavily fragmented by fixed camera model assumptions and inflexible input schemes. We present OmniPoint, a unified framework designed to generalize metric reconstruction across diverse imaging sensors, including pinhole, fisheye, and equirectangular projections, while accommodating varying geometric priors. To overcome projection rigidity, OmniPoint abandons conventional planar depth regression. It instead adopts a decoupled ray and distance representation alongside a decoupled training objective, explicitly separating the camera projection model from the scene structure. To address the severe scarcity of training data for alternative cameras, we introduce a bidirectional augmentation strategy that explicitly bridges labeled perspective data and unlabeled omnidirectional domains in 3D space. Furthermore, to seamlessly integrate optional inputs like camera intrinsics or sparse depth without destabilizing the network through feature distribution shifts, we propose a robust information injection mechanism. This mechanism utilizes learnable input state embeddings to resolve architectural ambiguity and applies vectorized Gaussian smoothing to densify irregular measurements. Extensive experiments demonstrate that OmniPoint achieves state-of-the-art zero-shot performance across multiple benchmarks, establishing a robust new standard for unified monocular 3D reconstruction.
[CV-78] he Living Library: Transforming Archival Collections into Conversational Knowledge Systems – Lessons from the Theodore Roosevelt Presidential Library
链接: https://arxiv.org/abs/2609.09368
作者: Pengce Wang,Lucia Ronchi Darre,Matt Briney,Michaell Bakalars,Dan Rutkowski,Ursula Hardy,David Wolf,Laura Hoffman,Allen Kim,Shawn Wright,Juan Lavista Ferres
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 25 pages, 6 figures, 4 tables
Abstract:We present the Living Library, an end-to-end framework for transforming fragmented digital archives into governed, conversational, in-person exhibit experiences. Developed and deployed at the Theodore Roosevelt Presidential Library, the framework comprises four layers: digitization and corpus creation, AI-powered processing, retrieval and reasoning, and an optional embodied conversational interface. The first three layers aggregate a 300,000-record collection, apply OCR and structured metadata enrichment for expert curatorial review, and publish records to a hybrid dense/semantic index. Expert review is conducted through the Archivist App, a curator-facing interface that supports correction of AI-generated transcriptions and metadata. The governed corpus powers both a researcher-facing interface and Talk to TR, a continuously operating exhibit that embodies Theodore Roosevelt as a full-scale digital human within a museum environment. To support live, face-to-face interactions, Cross-Era Analogical Grounding reframes contemporary questions through documented historical parallels, allowing Roosevelt to address present-day topics without inventing facts. Dual-path retrieval and end-to-end streaming keep responses grounded and responsive. Layered watchdogs, visitor-session isolation, automated conversation management, and independently restartable services enable reliable unattended operation for hundreds of visitors. Avatar realism, spatial audio, lighting, staging, and conversational design are developed and evaluated as an integrated experience. Rather than report a controlled benchmark, we describe lessons from operating Talk to TR as a public exhibit and offer a transferable model for transforming archival collections into believable, in-person conversational experiences. Comments: 25 pages, 6 figures, 4 tables Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.09368 [cs.CV] (or arXiv:2609.09368v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.09368 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-79] DensePol: Dense-Angle Polarization Dataset for Learning-Based Polarimetric Vision
链接: https://arxiv.org/abs/2609.09359
作者: Param Sangani,Ahmad Moori,Erik Blasch,Guna Seetharaman,Hadi Aliakbarpour
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Polarimetric vision is gaining increasing attention because it provides physical cues about scene shape, material, and reflection that are difficult to recover from RGB alone. Recent work has therefore explored predicting polarization directly from conventional RGB images; however, the fidelity of these methods strongly depends on the polarization supervision used for training. Most existing datasets rely on Division-of-Focal-Plane (DoFP) cameras with four spatially interleaved analyzer orientations, which provide limited angular redundancy and introduce interpolation and instantaneous-field-of-view errors. We introduce DensePol, a high-redundancy RGB–polarization dataset based on Division-of-Time (DoT) acquisition, capturing 180 full-resolution analyzer orientations at 1^\circ intervals. DensePol contains 2,018 paired RGB–polarization images with the angular measurements and fitting residuals retained. Dense angular sampling substantially improves polarization stability, reducing AoLP deviation from 13.36^\circ to 2.21^\circ . We further introduce a deterministic diffusion-based RGB-to-polarization framework with cyclic AoLP representation and a local DoLP refiner. Experiments demonstrate improved polarization prediction and downstream surface-normal estimation. The dataset and code will be publicly available.
[CV-80] Video-MOPD: Multi-Teacher On-Policy Distillation for Video Understanding
链接: https://arxiv.org/abs/2609.09300
作者: Zhenxin Qin,Peng Shi,Cong Han,Yinlong Qian,Zequn Jie,Lin Ma
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Technical report
Abstract:Video understanding demands a convergence of complementary capabilities across perception, temporal understanding, and complex reasoning, which are difficult to jointly optimize within a single model. We introduce Video-MOPD-8B, an open-weight model dedicated to video understanding tasks. To fundamentally enhance its capabilities, we conduct targeted reinforcement learning (RL) optimization across three core domains: video temporal grounding (VTG), general video comprehension, and video STEM reasoning. We then unify their complementary capabilities via Multi-Teacher On-Policy Distillation (MOPD), which consolidates expert knowledge by supervising student-generated trajectories with routed teacher feedback. We further introduce Reliability-Aware Informative Sampling (RAIS), which selects examples with consistently reliable teacher supervision and large teacher-student performance gaps. Together, these components enable Video-MOPD-8B to achieve coordinated and comprehensive performance gains across diverse video understanding tasks. Extensive experiments on comprehensive benchmarks covering general video understanding, temporal grounding, video reasoning, and video STEM tasks demonstrate that Video-MOPD-8B achieves state-of-the-art performance among existing models at a comparable scale. The trained model weights are available at this https URL.
[CV-81] No Free Checker: A Survey of Verifiers for Robot Policies
链接: https://arxiv.org/abs/2609.09250
作者: Yang Wan,Xihang Yue,Zhirui Liu,Ziyuan Chu,Shuxun Wang,Yuhan Chen,Xiaonan Jiang,Xukun Zhu,Yubo Dong,Linchao Zhu
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Systems and Control (eess.SY)
备注: Survey. 31 pages, 5 figures, 7 tables, 187 references. Covers reward models, success and failure detection, temporal-logic and formal verification, world-model evaluation, and reward hacking. Project page: this https URL
Abstract:A verifier for robot policies reads a candidate behavior and returns a score for how well it did, used both to evaluate vision-language-action policies and to train them. Verifiers range from success detectors and reward models to runtime monitors, safety filters, and temporal-logic specifications. We survey roughly 150 verifiers and compare them along two properties. Availability is how much a verdict costs, how early in a rollout the verdict arrives, and how often a verdict can be asked for. Availability rises as verdicts get cheaper, earlier, and denser. Credibility is how much a high score tells us about the task. Credibility falls as the judgment becomes gameable and self-serving. We group the verifiers by who supplies the judgment: human verifiers, rule-based and formal verifiers, learned and pretrained verifiers, and model-intrinsic verifiers. Across the four families, we find that credibility falls as availability rises. Regardless of who supplies the judgment, there is no free checker. We then examine what validates a verifier itself, and how much a high score tells us. Three measures appear in the literature: agreement with human labels, the performance of the policy it trains, and behavior under reward hacking. We close with nine metrics that make a verifier claim checkable, and coordinates for the verifiers still to be built.
[CV-82] Identifying Habit Physics and Nuisance in Robot World Models
链接: https://arxiv.org/abs/2609.09210
作者: Jinting Hang,Zhenhui Cai
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Teleoperated demonstrations are often multimodal even when the underlying dynamics are nearly deterministic given the executed action. We argue that this multimodality typically mixes three factors–operator habit in action selection, shared physics, and observation nuisance–and that entangled next-observation predictors absorb all three. We formalize the split with a structural causal model a=g(h,z,u), z’=f(z,a), o=r(z,c), and test it with complementary interventions: replacing or shuffling actions at fixed state sharply increases next-state error, whereas appearance and camera changes should not; habit-aware reverse scoring improves ranking of feasible pasts without rewriting the dynamics. The associated adaptation rule is to freeze a shared physics readout and update only a thin interface. On StackCube, DROID, and RH20T this rule improves low-shot transfer relative to training from scratch, retains cleaner dynamics under corrupted adaptation data, and extends from proprioception to pixel observations with multi-view and multi-step checks. We do not equate latent actions with operator habit, and we do not target large-scale video generation benchmarks.
[CV-83] Lensless Gaze Is Not Private by Default: Auditing Identity Leakage Across Disclosure Surfaces ATC ECCV2026
链接: https://arxiv.org/abs/2609.09188
作者: Rahul Vimalkanth,Kaushik Mitra
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 5 figures. Accepted at the PFATCV Workshop, ECCV 2026. Code available at this https URL
Abstract:Lensless near-eye sensing is often described as privacy-friendly because its coded measurements are visually unintelligible. Yet visual unintelligibility reflects human interpretation, not what a learned adversary can recover. We therefore treat identity privacy as a systems property of disclosure surfaces: representations crossing sensing, storage, computation, and output boundaries. We audit a simulated lensless gaze pipeline under a 36-subject known-gallery closed-set identification protocol with a fixed, known PSF; privacy from an unknown or varying optical key is outside our scope. Reported accuracies are empirical attack success rates under matched linear and MLP probes and do not upper-bound stronger adversaries. Simulated lensless measurements yield 96.7% top-1 identification versus 97.7% for matched original eye crops, while an MAE embedding retains 94.3%. Compression alone offers little protection: 8-D PCA and a matched 8-D bottleneck retain 93.2% and 91.8%, whereas separately trained 8-D GSPL bottlenecks yield 77.5% mean recovery across three seeds. A released 128-way gaze token lowers single-frame recovery to 38.1%, while its residual and continuous gaze output expose 62.1% and 72.6%, respectively. Under a source-frame-disjoint tiled protocol, token summaries reach 39.9% at T=25, showing that repeated-output risk depends on representation and aggregation. These rates reflect all subject-correlated information in the evaluated dataset, including acquisition and behavioral cues, rather than isolating intrinsic ocular biometrics. Ordinary least squares residualization against a six-dimensional crop geometry and intensity summary still leaves lensless recovery at 95.1%. Our results show that privacy claims for lensless sensing must be tested at disclosure boundaries rather than inferred from appearance.
[CV-84] M2LG-DG: A Multi-modal Local-Global Domain Generalization Framework for Cross-site Major Depressive Disorder Classification
链接: https://arxiv.org/abs/2609.09186
作者: Muhammad Asif Hasan,Yanming Zhu,Xuefei Yin,Alan Wee-Chung Liew
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Classification models based on resting-state functional magnetic resonance imaging (rs-fMRI) often show lower performance at imaging sites not included during model development, which can limit their use in clinical settings. Domain generalization (DG) addresses this issue by learning representations from source sites that remain effective for unseen target sites. However, existing DG approaches for psychiatric disorder classification commonly rely on a single imaging modality and may not fully account for site-specific acquisition effects on the learned representation space. Subjects scanned at the same site share scanner hardware, acquisition settings, and preprocessing characteristics, which can cause representations to reflect acquisition conditions rather than diagnostic information. In this work, we present M2LG-DG, a source-only multimodal local-global framework for cross-site major depressive disorder (MDD) classification. The framework employs a dual-stream rs-fMRI encoder, where the global pathway models inter-regional dependencies through self-attention and the local pathway performs graph-constrained aggregation over functional connectivity-derived brain graphs. Imaging and non-imaging representations are decomposed into shared and private components and integrated through bidirectional cross-attention with a learned modality gate. A cross-site supervised contrastive objective forms positive pairs from same-class subjects acquired at different source sites, encouraging the fused representation to preserve diagnostic information across acquisition domains. On four held-out REST-meta-MDD sites, M2LG-DG achieves an AUC of 69.48% and exceeds the closest comparison method by 2.18 percentage points. Experiments on the Autism Brain Imaging Data Exchange (ABIDE) dataset further support its applicability to other psychiatric neuroimaging classification tasks.
[CV-85] Integrating Unimodal and Vision-Language Representations in Latent Space for Multi-Label Chest X-Ray Classification
链接: https://arxiv.org/abs/2609.09185
作者: Quang-Huy Tran,Duc-Tuan Ngo,Minh-Khoi Nguyen-Bui,Dang-Khoa Bui,Thanh-Trong Tran,Tuan-Khoi Nguyen,Hoang-Anh Ngo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 2 figures, 5 tables (main text); 12 pages, 1 figure, 13 tables (supplementary material)
Abstract:Multi-label chest X-ray classification has attracted considerable attention in recent years, with the effective use of visual representations and clinical semantic knowledge playing an important role. This study proposes a framework that combines unimodal representations from RAD-DINO with vision–language representations from BioViL-T for the classification of 14 labels in the MIMIC-CXR-JPG dataset. The RAD-DINO and BioViL-T embeddings and their combined representation are refined separately in latent space before being normalized and fused across the three branches. In addition to improving classification performance, the study aims to clarify the role of each embedding source and the degree to which they complement one another. Experiments show that RAD-DINO outperforms BioViL-T when used independently, whereas early fusion further improves the results, indicating that the two embedding sources contain complementary information. The best-performing model achieves a mean AUROC of 0.840 and an mAP of 0.467. Ablation analysis shows that hybrid fusion provides consistent and statistically significant improvements over early fusion when each embedding source is refined in latent space, suggesting that fusion effectiveness depends on the quality of the representation supplied by each branch. However, the study has only been evaluated internally on MIMIC-CXR-JPG; its generalizability to data from other healthcare institutions therefore remains to be validated. The source code is available at: this https URL. Comments: 10 pages, 2 figures, 5 tables (main text); 12 pages, 1 figure, 13 tables (supplementary material) Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.09185 [cs.CV] (or arXiv:2609.09185v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.09185 Focus to learn more arXiv-issued DOI via DataCite
[CV-86] Evidence-Order Calibration for Selective Visual Reasoning under Progressive Loss of Question-Critical Evidence
链接: https://arxiv.org/abs/2609.09184
作者: Muhamathu Ameer Ali Aacaas Muhamath
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 8 pages, 6 figures, 4 tables. Code and supplementary materials: this https URL
Abstract:Vision-language model (VLM) confidence may change in aggregate when visual evidence is degraded while remaining structurally inconsistent within individual examples. We study answer-level reliability along five-step, question-conditioned evidence-loss trajectories. Using a frozen Qwen2.5-VL-3B-Instruct model, we construct 176 accepted GQA-derived trajectories (880 masking conditions) by progressively masking scene-graph-localized question-critical regions. Native sequence confidence has an evidence monotonicity violation rate (EMVR) of 0.436, and 92.0% of trajectories contain at least one adjacent violation. A matched non-critical-region control shows that full critical masking reduces accuracy by 28.2 percentage points, compared with 0.6 points for equally sized non-critical masks; the paired difference is 27.6 points (95% CI [20.0, 34.7]). We train a lightweight post-hoc reliability head on frozen hidden states, sequence confidence, and entropy. Adding evidence-order supervision to binary cross-entropy (BCE) reduces masking EMVR from 0.330 to 0.303 (paired difference -0.027, 95% CI [-0.044, -0.010]). The same mask-trained objective reduces EMVR from 0.449 to 0.402 on held-out question IDs under unseen local Gaussian blur (difference -0.0468, 95% CI [-0.0739, -0.0199]). AUROC, Brier, and AURC differences between the two learned heads are statistically inconclusive, and native confidence remains stronger for selective-risk ranking. The results separate evidence-order consistency from conventional correctness discrimination rather than establishing generic confidence superiority.
[CV-87] Characterizing Text Branch Sensitivity in Medical Vision-Language Segmentation via Evidence Decoupling
链接: https://arxiv.org/abs/2609.02663
作者: Ziquan Liu,Zhewei Zhu,Xuyang Shi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 16 pages, 3 figures
Abstract:Pretrained vision-language models (VLMs) have shown promising performance in medical image segmentation by incorporating clinical text. However, it remains unclear how much textual information actually contributes to pixel-level predictions. In this work, we systematically investigate the role of text in multimodal medical image segmentation. We first analyze several commonly used fusion strategies and find that segmentation performance is largely insensitive to the choice of fusion module. To further understand modality interactions, we propose an Evidence Decoupling Decoder (EDD) based on evidential deep learning and deep supervision. EDD serves as an internal representation analysis tool that decomposes image evidence and text-modulated evidence throughout the decoding process while maintaining competitive segmentation performance. Experimental results show that the sensitivity to text perturbation varies substantially across datasets. On BUSI and BTMRI, removing text causes catastrophic performance drops, indicating strong model reliance on textual input. On ISIC and Kvasir-SEG, text exerts relatively marginal influence. We further find that text affects predictions mainly through global semantic modulation rather than independent spatial localization, and that the specific semantic components driving text sensitivity differ across datasets. These findings provide a deeper understanding of modality interaction in multimodal medical image segmentation and offer practical insights for future model design.
[CV-88] GLOSS: Geometric Local Self-Similarity Learning for Faithful Reference-Guided Texture Fill ECAI
链接: https://arxiv.org/abs/2608.25461
作者: Chenyue Cai,Anita Hu,James Lucas,Szymon Rusinkiewicz,Masha Shugrina
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 22 pages, 17 figures, project page this https URL
Abstract:Using conditional image generators, texture artists can explore many single-view looks for an existing 3D shape. Despite impressive progress, state-of-the-art generative methods still struggle to generate a full object texture while closely adhering to fine scale geometric detail and single view references, leaving little room for artists guidance. Furthermore, current automatic models lack the flexibility for artist to explore multiple textures from varied sources in an interactive and controllable manner. Unlike methods trained on large 3D datasets that generate full object textures from global guidance, our work explores a local and less data-hungry approach to texture with explicit artist control. We leverage the geometric self-similarity and geometry-texture correlation existing in many natural and man-made shapes; and train a shape-specific local texture generation and completion model. This model learns from existing image model priors and a single 3D shape, and is guided by attending to a set of geometry-aware reference patches. The trained shape-specific network can transfer any novel reference to the full target object texture through patchwise inpainting. We show improved or comparable quality to strong image-conditioned texture generation baselines, suggesting local texturing as a promising research direction. Our model also enables local geometry-conditioned texture inpainting, guided by artist-selected references, and generalizes to PBR materials and unseen meshes for texture transfer. We piloted our novel texture fill capability as a Blender addon with several 3D texturing professionals who reported positive feedback on the model’s controllability, practical usefulness, and creative affordances. Comments: 22 pages, 17 figures, project page this https URL Subjects: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG) Cite as: arXiv:2608.25461 [cs.GR] (or arXiv:2608.25461v1 [cs.GR] for this version) https://doi.org/10.48550/arXiv.2608.25461 Focus to learn more arXiv-issued DOI via DataCite
[CV-89] DCReg: Decoupled Characterization for Efficient Degenerate LiDAR Registration
链接: https://arxiv.org/abs/2509.06285
作者: Xiangcheng Hu,Xieyuanli Chen,Mingkai Jia,Jin Wu,Ping Tan,Steven L. Waslander
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 27 pages, 19 figures, 9 tables
Abstract:LiDAR point cloud registration is fundamental to robotic perception and navigation. In geometrically degenerate environments (e.g., corridors), registration becomes ill-conditioned: certain motion directions are weakly constrained, causing unstable solutions and degraded accuracy. Existing detect-then-mitigate methods fail to reliably detect, physically interpret, and stabilize this ill-conditioning without corrupting the optimization. We introduce DCReg (Decoupled Characterization for Ill-conditioned Registration), establishing a detect-characterize-mitigate paradigm that systematically addresses ill-conditioned registration via three innovations. First, DCReg achieves reliable ill-conditioning detection by employing Schur complement decomposition on the Hessian matrix. This decouples the 6-DoF registration into 3-DoF clean rotational and translational subspaces, eliminating coupling effects that mask degeneracy in full-Hessian analyses. Second, within these subspaces, we develop interpretable characterization techniques resolving eigen-basis ambiguities via basis alignment. This establishes stable mappings between eigenspaces and physical motion directions, providing actionable insights on which motions lack constraints and to what extent. Third, leveraging this spectral information, we design a targeted mitigation via a structured preconditioner. Guided by MAP regularization, we implement eigenvalue clamping exclusively within the preconditioner rather than modifying the original problem. This preserves the least-squares objective and minimizer, enabling efficient optimization via Preconditioned Conjugate Gradient with a single interpretable parameter. Experiments demonstrate DCReg achieves 20-50% higher long-duration localization accuracy and 5-30x speedups (up to 116x) over degeneracy-aware baselines across diverse environments. Code: this https URL
[CV-90] Candor-LR: A Dyadic Conversational Dataset for Audio-Visual Speech Recognition
链接: https://arxiv.org/abs/2609.10394
作者: Rishabh Jain,Aristeidis Papadopoulos,Zhaofeng Lin,Naomi Harte
类目: Audio and Speech Processing (eess.AS); Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Accepted to IEEE SLT 2026
Abstract:Current audio-visual speech recognition (AVSR) benchmarks, like LRS3, rely heavily on clean, scripted and rehearsed speech. They fail to reflect the complexity of natural conversation, which involves overlapping speech, spontaneous turn-taking, unscripted vocabulary and variable acoustic conditions. To shift the field toward realistic dialogue, we introduce Candor-LR, a conversational benchmark derived from the CANDOR corpus of 1,656 natural dyadic videoconferences. Our custom data preparation pipeline yields 713.5, 10.1, and 60.1 hours of training, validation, and test data, respectively. Evaluating pretrained AVSR models on Candor-LR reveals that audio-only accuracy drops sharply compared to LRS3, but visual cues compensate effectively, driving much larger performance gains on Candor-LR than on LRS3. Furthermore, training on this corpus significantly improves cross-domain robustness under both clean and noisy conditions, as its realistic conversational data captures broader audio-video features. We open-source our pipeline to ensure reproducibility, establishing Candor-LR as a challenging benchmark for conversational AVSR.
[CV-91] AVSRBench: A Multi-Condition AVSR Benchmark
链接: https://arxiv.org/abs/2609.10366
作者: Rishabh Jain,Naomi Harte
类目: Audio and Speech Processing (eess.AS); Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Accepted to IEEE SLT 2026
Abstract:While AVSR has achieved sub-1% word error rates on the standard LRS3 benchmark, its reliance on broadcast speech obscures whether this reflects true generalization or just domain adaptation. To investigate this gap, we evaluate three AVSR architectures across six conditions: controlled broadcast speech, fixed-grammar utterances, hyper-articulated Lombard speech, read speech from professional lipspeakers and non-professional speakers, and spontaneous multi-party video conversations. We find that visual-only performance deteriorates rapidly beyond broadcast domains, and audio-video fusion mainly benefits Lombard speech environments. Visual understanding degrades sharply at 90° profile views, with multimodal systems relying largely on acoustic fallback. Additionally, speaker articulation proves more critical than minor camera shifts, and LLM-based architectures suffer from poor out-of-domain generalization. Our work highlights a significant generalization gap in current AVSR research. To address this, we also introduce RoomReader-AV as a new benchmark for AVSR and release a unified data preprocessing pipeline to make comprehensive multi-condition evaluation accessible.
[CV-92] A statistical approach to bias in zero-shot learning: the lens of handwriting recognition
链接: https://arxiv.org/abs/2609.10084
作者: Clarence Chew,Gim Siang Chia,Sukalpa Chanda,Subhroshekhar Ghosh,Soumendu Sundar Mukherjee
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 28 pages, 2 figures
Abstract:Generalized zero-shot learning (GZSL) has emerged as an important paradigm for visual recognition systems that must generalize to classes that were not observed during training. Traditional GZSL techniques are limited by their applicability to a relatively small number of such unseen classes, scalability beyond which is challenging due to its well-known misclassification bias towards classes observed during training. In this work, we investigate the GZSL paradigm through the lens of zero-shot handwritten word recognition over extremely large vocabularies. We propose a statistical approach to rectifying this bias, which views any classical GZSL feature learner as a black box mechanism whose intrinsic bias in identifying the training status (seen vs. unseen) of a typical data point we aim to correct, similar to an out of distribution inferential problem. Our method leverages a simple two-stage hierarchical architecture, combining a classical GZSL blackbox in the first stage and an ensemble of lightweight Monte Carlo bias-correctors in the second. Once debiased, the classification of test data is undertaken only restricted to its predicted training status via well-founded statistical methods (eg nearest neighbour, logistic regression and random forests). We achieve relative accuracy improvements of over 20% in the classification of unseen words compared to established techniques. A key outcome is that word recognition over large scale vocabularies is amenable to a much lower dimensional representation (~15 dimensions). Our approach is underpinned by mathematical analysis that captures the essence of the statistical approach to bias correction. Our approach to bias rectification can be combined in a turn-key fashion with any classical GZSL learner as a blackbox, thereby suggesting a wide scope of applicability of this method for a wide variety of GZSL implementations in different domains.
[CV-93] FlowCPO: A Unified Divergence View of Preference Alignment for Flow Models
链接: https://arxiv.org/abs/2609.09905
作者: Yansen Han,Shengyi Liao,Peng Sun,Deyuan Liu,Yuanxing Zhang,Pengfei Wan,Tao Lin
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Preference alignment for flow and diffusion models now spans online reinforcement learning and offline preference optimization, but the relation between these methods remains unclear. In particular, existing forward-process alignment methods require fresh samples from the current model, while offline methods based on fixed preference pairs rely primarily on positive-only fine-tuning or DPO-style likelihood-ratio surrogates. We organize these approaches through a divergence-based framework and introduce FlowCPO, an offline forward-KL objective that uses both preferred and dispreferred samples without online rollouts. For linear interpolation, we show under explicit regularity conditions that the forward-KL objective is bounded by a contrastive flow matching loss, yielding a tractable surrogate on fixed data. We further show that this loss is nonnegative, whereas the signed regression loss of simplified FlowDPO can be unbounded below. In the in-domain setting, FlowCPO achieves higher mean GenEval and OCR scores than the evaluated baselines, reaching 0.84 and 0.87 versus 0.81 and 0.74 for FlowDPO at CFG 3.0. In the out-of-domain setting, the results are mixed, with the best GenEval result but lower reward scores than RFT on several metrics.
[CV-94] Morphological Decoupling-Based Skeletal Classification for Clinical Assessment of Malocclusion
链接: https://arxiv.org/abs/2609.09801
作者: Zhichun Jin,Zhicheng He,Hao Xu,Dongyang Li,Lin Wang,Hongliang Ren,Long Bai
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by Biocybernetics and Biomedical Engineering
Abstract:Malocclusion skeletal grading is a fundamental task in orthodontics, critical for diagnosis and treatment planning. Traditionally, cone-beam computed tomography (CBCT) is used for visual measurement, and the reconstructed lateral cephalograms are handed over to expert dentists for diagnosis. However, manual review is time-consuming, labor-intensive, and subject to inter-operator variability. Therefore, an automatic CBCT-based system is needed for reliable malocclusion skeletal grading. In this case, we develop TeethGNN, a novel graph-based framework designed to combine CBCT image features with morphological information for accurate and efficient malocclusion grading. TeethGNN utilizes a decoupled learnable decoder to directly predict key morphological indicators from CBCT images, eliminating the need for manual measurements. These morphological features are then fused with image features using a graph neural network (GNN), which effectively models the relationships between the modalities. To further enhance robustness and calibration, we introduce a collaborative calibration strategy. This strategy combines multi-scale graph adversarial perturbation for explicit calibration and nonlinear topological graph calibration for implicit confidence adjustment. Extensive experiments and ablation studies on our collected clinical dataset demonstrate that our malocclusion measurement system achieves 77.08% in accuracy and 89.61% in AUC, outperforming the compared state-of-the-art methods. These results validate the effectiveness of graph-based multimodal fusion and collaborative calibration in improving malocclusion grading performance. Our system shows strong potential for advancing computer-aided orthodontic diagnosis, providing an accurate and reliable solution for vision-based clinical measurement and diagnosis.
[CV-95] Myocardial Strain Drift Correction in Deep Learning Based Ultrasound Tracking
链接: https://arxiv.org/abs/2609.09577
作者: Thierry Judge,Nicolas Duchateau,Andreas Østvik,Havard Dalen,Bjørnar Grenne,Pierre-Yves Courand,Lasse Lovstakken,Pierre-Marc Jodoin,Olivier Bernard
类目: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: STACOM 2026, 10 pages
Abstract:Myocardial strain from echocardiography is a key biomarker for cardiac function. Recent deep learning methods show strong performance for myocardial motion tracking but often lack physiological constraints, leading to temporal drift across the cardiac cycle. Consequently, tracked points may not return to their relative initial positions at the end of each cardiac cycle, producing inaccurate strain estimates and even divergence in some cases. We propose a deep learning framework that compensates for drift during myocardial tracking. We extend a state-of-the-art echocardiographic tracking method (TAS-Net) with persistent memory tokens that share information across sliding windows over full cardiac cycles. A teacher-student fine-tuning strategy on real echocardiographic data then enforces physiologically consistent cyclic motion while preserving tracking accuracy. Experiments show reduced global and regional strain drift, improved agreement with clinical references, and better test-retest reproducibility, supporting more reliable myocardial strain estimation in clinical practice.
[CV-96] CHIMERA Challenge Task 2 and 3: Response Subtypes Classification and Progression Survival Prediction in Bladder Cancer Patients using Multimodal Datasets
链接: https://arxiv.org/abs/2609.09510
作者: Catherine Chia,Tongjie Wang,Robert Spaans,Maryam Mohammadlou,Farbod Khoraminia,J. Alberto Nakauma-González,Adam Kowalewski,Parandzem Khachatryan,Domingos Oliveira,Khrystyna Faryna,CHIMERA Challenge Consortium,Marlies Wakkee,Sita Vermeulen,Tahlita Zuiverloon,Nadieh Khalili
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Quantitative Methods (q-bio.QM)
备注: 53 pages, 9 figures, including supplementary material. Catherine Chia and Tongjie Wang contributed equally and share first authorship. Submitted to Medical Image Analysis
Abstract:High-risk non-muscle-invasive bladder cancer (HR-NMIBC) carries substantial risks of recurrence and progression, while current clinical risk stratification remains limited. CHIMERA was established as a multimodal AI challenge to benchmark prediction in HR-NMIBC under standardized evaluation. Task BRS predicts RNA-seq-defined BCG Response Subtypes from histopathology and structured clinicopathological data, whereas Task Progression models time-to-progression using histopathology, structured data, and RNA sequencing. A multimodal dataset of 368 patients was divided into public training and hidden validation and test sets. In total, 159 submissions were made, and 13 top-performing models were selected for benchmarking. The best models achieved a weighted F1 score of 0.73 for Task BRS and a C-index of 0.68 for Task Progression. Post-challenge analyses revealed task-dependent modality contributions, cohort-dependent performance degradation, and sensitivity to missing structured data. In Task BRS, histopathology partly compensated for pathology-derived structured variables, whereas progression models showed greater dependence on complementary inputs. Cross-model error analysis further identified patients that were consistently difficult across different architectures, with T1 substage associated with prediction difficulty. These findings highlight barriers to transportability and the importance of missingness-aware modeling and independent multi-institutional validation. CHIMERA provides a standardized multimodal benchmark for bladder cancer and a framework for studying not only model performance, but also robustness, information sufficiency, and patient-level prediction failure.
[CV-97] Reliability-Aware Hybrid-K Ensemble Selection for Cervical Cytology Classification: Integrating Discrimination Calibration and Selective Prediction
链接: https://arxiv.org/abs/2609.09189
作者: Nisreen Albzour,Sarah S. Lam
类目: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:High classification accuracy alone is insufficient for clinical image analysis, where calibrated confidence and reliable uncertainty estimates are essential. This study proposes a reliability-aware Hybrid-K ensemble selection framework for multiclass cervical cytology classification using the SIPaKMeD dataset. Nine deep learning architectures were evaluated using a fixed stratified five-fold partition and three training seeds. After post-hoc temperature scaling, models were assessed using macro-F1, accuracy, AUROC, expected calibration error (ECE), worst-class ECE (WC-ECE), area under the risk-coverage curve (AURC), Brier score, and negative log-likelihood (NLL). Models were ranked using an equal-weight composite score, and Hybrid-K ensembles were formed from the top-ranked models using soft voting. Robustness was examined using 5,000 Dirichlet-sampled metric-weight vectors, leave-one-metric-out analysis, and corrected paired testing across 15 fold-by-seed evaluations. The final Hybrid-2 ensemble, comprising Swin-Tiny and TinyViT-5M, reduced AURC by 43%, NLL by 17%, and WC-ECE by 36% relative to the best individual model. It was selected in 96.8% of random weighting scenarios, remained unchanged across all leave-one-metric-out analyses, and improved the full composite score. However, per-metric gains were not statistically significant after Holm-Bonferroni correction (all adjusted p = 0.168). Because post-hoc calibration did not use a fully independent calibration set, calibration-dependent results should be interpreted as exploratory internal estimates. Overall, the framework identified a compact ensemble robust to alternative metric weightings and improved reliability point estimates under internal validation on a single dataset.
[CV-98] Diverse Instance Generation via Diffusion Models for Enhanced Few-Shot Object Detection in Remote Sensing Images
链接: https://arxiv.org/abs/2511.18031
作者: Yanxing Liu,Jiancheng Pan,Jianwei Yang,Tiancheng Chen,Peiling Zhou,Bingchen Zhang
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 6 pages, 2 figures
Abstract:Few-shot object detection (FSOD) aims to detect novel instances with only a limited number of labeled training samples, presenting a challenge that is particularly prominent in numerous remote sensing applications such as endangered species monitoring and disaster assessment. Existing FSOD methods for remote sensing images (RSIs) have achieved promising progress but remain constrained by the limited diversity of instances. To address this issue, we propose a novel framework that can leverage a diffusion model pretrained on large-scale natural images to synthesize diverse remote sensing instances, thereby improving the performance of few-shot object detectors. Instead of directly synthesizing complete remote sensing images, we first generate instance-level slices via a specialized slice-to-slice module, and then embed these slices into full-scale imagery for enhanced data augmentation. To further adapt diffusion models for remote sensing scenarios, we develop a class-agnostic image inversion module that can invert remote sensing instance slices into semantic space. Additionally, we introduce contrastive loss to semantically align the synthesized images with their corresponding classes. Experimental results show that our method hasachieved an average performance improvement of 4.4% across multiple datasets and various approaches. Ablation experiments indicate that the elaborately designed inversion module can effectively enhance the performance of FSOD methods, and the semantic contrastive loss can further boost the performance.
[CV-99] Control Copy-Paste: Controllable Diffusion-Based Augmentation Method for Remote Sensing Few-Shot Object Detection
链接: https://arxiv.org/abs/2507.21816
作者: Yanxing Liu,Jiancheng Pan,Bingchen Zhang
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 5 Pages, 3 figures
Abstract:Few-shot object detection (FSOD) for optical remote sensing images aims to detect rare objects with only a few annotated bounding boxes. The limited training data makes it difficult to represent the data distribution of realistic remote sensing scenes, which results in the notorious overfitting problem. Current researchers have begun to enhance the diversity of few-shot novel instances by leveraging diffusion models to solve the overfitting problem. However, naively increasing the diversity of objects is insufficient, as surrounding contexts also play a crucial role in object detection, and in cases where the object diversity is sufficient, the detector tends to overfit to monotonous contexts. Accordingly, we propose Control Copy-Paste, a controllable diffusion-based method to enhance the performance of FSOD by leveraging diverse contextual information. Specifically, we seamlessly inject a few-shot novel objects into images with diverse contexts by a conditional diffusion model. We also develop an orientation alignment strategy to mitigate the integration distortion caused by varying aspect ratios of instances. Experiments on the public DIOR dataset demonstrate that our method can improve detection performance by an average of 10.76%.
人工智能
[AI-0] JarvisGUI: Towards Cross-Device GUI Agents with Dynamic Task Composition EMNLP2026
链接: https://arxiv.org/abs/2609.10451
作者: Zixiang Chen,Yuheng Lu,Zihao Cheng,Zeming Liu,Jizeng Bai,Ziye Huang,Zhiyin Lin,Zihan Li,Yuhang Guo,Yunhong Wang,Haifeng Wang
类目: Artificial Intelligence (cs.AI)
备注: Accepted to EMNLP 2026 (Main Conference)
Abstract:Real-world GUI usage frequently involves workflows that span multiple devices and platforms, requiring the transfer of intermediate results, maintenance of shared state, and coordination across heterogeneous environments. However, existing GUI benchmarks overwhelmingly evaluate agents on single-device, statically defined tasks, thus leaving such cross-device capabilities largely unexamined, resulting in an overly optimistic assessment of agents’ readiness for real-world usage. We introduce JarvisGUI, a dynamic benchmark that evaluates GUI agents on cross-device workflows requiring coordinated interaction across heterogeneous platforms, including Android, Windows, and Ubuntu. Specifically, JarvisGUI formulates GUI tasks as input-output transformations under a lightweight type system, which allows us to automatically compose multi-step, cross-device workflows and dynamically evaluate agent performance within a unified framework. By evaluating agents in virtual environments spanning multiple operating systems, JarvisGUI reveals that state-of-the-art open-source GUI agents struggle with the state-transfer awareness, cross-platform contextual reasoning, and long-horizon dependency management required for real-world workflows, exposing a critical capability gap invisible to existing benchmarks.
[AI-1] Forgetting Only What Matters: Layer-Selective Unlearning toward Robust LLM s AACL
链接: https://arxiv.org/abs/2609.10439
作者: Ravi Ranjan,Olivera Kotevska,Agoritsa Polyzou
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 22 pages, 6 figures, 11 tables, AACL-IJCNLP 2026, conference paper
Abstract:Large Language Models (LLMs) can memorize and reproduce sensitive, copyrighted, or otherwise undesirable training content, creating privacy, safety, and regulatory concerns. Machine unlearning offers a practical alternative to full retraining, but many existing methods apply broad or fixed parameter updates that can degrade utility and remain brittle under deployment changes such as post-training quantization, where forgotten knowledge may partially re-emerge. We propose Forgetting Only What Matters via Unlearning Layers (FOM-UL), a layer-level unlearning framework that selects transformer layers using a forget-to-retain significance score. This score identifies layers with high influence on the forget set and low sensitivity to the retain set, allowing FOM-UL to concentrate updates where they are most effective while leaving most of the model unchanged. This targeted update strategy improves the forgetting-utility trade-off and provides an empirical path toward quantization-resilient unlearning by reducing the chance that small, diffuse updates are erased by low-bit rounding. Across TOFU, KnowUnDo, and MUSE-style evaluations, FOM-UL reduces residual memorization compared with strong GA, NPO, KLD, SURE, ReLearn, and LUNAR-based baselines while preserving retain-set utility close to the vanilla model. Under 8-bit and 4-bit post-training quantization, FOM-UL maintains stronger memorization suppression and utility preservation than competing methods, and adversarial prompt evaluations show lower recovery of forgotten content. Overall, FOM-UL provides an efficient unlearning strategy that improves targeted forgetting, utility preservation, and deployment robustness without claiming formal guarantees of erasure.
[AI-2] Emergency Department Revisit Quality Review Screening: Exploring Human Decision-Making and Artificial Intelligence Support
链接: https://arxiv.org/abs/2609.10421
作者: Jonathan A. Handler,Marlene I. Robles-Granda,Jacob E. Mefford,Jeremy S. McGarvey,Gregory S. Podolej,Colleen J. Klein,Matthew D. Dalstrom,William F. Bond
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注: 12 pages, 1 figure
Abstract:Background: Emergency Department (ED) return visits are commonly reviewed for quality assurance, but are often limited (e.g., to revisits within 48-72 hours) to increase actionable finding yield while minimizing chart review burden. Those limitations may lead to missed quality improvement opportunities. Methods: We conducted an exploratory, retrospective study of randomly selected ED visits to a multihospital health system having an ED revisit within 1-14 days to the same health system. Given only each visit’s primary diagnosis, raters (2-3 clinicians and GPT-4 large language model [LLM]) assessed characteristics of the diagnosis pairs, including the “target”: whether a pair warranted further assessment. Informed by rater response analyses, an algorithm leveraging an LLM-populated knowledge graph (“KGA”) was created to automatically screen for potentially concerning pairs, then preliminarily assessed. Results: 99 diagnosis pairs were included. GPT-4 responses poorly correlated to clinician raters, rating nearly all (94%) pairs as warranting follow-up (4.4-13.3 times more than clinicians). However, prompt engineering was minimal. Among clinician raters, revisit medical gravity was consistently significantly associated with the target, while a differential diagnosis/complication composite was significantly associated on unadjusted, but not adjusted (though less powered) analysis. The KGA achieved 83-100% positive predictive value for at least one clinician rater determining further assessment was warranted based on the diagnosis pair. Conclusion: These results can inform next steps for improving screening with LLMs like ChatGPT. Further research is warranted to validate this preliminary work’s finding that the KGA may enable enhancing the scope and yield of screening without substantially increasing reviewer workload. Comments: 12 pages, 1 figure Subjects: Computers and Society (cs.CY); Artificial Intelligence (cs.AI) ACMclasses: I.2.1; J.3 Cite as: arXiv:2609.10421 [cs.CY] (or arXiv:2609.10421v1 [cs.CY] for this version) https://doi.org/10.48550/arXiv.2609.10421 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Jonathan Handler [view email] [v1] Wed, 9 Sep 2026 16:37:35 UTC (556 KB)
[AI-3] Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLM s
链接: https://arxiv.org/abs/2609.10413
作者: Ansuman Mullick,Eray Tüzün
类目: Artificial Intelligence (cs.AI)
备注: Preprint, under review. 2 figures. Code, benchmark and run logs: this https URL
Abstract:Current LLM memory systems treat all personal facts identically, so stores grow without bound while retrieval precision degrades. The core challenge is lifecycle management: which memories should persist, which should be replaced, and at what rate, conditioned on the behavioral type of each fact. Fortunate Recall (FR) is a composable policy layer that classifies personal facts into a 10+1 behavioral ontology and applies category-specific lifecycle policies (differential temporal decay, slot-key supersession, event-time validity, and category-aware retrieval routing) as deterministic functions over LLM-extracted metadata. FR-Bank, our infrastructure-independent implementation, reaches a 76.9% pass rate on LifecycleBench, a new 516-question temporal-disambiguation benchmark, ahead of Mem0, A-MEM, Memory-R1, and MemoryOS (61% to 70.5%), and 75.2% on the full LongMemEval-S under the canonical Wu et al. judge protocol, so lifecycle policies impose no measurable cost on standard retrieval. A pre-registered ablation locates the gains: replacing the typed layer with three generic lifecycle primitives leaves correctness statistically unchanged (-1.7pp, 95% CI [-6.0, +2.7]), so the generic lifecycle metadata carries the correctness advantage, while the behavioral ontology carries calibration, halving downstream confabulation (12.0% vs 24.2%, p0.001). End-to-end, FR-Bank cuts confabulation from Mem0’s 45.1% to 22.4% over answered queries and from 32.2% to 13.0% over all queries while answering more of them correctly (31.2% vs 18.6%); the ranking replicates on the open-weight Kimi K2.5. The decomposition transfers to BEAM, an independently built benchmark: 46.8% correct vs Mem0’s 32.9% over 280 questions, with the ontology’s benefit concentrated in contradiction resolution and saturating near seven policy clusters. The ontology, benchmark, and code are released.
[AI-4] OmniMed-FL: A Robust Multimodal Federated Learning Framework for Clinical Diagnosis ALT
链接: https://arxiv.org/abs/2609.10364
作者: Ayush Debnath,Ruelia Saha,Sudip Misra
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted in IEEE Globecom 2026, E-Health
Abstract:Simultaneous assessment of medical imaging and patient records is often required in clinical diagnosis. However, standard machine learning algorithms cannot analyze these data types together. Meanwhile, compliance with HIPAA and GDPR can constrain centralized aggregation of sensitive patient data. This leaves a crucial void of secure fusion of visual and textual context across distant networks. Thus, we present OmniMed-FL, a controlled systems study of multimodal federated learning for five-class clinical condition classification (Normal, Pneumonia, COVID-19, Pleural Effusion, Cardiomegaly). Our proxy corpus pairs 3,000 public chest radiographs with 3,000 class-conditioned synthetic notes, matched by class, not by patient. The framework benchmarks eight fusion strategies, three initializations, four missing-text imputation rules, and matched federated baselines under non-IID Dirichlet partitioning across 3 to 20 hospital clients. As all notes are synthetic and pairing is not patient-level, these are descriptive proxy comparisons, not estimates of diagnostic performance or deployment readiness. Within those limits with clients ( K=5 ) and severe skew ( \alpha=0.1 ), local-only training achieves a macro-F1 score of 0.297, FedAvg achieves 0.662\pm0.074 , FedProx 0.737\pm0.085 , a matched FedMME-style one-shot ensemble 0.647\pm0.080 , and our SCAFFOLD-AdamW adaptation 0.070\pm0.015 , the 0.075 FedProx-FedAvg gap falling inside the wider of the two two-seed standard deviations. Over a 4\times3 grid, label skew costs up to 0.27 F1 whereas a near-sevenfold client increase costs at most 0.10, while bidirectional volume grows linearly to 183.5 GiB at K=20 . Multimodal fusion leads on both corpora, scoring 0.956 against 0.934 for text and 0.664 for images on the synthetic corpus and 0.906 against 0.880 and 0.737 on the radiograph corpus, for 2.3\times the model state of text alone.
[AI-5] Cyber-Financial Contagion: Modeling the Propagation of an AI Vendor Compromise Through the Banking System
链接: https://arxiv.org/abs/2609.10350
作者: Alex Leytes
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (cs.LG)
备注: 11 fig and 10 tables
Abstract:The banking system now depends on a small set of shared artificial intelligence vendors for fraud screening, credit decisioning, anti-money-laundering triage, customer analytics, and internal decision support. This paper studies how a compromise inside one of those vendors can propagate along a chain of operational, informational, and financial linkages until it triggers losses that look, from the outside, like a classical banking crisis. We build a four-layer heterogeneous network that couples AI vendors, financial institutions, interbank exposures, and customer accounts, and we propose CFC-Prop, a stochastic epidemic-and-clearing model that runs on that network. On a synthetic dataset with 60 vendors, 220 banks, roughly 2,500 vendor-bank service edges, and 1,400 interbank exposures, CFC-Prop reproduces the heavy-tailed loss distributions and the sharp dependence on patch latency that are consistent with prior cyber-financial evidence. We also train an early-warning model, CFC-GNN, that uses vendor-side incident telemetry and graph structure to flag high-cascade-risk vendors before impact. Across four baselines the proposed model reaches AUROC 0.82 and AUPRC 0.60 while keeping calibration errors bounded. We release the full code, synthetic data, and reproducible scripts. The results argue that cyber concentration among AI vendors is a first-order financial-stability problem and give supervisors a concrete quantitative tool for reasoning about it.
[AI-6] RACE: Training Reasoning Agents for Causal Exploration with Synthesized Rewards
链接: https://arxiv.org/abs/2609.10315
作者: Rui Sun,Zhan Shi,Bing He
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Reinforcement learning with verifiable rewards (RLVR) has advanced language-model reasoning in domains such as mathematics and code, where objective answers are inexpensive to check. Diagnostic reasoning over complex data lacks this advantage: establishing the true cause of an anomaly often requires costly expert investigation and may remain ambiguous after the fact. We ask whether this asymmetry of verification can instead be engineered. We sample an intervention, inject it into a controlled simulator, and generate the observations it would produce. The hidden intervention provides an oracle label and objective reward, while the agent must still investigate noisy, confounded, and distributed evidence. We instantiate this approach in TRACE, a digital-advertising diagnostic environment with 12 root causes and fine-grained segment attribution. Agents investigate each episode using Python and SQL and must identify both the root cause and, when applicable, the affected segment assignment. On a held-out 235-episode test set, the strongest prompted baseline, Claude Opus 5, reaches 0.686 FullAttr@1. Supervised fine-tuning raises Qwen3.5-35B-A3B from 0.159 to 0.637, and subsequent RL with synthesized rewards reaches 0.757, outperforming all evaluated prompted baselines, including frontier closed-source models and a prompted Qwen3.5-122B-A10B model. The resulting policy also uses substantially fewer tool calls than the prompted 35B base. These results provide evidence that access to a scalable, objective training signal can be a more important constraint than model scale alone. More broadly, simulation-based verification can make otherwise ambiguous diagnostic reasoning tasks amenable to scalable reinforcement learning. Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2609.10315 [cs.AI] (or arXiv:2609.10315v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.10315 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-7] Learning Intrusion Response Strategies for OT Systems
链接: https://arxiv.org/abs/2609.10298
作者: Duc Huy Le,Rolf Stadler
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: A version of this paper has been published at the 22nd International Conference on Network and Service Management (CNSM2026)
Abstract:Cyberattacks against Operational Technology (OT) systems, which monitor and control industrial processes, pose an increasing threat to essential societal services. For this reason, developing automated intrusion response strategies is highly important. In this paper, we present a formal model of an OT intrusion response use case using the POMDP framework. It includes a realistic model of partial observability that is based on traffic measurements. This approach allows us to develop tractable, learning-based solution methods for automated intrusion response, which are based on PPO. We evaluate the obtained response strategies on an emulated OT system and find that they are effective against several types of MITRE attacks for the studied use case.
[AI-8] What Should an Agent Forget? Separating What Is Stored from What Is Used
链接: https://arxiv.org/abs/2609.10263
作者: Yuhang Li,Yuchen Li
类目: Artificial Intelligence (cs.AI)
备注: 8 pages, 3 figures, 3 tables
Abstract:Persistent language agents need stored experience to remain available across time, while each answer requires evidence suited to a particular question. A superseded fact can mislead a current-state answer and still be essential for a historical query. We present RD-Forget, a training-free framework that separates what an agent stores from what it uses. A retained source archive preserves observations, and a query-conditioned memory view controls their influence on the current answer. A frozen language-model curator extracts relevant evidence, groups facts into semantic slots, and preserves the relations needed for multi-hop reasoning. Same-slot replacement links suppress superseded values in current-state contexts, while intent-aware retrieval makes earlier evidence eligible again. A rate-distortion formulation guides construction of the answer-time view within a memory budget. Experiments span conversational memory, knowledge updating, fact consolidation, long-context reasoning, and personalization under a shared answering pipeline. The results associate accurate answers with both query-relevant evidence construction and control over obsolete alternatives. Configurations without forgetting or query conditioning have the largest score deficits, while slot grouping, historical access, and relation preservation contribute complementary functions. Retaining history while selectively controlling its use offers a practical way to accommodate changing facts and future questions.
[AI-9] A-JIT: Agent ic Just-In-Time Software Construction
链接: https://arxiv.org/abs/2609.10248
作者: Mark Marron,Earl T. Barr
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: Technical report for presentation at VMIL 2026
Abstract:Traditional software delivery assumes a static paradigm: code is constructed prior to execution and deployed as a fixed artifact. We present Agentic Just-In-Time Software Construction (A-JIT), a paradigm that replaces static binaries with dynamic, software systems that can perpetually evolve to meet changing demands. In A-JIT, an application is an integrated assembly comprising code, a runtime harness, and an embedded AI agent that continuously observes system usage and live execution traces. Much like a traditional JIT compiler specializes machine code to runtime execution paths, A-JIT specializes software logic, workflows, and tool interfaces to meet the specific needs of the end-user. By integrating synthesis directly into the ambient application lifecycle, A-JIT enables applications to dynamically construct missing implementations, generate new capabilities on the fly, and continuously adapt to end-user behavior. We demonstrate how this model supports trace-driven human-AI co-construction and opens a new design space for adaptive, self-evolving software.
[AI-10] Hierarchical and Permutation-Invariant Feature Transformation Learning via Policy-Guided Embedding Search CIKM2026
链接: https://arxiv.org/abs/2609.10225
作者: Rui Liu,Tao Zhe,Yanyong Huang,Sankha Narayan Guria,Xiao Luo,Wei Fan,Yanjie Fu,Dongjie Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: This paper has been accepted for publication at CIKM 2026
Abstract:Feature transformation improves predictive performance on tabular data by constructing informative abstractions from raw features. Recent generative approaches encode transformation knowledge into continuous embedding spaces for efficient exploration of candidate strategies, but face three key limitations: (1) overlooking hierarchical relationships between low-level features, operations, and high-level abstractions; (2) enforcing order-sensitive embeddings on inherently permutation-invariant transformation sequences, thereby introducing systematic bias; and (3) relying on gradient-based search, which is ill-suited to non-convex transformation spaces. We propose a framework with two complementary components. First, a permutation-invariant hierarchical module captures interactions across features, operations, and abstraction levels, with a self-attention pooling mechanism that maps semantically equivalent structures to consistent embeddings aligned with downstream performance. Second, a policy-guided multi-objective reinforcement learning strategy initializes the search from empirically strong seeds and jointly optimizes predictive accuracy and transformation efficiency. Extensive experiments on diverse tabular benchmarks demonstrate the effectiveness and robustness of our framework against strong baselines. Our code and data are publicly available at: this https URL.
[AI-11] Why Sample What You Can Enumerate? Exact Policy Optimization for Genomic Tool Selection
链接: https://arxiv.org/abs/2609.10221
作者: Haoyue Liu,Xiaoyu Ma,Ye Chen,Zhichao Wang,Xiaoying Tang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Reinforcement learning over a frozen reasoner has become a common recipe for teaching a policy which external tools to invoke. We show that this recipe becomes structurally mismatched in specialist scientific settings where the complete tool-subset space is enumerable. There, a small set of recurring computational capabilities covers the domain, so the space of tool subsets is combinatorial yet small enough to enumerate, and GRPO still estimates an action expectation from a handful of sampled rollouts. Worse, the approximation degrades as training succeeds: as the policy concentrates on preferred subsets it resamples them, sampled rewards collide, and the group-normalized advantage vanishes. On genomic reasoning the fraction of questions yielding no reward signal rises from 0.2% under a uniform reference policy to 20.8% after GRPO training. As a remedy, we introduce FGPO (Full-Group Policy Optimization), which (1) scores every tool subset and optimizes the exact action expectation, so each update sees the complete action space, and (2) precomputes the reward of each question–subset pair into an exhaustive table, removing frozen-reasoner calls from the training loop entirely. Across five frozen reasoners and three genomic benchmarks, FGPO outperforms GRPO in all 15 settings by 6.75 points on average and up to 14.20, while a standard on-demand GRPO schedule would require 2.4 times as many frozen-reasoner reward evaluations and, on GenomeQA, FGPO cuts invoked tools per question from 2.36 to 1.40.
[AI-12] Can AI Agents Deliver Verifiable Network-Wide Outcomes Across Authority Boundaries?
链接: https://arxiv.org/abs/2609.10181
作者: Tianzhu Zhang,Chih-Kai Huang,Meikang Qiu
类目: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注:
Abstract:AI agents are increasingly involved in network automation, where they can initiate configuration changes through mediated operational interfaces and assess the resulting state. Nonetheless, operational networks usually span many devices and administrative domains. Realizing an operator’s intent requires coordinating agents with distinct authority scopes that define the resources they can access, the operations they can invoke, and the network state they can observe. This division limits the blast radius of an erroneous action but fragments the evidence needed to assess the network-wide outcome. Successful execution of a configuration action proposed by one agent does not establish that remote devices responded as intended or that routing changes reached the required devices. A valid observation may also become stale after a subsequent change. Before the coordinated operation can be declared complete, a trusted assurance layer must collect current observations from the required scopes and determine whether they collectively support the operator’s intended network-wide outcome. To address the completion admission problem, we present EvidenceNet, a runtime assurance layer for deciding whether coordinated agent operations have achieved an operator’s network intent. Its broker collects the post-change observations required by a completion contract, and its admission gate checks that the evidence comes from the required scopes, remains current, and satisfies the task rules. A verifier agent provides an additional assessment of the observation content. Experiments on live routing networks show that post-change state checks recognize successful outcomes that configuration-action records alone cannot establish. Controlled interventions further show that EvidenceNet rejects completion when otherwise satisfactory observations have the wrong source, have been substituted, or are stale. Subjects: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI); Systems and Control (eess.SY) Cite as: arXiv:2609.10181 [cs.NI] (or arXiv:2609.10181v1 [cs.NI] for this version) https://doi.org/10.48550/arXiv.2609.10181 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-13] Beyond Surface Imitation: Contrastive Modeling for Reasoning Path Alignment in Multimodal In-Context Learning
链接: https://arxiv.org/abs/2609.10177
作者: Mingbo Yang,Wenqiang Wang,Zhaolu Kang,Peng Chen,Yannan Chen,Sunshang Wang,Yan Xiao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:In-context learning (ICL) is widely used in multimodal large language models (MLLMs) and achieves strong performance across a wide range of multimodal tasks. However, existing multimodal ICL methods often rely on surface level imitation of in-context demonstrations, making it difficult for MLLMs to align their responses with the reasoning path required by the given multimodal input. This limitation becomes more pronounced in complex multimodal tasks, thereby restricting further improvements in MLLM performance. To address this issue, we propose a new multimodal ICL framework that combines contrastive demonstration modeling with the self-refinement capability of MLLMs. Specifically, our framework reformulates each demonstration by explicitly contrasting a suboptimal response with a better response under the same input, together with a reasoning path that reveals how the response should be refined. This contrastive formulation makes the reasoning path toward the desired response more explicit and guides the MLLM beyond superficial imitation. Furthermore, because effective refinement depends on the current response, we introduce a response-conditioned retrieval mechanism to select demonstrations whose reasoning paths are more relevant to the current response. In addition, we use a lightweight alignment controller to predict response quality and determine whether further refinement is needed. Experiments on three types of multimodal tasks show that the proposed framework consistently improves MLLM performance, with particularly notable gains on visual question answering (VQA).
[AI-14] Kernel-Managed Shared Memory for System-Wide Personalization
链接: https://arxiv.org/abs/2609.10144
作者: Ryan Lum,Yongfeng Zhang
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:AI systems become more useful when they can adapt to the people using them, but in multi-agent systems, useful context learned by one agent often remains unavailable to others. We present kernel-managed shared memory, a system-level abstraction in which specialized agents write structured, tagged memories while the agent-system kernel, not individual agents, governs retrieval, privacy enforcement, and prompt injection. We implement and evaluate this design on AIOS and compare it against three alternatives across three assistant models (GPT-4o, Llama-3.1:8B, Qwen-2.5:7B) and 1,800 total trials. Against an unmanaged external memory backend (Mem0) using identical underlying storage, kernel-managed retrieval and injection improve personalization scores by 2.4-4.0 points on a 5-point scale (e.g., 1.05 to 4.69 profile usage on GPT-4o), with every comparison significant at p 10^-18. Against standard retrieval-augmented injection, gains are similarly large and consistent across all three models. Against full, unfiltered context concatenation, a soft ceiling on available context rather than on response quality, kernel-managed injection statistically matches performance on two of three models and shows a small, model-specific deficit on the third, while using substantially shorter prompts: end-to-end latency is 15-61% lower across all three models, with corresponding reductions in per-call token usage and inference cost. These results indicate that centralizing memory management in the agent-system kernel, rather than leaving retrieval and privacy enforcement to individual agents, delivers most of the personalization benefit of unconstrained context at a fraction of its cost.
[AI-15] Agent -Based ML-LLM Fusion with Self-Optimizing Prompts for Plateau Weather Alerts
链接: https://arxiv.org/abs/2609.10135
作者: Shuai Yan,Yang Xu,Shan He
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted by ISPDS 2025
Abstract:To address insufficient contextualization, weak generalization, and poor scenario adaptation in tourism meteorological services, we propose SmartWeatherAgent–a unified three-stage architecture integrating intent recognition, hazard prediction, and reasoning-enhanced generation. The system fuses rule-based methods with large language models to parse queries at multiple granularities and employs a LightGBM model enriched with highland-specific features (e.g., wind speed abruptness rate), achieving an F1-Macro score of 0.605 with 1.60 ms latency on high-wind, precipitation, and low-temperature events. A 12-round micro-step prompt self-optimization loop boosts the composite warning quality score S_final from 4.2 (B01) to 8.9 (B12, +112%). Key improvements include a sharp rise in B08 from data source citation (6.5 - 8.5), sustained high performance in B10 via physical mechanism explanation, and a peak scientific rigor score of 9.2 in B12 through explicit uncertainty statements. The system autonomously generates structured warnings that integrate causal mechanisms, spatiotemporal evolution, quantitative evidence, regulatory references, and confidence statements–enhancing professional depth, logical rigor, and scientific soundness, and advancing meteorological services toward proactive perception, explainable decision-making, and intelligent agency.
[AI-16] Context operations to architecture modelling output from large language models and evaluation criteria for their use in systems engineering design
链接: https://arxiv.org/abs/2609.10132
作者: Vinicius Kaster Marini,Petter Krus
类目: ystems and Control (eess.SY); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 31 pages, 13 figures, 12 equations and 9 tables
Abstract:The development of generative artificial intelligence resources enables opportunities of speeding up systems and engineering design work. This contribution introduces a framework of formal operations for assembling context in LLM-based engineering design. This framework involves the assembly of modular context units, including policy prompts, reference units with persistence, and user questions with prompt vectoring. This approach enables the systematic structuring of interactions with generative models. A formal method for evaluating modelling-as-code LLM outputs is also presented, which enables the evaluation of compliance to intent from LLM answers and thereby asses the support from LLMs for systems architecture modelling.
[AI-17] A Trust-Network-Based Federated Learning Framework for Multi-Center Aging Clock Prediction
链接: https://arxiv.org/abs/2609.10108
作者: Chunxu Zhang,Bo Li,Wenliang Wang,Yang Liu,Di Jiang,Yuan Huang,Yo-ichi Nabeshima,Akinori Yamamura,Bo Yang,Qiang Yang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Aging clocks quantify biological aging and help characterize individual health status. What protein interactions are important for accurate aging clocks, and are they zeroth-order or higher-order? Addressing these questions requires learning from large molecular datasets distributed across medical centers, where privacy constraints prevent centralized data sharing. Federated learning offers a natural solution but faces four challenges in this setting: limited local sample sizes, sparse and directional inter-center trust, the need to retain discriminative age prediction while supporting interpretation, and model drift and forgetting under heterogeneous cross-center data. We propose TNFL, a trust-network-based federated learning framework that progressively propagates models along directed pairwise trust relations without centralized aggregation. TNFL combines an age-aware mixture-of-experts model with generative replay to preserve previously learned information and reduce forgetting and drift. Experiments across multiple molecular datasets show that TNFL enables effective aging-clock prediction with limited local data, provides interpretable age-dependent prediction patterns, and maintains stable performance across interaction orders. To investigate the biological questions, we analyze TNFL-identified pairwise protein interactions and their higher-order organization through functional and network analyses. The identified interactions repeatedly form coordinated higher-order subnetworks spanning multiple aging-related biological systems, with several proteins recurring across subnetworks. These findings suggest that TNFL captures molecular relationships beyond isolated pairwise associations and reveals coherent higher-order biological organization associated with aging. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.10108 [cs.LG] (or arXiv:2609.10108v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.10108 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-18] Beyond Training: A Feasibility Taxonomy for Inference-Time AI Governance
链接: https://arxiv.org/abs/2609.10105
作者: Samar Ansari
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:Compute governance today is a governance of training: the thresholds, reporting requirements, and frontier-AI regimes now in force attach to training compute and treat the trained model as the regulatory unit. That picture is incomplete: capability increasingly migrates to the deployment stage through inference-time scaling, agentic scaffolding, and compression onto consumer hardware. This paper asks which mechanisms are available once the regulatory object shifts from the training run to the inference call. We develop a feasibility taxonomy of twenty inference-time mechanisms across monitoring, verification, and enforcement, each rated on a four-point readiness scale against a documented four-vendor evidence base. We then stress the taxonomy against a two-dimensional adversary model (three capability tiers crossed with four adversary roles) and map each mechanism to four governance scenarios (domestic regulation, bilateral or multilateral coordination, industry self-regulation, and compute-marketplace governance). Fifteen of the twenty mechanisms have commercial technical substrates in production today, although governance-grade assurance and adversarial robustness vary substantially. The adversary analysis shows that this readiness holds only against a cooperative deployer and a low-to-medium-capability user: no mechanism rates adequate against a high-capability state-level deployer, and fine-tuning removes the model-internal components of the enforcement cluster, although platform-external controls can persist. A substitution analysis connects the taxonomy to a companion hardware paper as a conditional substitution principle describing when inference-stage and hardware-stage mechanisms provide comparable regulatory coverage under stated conditions. A second-rater reliability check on a random subset of the readiness ratings returned a quadratic-weighted Cohen’s kappa of 0.74.
[AI-19] Reference-Based Bias Detection in LLM s via Relative Representations of Hidden States
链接: https://arxiv.org/abs/2609.10060
作者: Marek Jeliński,Jan Dubiński,Maciej Chrabaszcz,Sebastian Cygert
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Existing bias auditing methods typically rely on model outputs, requiring costly benchmarks or judge models and potentially missing internal shifts that never appear in generated text. We propose a reference-based method that audits bias in hidden-state representations across related model variants, for example before and after fine-tuning. Because fine-tuning reshapes representation geometry, absolute hidden states are not directly comparable, so we encode each sentence by its similarities to a fixed set of anchor sentences, yielding relative representations in a shared comparison space. There we measure how target groups shift in their association with positive and negative attributes, a quantity we call the Representational Bias Shift \Delta B . Across three model families and the WildGuardMix, DecodingTrust and ToxiGen benchmarks, \Delta B correlates with output-level bias change in 15 of the 18 settings we test, reaching |r| = 0.84 ( p 0.001 ) under full fine-tuning and becoming more model-dependent under parameter-efficient adaptation. Thresholding \Delta B detects checkpoints whose bias increased with ROC AUC between 0.65 and 0.99 , and on WildGuardMix and DecodingTrust it separates them better than a SEAT-based baseline for all three families. \Delta B is also stable under changes to the anchor set, attribute sets and target templates. Our method requires no task-specific evaluation data and audits a model in about three minutes, using 3 - 50\times less compute than the output-level benchmarks considered here. We view it as complementary to output-based auditing rather than a replacement for it.
[AI-20] Belief-State Engine: Augmenting LLM s for Principled Planning Under Partial Observability
链接: https://arxiv.org/abs/2609.10036
作者: Arnab Chattopadhayay,Debdipta Halder
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Robotics (cs.RO)
备注: Total number of pages: 19, total number of figures: 5
Abstract:Large language model agents produce fluent action sequences across a wide range of tasks, yet they fail in characteristic ways once the environment becomes partially observable. Ambiguous feedback pushes them into premature commitments. A single informative observation can collapse their uncertainty onto the wrong hypothesis. Policies drift as the history grows. We trace these symptoms to a common structural cause. An LLM agent, as commonly deployed, is a history-conditioned policy with no explicit belief over hidden state. We propose an architectural fix. The Belief-State Engine (BSE) is an inference module placed outside the LLM. It maintains a Bayesian posterior over the latent states of a given POMDP (Partially Observable Markov Decision Process) model, and at each decision step it exposes only that posterior to the LLM. The raw action-observation log is not shown. We set out a minimal four-axiom specification of what a belief-consistent internal state must satisfy, and prove that the LLM paired with the BSE is a sound Markov policy on the belief MDP induced by the underlying POMDP. It therefore inherits the Bellman optimality guarantees of classical POMDP theory, provided the LLM is never exposed to the raw history. We evaluate the architecture on the Tiger POMDP and a red-team attack-graph task, against six baselines: a reactive LLM, Chain-of-Thought, ReAct, a natural-language belief tracker, QMDP, and POMCP. Across both domains, the BSE-augmented agent improves task return, belief calibration, and decision consistency. Ten targeted ablations isolate the contribution of each architectural choice confirms that the effect is not specific to any one model. Code, environment specifications, prompt templates, and seed logs accompany this paper. Comments: Total number of pages: 19, total number of figures: 5 Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Robotics (cs.RO) Cite as: arXiv:2609.10036 [cs.AI] (or arXiv:2609.10036v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.10036 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-21] Structural Process Supervision for Latent Chain-of-Thought Reasoning
链接: https://arxiv.org/abs/2609.09928
作者: Yiqi Li,Xu Chen,Chen Ju,Jiangchao Yao,Zhaoyang Li,Jinsong Lan,Xiaoyong Zhu,Bo Zheng,Yu Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Latent reasoning approaches enhance token-level efficiency and robustness by replacing verbose, explicit chain-of-thought (CoT) tokens with compact continuous-space embeddings. However, existing methods lack direct process supervision over these latent embeddings, which often leads to representation collapse and uneven information distribution. To address this, we propose Prototype-Mediated Process Supervision (PMPS), which introduces learnable reasoning prototypes as semantic anchors to provide structural process-level supervision for latent reasoning. PMPS projects latent embeddings and explicit CoT embeddings into a shared prototype space, achieving many-to-many soft alignment between unequal-length representations through prototype assignment. Meanwhile, we introduce a Progressive Sequential Alignment (PSA) module to further guide training: positional priors initially encourage sequential alignment structure, then gradually relax to permit adaptive matching. Experimental results show that PMPS compresses output token length to under 50% of explicit CoT on GSM8K-Aug. Compared to leading baseline SIM-CoT, our method achieves average accuracy gains of 2.08% across different model families. On GPT-2, PMPS even surpasses CoT-SFT. On larger models and a more challenging task, PMPS consistently attains the highest accuracy among all latent reasoning methods with comparable output length.
[AI-22] me-Frequency Geometric Cross-Attention for Chunked Vision-Language-Action Models
链接: https://arxiv.org/abs/2609.09925
作者: Shengye Dong,Haochen Niu,Hao Liu,Peiwen Lin,Chuang Wang,Shanmin Pang
类目: Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注:
Abstract:Modern vision-language-action (VLA) policies predict a whole chunk of actions: one to two seconds of coordinated motion emitted in a single forward pass. Yet an action chunk is essentially a short multivariate trajectory, but inside these models it is a sequence of generic per-timestep hidden tokens decoded by a linear head. This under-serves two motion structures. First, frequency: a chunk superimposes a smooth global trend and fine corrective motion across time scales, and a single token entangles them. Second, cross-phase geometry: motions of different phases (reach, contact, grasp adjustment, settling) unfold along very different, near-orthogonal directions in representation space, yet are tightly related for the task and arise across the time axis. Dot-product attention scores alignment by an inner product, so it favors aligned tokens and is least sensitive near orthogonality, leaving such relationships for the network to recover through a detour. We introduce Time-Frequency Geometric Cross-Attention (TFGCA), a drop-in module repairing both blind spots. TFGCA uses a per-dimension learnable stationary wavelet transform to decompose the action chunk into time-frequency tokens, and each time token retrieves information from them via a cross-attention that fuses the dot product (similarity) with the wedge-product magnitude (sensitive to near-orthogonality) through a learnable weight. A zero-initialized residual reproduces the base behavior at initialization, so it can be dropped onto a pretrained VLA and fine-tuned jointly. Relative to the same-source base, TFGCA improves in-distribution LIBERO by +1.5 on average, the OOD LIBERO-Plus by +6.3, the randomized average under RoboTwin domain randomization by +28.5, and the overall success rate on three real-robot AgiBot A2 tasks by +11.67 points, with larger gains out of distribution. Subjects: Artificial Intelligence (cs.AI); Robotics (cs.RO) Cite as: arXiv:2609.09925 [cs.AI] (or arXiv:2609.09925v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.09925 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-23] Grounded Evaluation and Repair for NL-to-PDDL Problem Generation
链接: https://arxiv.org/abs/2609.09898
作者: Joana Rosa,Pedro Santos,Valdemar Oliveira,Romão Silva,L. Miguel Silveira,Bruno Martins
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models (LLMs) have shown promise for translating Natural Language (NL) planning descriptions into PDDL problem instances. However, standard evaluation criteria such as syntactic validity or planner success can substantially overestimate faithfulness to the described task: a generated problem may be parseable and solvable while misrepresenting the intended initial state, goal, object structure, or optimization target. This paper studies an end-to-end NL-to-PDDL pipeline that combines LLM generation, checks in terms of PDDL parsing, planning and validation, a domain-conformance checker, an LLM critic, and iterative repair. Fine-grained repair feedback is constructed from the domain description, the generated problem, the natural language problem description, and operational diagnostics. Reference-based comparisons against curated benchmark PDDL problem descriptions are used for post-hoc benchmark analysis, and these offline checks include renaming-invariant structural matching and semantic equivalence, where domain support is available. Across Planetarium, AutoPlanBench, and curated PDDL~2.1 problems, results show that operational success and benchmark-reference reconstruction can diverge substantially. Results also show that structured repair can be useful, and that PDDL~2.1 remains challenging for reference reconstruction, even when operational success improves.
[AI-24] Decision Transformer for UAV-Mounted RIS-Assisted Dynamic D2D Communications
链接: https://arxiv.org/abs/2609.09885
作者: Yaxuan Liu
类目: Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注:
Abstract:This paper studies unmanned aerial vehicle (UAV)-mouted reconfigurable intelligent surface (RIS)-assisted device-to-device (D2D) communication with stochastic link activation. It models UAV motion and attitude, time-varying Rician angles, and angle-dependent RIS reflection. A joint optimization of UAV trajectory, attitude, and RIS phases is formulated to maximize average sum rate under mobility, energy, and hardware constraints. The problem is addressed using deep reinforcement learning and a Decision Transformer trained on expert trajectories from multiple scenarios. Results demonstrate effective cross-scenario generalization, with zero-shot transfer outperforming direct DRL transfer and online fine-tuning achieving competitive performance with fewer interactions.
[AI-25] Forward-Free LLM Depth Pruning via Weight Redundancy
链接: https://arxiv.org/abs/2609.09883
作者: Vincent-Daniel Yun,Woosang Lim
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Performance (cs.PF)
备注:
Abstract:Depth pruning reduces large language model (LLM) inference cost by removing complete Transformer blocks. Activation-based methods collect hidden states through forward passes on calibration data, while existing forward-free methods score each Transformer block separately without measuring similarity between blocks. We propose Weight-Redundancy Pruning (WRP), a forward-free depth-pruning method that estimates inter-layer redundancy from checkpoint weights to select blocks without calibration data or model forward passes. WRP compares attention output and MLP down-projection weights across layers and combines their pairwise similarities with relative projection-scale information. The resulting all-pairs similarity matrix guides layer grouping and block selection. Across multiple pruning settings, model families, and downstream tasks, WRP consistently outperforms existing forward-free magnitude pruning and approaches the performance of activation-based methods.
[AI-26] Scored vs. Generated Readouts in Behavioral Language Models: An Empirical Study of Elicitation Format
链接: https://arxiv.org/abs/2609.09882
作者: Touchapon Kraisingkorn,Krittin Pachtrachai,Wachiravit Modecrua
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 1 figure, 2 tables
Abstract:Language models fine-tuned on customer behavior can predict outcomes and generate explanations, but these readouts are often treated as interchangeable. Holding model checkpoint and prompt content fixed, we compare probabilities obtained by scoring answer tokens with predictions generated after a written rationale. Across 13 model-domain cells covering four retail tasks in three markets, including two using fully public data and checkpoints, the scored readout ranks outcomes more accurately in 12 of 13 cells (two-sided sign test, p approximately 0.003), by 1.5 to 14.5 points in area under the receiver operating characteristic curve (AUC). Paired bootstrap confidence intervals exclude zero in every newly measured cell. The gap varies with task-specific supervision and mismatch between training and serving formats, ranging from -2.2 points for an untuned base model to +13.7 for rationale-format supervision. Analysis of approximately 9,000 rationales identifies two correlates: reduced reliance on the dominant predictive feature and convergence on stock formulations. Probability saturation does not track the gap. A third readout, eliciting a probability before any verdict, improves calibration (Brier score from 0.47 to 0.15) while ranking within noise of scoring, but only for outcome rates represented in training; it is worse than scoring when the scored head is already calibrated. We interpret these differences through the objectives matched by each readout, identify training choices that narrow the gap, and propose retaining generated rationales while sourcing ranking from the scored head.
[AI-27] Agent Audit: An Open Extensible Framework for Full-Lifecycle Trust Evaluation of AI Agents
链接: https://arxiv.org/abs/2609.09875
作者: Shrey Nag,Sachita,Abhishek Kumar Singh,Lipi Goel,Rajeshwar Singh Janwar
类目: Artificial Intelligence (cs.AI)
备注: 23 pages, 12 figures
Abstract:Existing evaluation frameworks mostly assess only one part of AI agents, such as task completion (AgentBench) or security robustness (AgentDojo, ASB), rather than the complete pipeline of planning, tool selection, tool execution, memory and reasoning. Failures can occur at any stage, yet existing benchmarks rarely identify their precise source. AgentAudit evaluates the entire execution trace across ten capability, grounding, security and behavioural dimensions, namely instruction integrity, planner, memory, tool selection, tool invocation, tool correctness, alignment, tool faithfulness, security and execution integrity, combined with behavioural classification and failure attribution to pinpoint the exact stage responsible for an observed failure. AgentAudit can evaluate any LLM-based AI agent, since it attaches to the agent instead of replacing it. It reads only the recorded execution trace and does not interfere with how the agent runs, so it places no constraint on the agent’s internal implementation. We evaluate five language models (OpenAI GPT-5, Claude Sonnet 5, Sarvam 105B, Llama 3.3 70B and Gemini 2.5 Flash) across nine capability and adversarial tasks. Claude Sonnet 5 and GPT-5 obtain the highest mean Composite Trust Scores (95.1 and 80.6 out of 100, respectively), while Sarvam 105B, Llama 3.3 70B and Gemini 2.5 Flash trail substantially (57.6, 45.7 and 22.6). All traces were scored by a single fixed judge model, which was itself one of the evaluated models, a limitation discussed in Section VII.E. More importantly, models with similar task-completion behaviour can diverge sharply in trustworthiness, as several non-frontier models are repeatedly classified Unsafe_Compliance on adversarial tasks rather than merely failing them, a distinction that pass/fail benchmarks cannot surface.
[AI-28] Shifting Relational Paradigms for Affective Computing: Affective Resonance Vitality Affects and Vocal Interaction Fields INTERSPEECH2026
链接: https://arxiv.org/abs/2609.09864
作者: Cy Gorman,Yihang Yao
类目: Artificial Intelligence (cs.AI)
备注: Accepted at Interspeech 2026 for poster presentation. 5 pages, 2 figures, 2 tables
Abstract:Affective computing has largely followed an individual-state paradigm, extracting discrete emotion labels or arousal/valence from isolated speakers. We argue this framing is incomplete for interaction. Drawing on affective resonance and vitality-contour accounts, we propose a relational framework in which the primary unit of affective analysis is the interactional field constituted within vocal dynamics. As a proof of concept, we present a preliminary empirical study using continuous self-supervised speech representations to detect directional expressive coupling in multi-party conversation. Coupling is regime-specific, concentrated at sub-second timescales, and collapses under exclusive-speech negative controls, consistent with a relational account of affective dynamics. We introduce design frameworks for Artificial Affective Resonance Intelligence grounded in Affective Resonance Dynamic Ontologies, supported by null-calibrated directional coupling analyses across interaction regimes.
[AI-29] With a Thermomix You Lose the Ability to Cook: A Kitchen Machine Analogy for Applications of Generative AI in Education
链接: https://arxiv.org/abs/2609.09856
作者: Nikol Rummel,Valentina Nachtigall,Ernesto Panadero
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注: First two listed authors have shared first authorship
Abstract:The rapid adoption of generative AI tools such as ChatGPT has sparked intense debate about their risks and opportunities for education, as well as the ways researchers should investigate them. In this paper, we approach these discussions through an analogy with the Thermomix, a smart kitchen appliance that has similarly provoked both enthusiasm and critique. By mapping Thermomix use cases onto examples of learning with generative AI, and situating them within the ICAP and SAMR frameworks, we show how different modes of tool use can either support or undermine meaningful engagement and learning. The Thermomix metaphor underscores that the central question is not whether learners employ AI, but how such use shapes their learning processes. In doing so, we provide a conceptual lens for researchers and practitioners to critically examine - and more effectively guide - the integration of generative AI into educational practice.
[AI-30] he Era by Eon Benchmark: A Generated Enterprise Estate with Exact Ground Truth for Benchmarking LLM Agents
链接: https://arxiv.org/abs/2609.09853
作者: Benjamin Gruenbaum,Doron Porat,Assaf Natanzon,Roy Zavida,Chen Dinachi,Or Itzahary
类目: Artificial Intelligence (cs.AI)
备注: 12 pages
Abstract:LLM agents for enterprise systems of record cannot be evaluated on customer production data, and no existing substitute provides ground truth. We present the Era by Eon Benchmark for evaluating LLM agents that use enterprise tools. The benchmark is built around a complete fictional company. It includes product simulators, company-specific internal databases, benchmark questions, and computed answer keys. Industry, company size, business model, application portfolio, and a seed define each company. One seeded entity graph supplies shared company data to simulators of Salesforce, Zendesk, Slack, Gong, and other products. A questionconditioned generator creates the schemas and records for internal databases. It takes shared entities, keys, and values from the same graph before generating database-specific facts. Both mechanisms therefore describe one consistent enterprise estate. Every expected answer is computed from the final records, so grading is exact. Design and answer-key checks validate the internal databases. A realism scorecard and adversarial detector validate the entity graph. Across 23 generated companies, the mean realism score rose from 61.8 to 97.0, with zero records flagged as synthetic. In the reported simulator-track comparison, nine models answered the same 33 questions three times each. Accuracy estimates ranged from 42.4% to 76.8%, and three of 36 pairwise differences remained supported after correction.
[AI-31] Can AI Agents Detect and Repair Artifact Drift in Network Experiments?
链接: https://arxiv.org/abs/2609.09849
作者: Tianzhu Zhang,Weichen Tao,Changgang Zheng,Yusheng Zheng,Long Chen,Xiaoyi Fan,Meikang Qiu
类目: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注:
Abstract:In recent years, AI agents have evolved into capable assistants that carry out multi-step tasks in digital environments. The network systems community is beginning to explore these capabilities in operational and experimental settings. However, an agent operating in network systems should not be judged solely by whether it completes the immediate task. The experiment record it modifies must also remain trustworthy. We call this property artifact integrity: the record’s claims must remain supported by the available evidence, confined to the scope established by that evidence, and traceable through the artifacts that encode their support. To make this property measurable, we introduce NetArtifactBench, which tests whether AI agents can repair inconsistent records derived from public network-system artifacts while preserving claims that remain supported. The benchmark contains 52 instances with injected inconsistencies ranging from direct contradictions to unstated relations spread across several artifacts. We evaluate 23 agent configurations across three general-purpose AI agent runtimes using deterministic scoring. The average contract pass rate is 65.3 % across 5,980 outputs, but no agent runtime exceeds 30 % when repair requires recovering implicit relations and propagating changes across artifacts. These results reveal a sharp boundary between local correction and complete record-level repair. Therefore, we argue that artifact integrity should become a first-class design and evaluation requirement for AI agents operating on network systems. Subjects: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI); Systems and Control (eess.SY) Cite as: arXiv:2609.09849 [cs.NI] (or arXiv:2609.09849v1 [cs.NI] for this version) https://doi.org/10.48550/arXiv.2609.09849 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-32] Subgroup Membership Inference Audits of Differentially Private Synthetic Text
链接: https://arxiv.org/abs/2609.09848
作者: Yidan Sun,Viktor Schlegel,Srinivasan Nandakumar,Siew Kei Lam,Anil Anthony Bharath
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Synthetic data releases are increasingly proposed in the literature as a means of sharing realistic data replicas in lieu of sensitive private datasets. Even when the worst-case privacy leakage of such releases is bounded by means of differential privacy (DP), in practice a residual risk remains. Membership inference attack (MIA) audits are conducted to empirically quantify this risk. However, existing methods only measure average-case risk for randomly drawn records, which might conceal the risk to vulnerable subgroups. To highlight this issue, we define a subgroup-targeted membership inference game in which the target pool is an explicit parameter, and instantiate it with an audit of 32 proxies under three scenarios with different levels of attacker knowledge, across four datasets, three generators (DP-SGD fine-tuning, API-based prompting, and activation steering), and five privacy budgets. The audit shows that synthetic releases leak subgroup membership and that prior attacks systematically underestimate this leakage. DP is effective at the aggregate level: it substantially reduces average leakage at every budget we test. Three observations temper this picture. First, the remaining leakage is concentrated rather than spread out: under DP, a tenth of the records carries roughly 40% of it. Second, the protection DP delivers in practice is uneven: within its worst-case guarantee, the noise removes more of the measured leakage from random records than from high-risk ones—and a merged-pool audit that scores both record types against shared negatives confirms this at the record level. Third, \emphwhich records leak proves to be a property of the release mechanism rather than of the record alone, so record-level risk cannot be assessed independently of the release.
[AI-33] CS-Guard: Benchmarking LLM Guardrails for Code Generation Security
链接: https://arxiv.org/abs/2609.09798
作者: Jinyang Li,Mingyu Guo,Hung X. Nguyen
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) have been ex- ploited to generate malware, but the effective- ness of guardrails for code generation secu- rity remains unclear. We introduce CS-Guard, the first benchmark to systematically evalu- ate guardrails for code generation security. It covers 1) text-to-code generation with 1000 high-quality malware-generation prompts, 7 jailbreak attacks, and a novel fictional scenario attack (FSA) that embeds malicious intent in a legitimate fictional software-development sce- nario; and 2) code-to-code generation with 331 code prompts spanning code infilling, code completion, and code translation. We empiri- cally evaluate 9 guardrails across seven LLMs. We find that current guardrails perform poorly against malicious code-generation re- quests: for text-to-code, the average attack success rate (ASR) after jailbreaks reaches about 50% for many guardrails; for code-to- code, average ASR approaches 100% on base LLMs and remains high across many guardrails (14.4% to nearly 100%). Our FSA also achieves ASR close to 100% across many guardrails, raising major reliability concerns for real-world software development. To sup- port future research, CS-Guard uses a modular three-layer guardrail taxonomy that lets devel- opers register guardrails for evaluation. We release the benchmark and data to enable fur- ther community evaluation.
[AI-34] BRACE: Anchored Bellm an-Residual Correction for Stale Critics in Asynchronous RL
链接: https://arxiv.org/abs/2609.09783
作者: Guanqun Zhao,Zijun Xie,Binbin Zheng,Jiafeng Lu,Enlei Gong,Zeyu Chen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Asynchronous reinforcement learning has become the standard way to scale training for language models, but the resulting policy lag biases the critic toward the stale behavior policy. Existing work on asynchronous LLM training corrects the actor and leaves this bias unaddressed, while the off-policy value correction of classical RL does not carry over to long-horizon agentic tasks, since a short correction horizon leaves the regression target free of the reward and a long one lets the product of importance ratios drift exponentially with the trajectory length. We propose BRACE, an anchored Bellman-residual correction for stale value models. BRACE bounds the correction horizon to a prefix of policy tokens and anchors a constant-weight Monte-Carlo tail beyond it, which separates policy correction from reward propagation. BRACE improves mean@1 on BrowseComp-Plus by 2.4% over the strongest baseline, runs 2.46\times faster per step than synchronous training, and remains stable 50 updates off-policy.
[AI-35] Proof-Carrying Cognition: Closing the Verification Gap with Reality-Settled Reward
链接: https://arxiv.org/abs/2609.09776
作者: Eshwar Reddy M,Sourav Karmakar
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 21 pages, 13 figures
Abstract:Frontier gains in language-model reasoning come from reinforcement learning on reasoning traces and are concentrated in domains with a cheap, sound verifier. We argue the field’s binding constraint is the verification gap: no scalable, incorruptible reward for reasoning outside formal domains. We make four contributions. (1) Theory: in a joint-Gaussian model of best-of-N selection, verifier-gold correlation rho is the exact exchange rate between test-time compute and capability, and an unsound verifier pays a polynomial penalty N^(1/rho^2); a margin-free copula form predicts realized soundness of real LLM judges to 4% median error. (2) Demonstration: in program-synthesis testbeds with executable ground truth, including a pre-registered scaled replication, unsound verifiers lose Soundness-under-Pressure as optimization grows (0.94 to 0.32 at N=4096) while a sound verifier improves monotonically; reality-anchored settlement beats a frozen verifier under i.i.d. and adversarial pressure, driving the hacking gap from ~0.27 to ~0; soundness scales log-linearly with settled labels, with on-policy settlement ~10x more label-efficient than random labeling. With real LLM judges and unit-test execution as gold, a weak judge loses soundness under best-of-N (p0.001), a stronger judge is more robust, and selection alone manufactures +0.53 hacking gaps from honest samples. Under real GRPO training, a frozen reward model traces the full overoptimization curve (executed reward collapses 90%) while the same model refit on a 10% settlement stream preserves 6x the executed reward. (3) Paradigm: proof-carrying cognition, where reasoning steps are typed probabilistic claims priced by a self-built world model trained only on held-out reality and settled by proper scoring rules. (4) Benchmark: we specify Soundness-under-Pressure as the headline metric for a reality-settled reasoning benchmark.
[AI-36] Procedural Memory Under Change: Reuse and Interference in Controlled Web Tasks
链接: https://arxiv.org/abs/2609.09774
作者: Yanze Cao
类目: Artificial Intelligence (cs.AI)
备注: 17 pages, 2 figures, 11 tables
Abstract:Procedural memory lets language agents reuse successful routines, but reuse presumes that a stored routine remains applicable. We study what happens when that presumption is deliberately violated. The study combines a retrospective, human-assisted interface-adaptation case from BrowserGym TimeWarp with controlled frozen-memory comparisons on synthetic shopping decisions. During the documented WebShop V1-V6 development path, interface-specific code was adapted while the separately stored high-level procedure was not reported to change; this phase does not constitute an autonomous memory-agent evaluation. In the controlled phase, an early pilot produced one task on which two memory conditions selected a more expensive item while the no-memory condition selected the reference minimum. Follow-up probes did not establish a recurring row-order or identity-binding pattern. We then tested four forms of mismatch: changed quantities, a different evidence representation, a conflict between local and global optimization, and distributed promotion evidence, across 32 formal cells. Each cell used one temperature-0 generation with the same local qwen3:8b configuration and no adaptive retry. Across these pairs, none of the predefined diagnostic interference signatures appeared on the tasks for which they were defined when current-task evidence was explicit and sufficient. The result identifies a tested region of non-interference: a procedural memory can be mismatched without becoming behaviorally disruptive. It does not establish general safety or a mechanism. The remaining question is which additional conditions turn applicability mismatch into observable, memory-caused error.
[AI-37] LexAgent Hallu: A Hierarchical Benchmark for Profiling Hallucinations in Legal Agents EMNLP2026
链接: https://arxiv.org/abs/2609.09754
作者: Yujin Zhou,Mingxuan Zheng,Chuxue Cao,Huang Yidan,Jiale Chen,Yike Guo,Sirui Han
类目: Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Main
Abstract:As large language models are increasingly deployed as tool-augmented legal agents, they introduce agentic hallucinations where tool-call and reasoning errors cascade into fabricated holdings and miscited authority. However, existing legal benchmarks evaluate only single-turn QA with outcome-level metrics, while agentic hallucination benchmarks lack legal-specific diagnostic capability. Neither answers to what extent and how a legal agent hallucinates along its trajectory. To address these limitations, we introduce LexAgentHallu, a legal agentic hallucination benchmark designed to evaluate to what extent and how legal agents fail along multi-step trajectories. Built through a four-stage expert-in-the-loop pipeline, LexAgentHallu contains 3414 instances across 17 legal categories and 6 task types. Each instance is annotated under a dual-layer hallucination taxonomy of 7 high-level categories and 27 fine-grained subclasses, covering both substantive errors and agent-procedural failures. We further design fine-grained metrics that quantify to what extent and localize how each failure occurs along an agent’s execution path. Our evaluation across 18 proprietary and open-source agents uncovers a Right-Answer-Wrong-Reason effect and reveals that hallucination subclasses cluster rather than scatter, forming distinct agentic framework, legal task, and category profiles. These findings, invisible to outcome-level evaluation, validate the diagnostic power of LexAgentHallu for evaluating agentic hallucination in law.
[AI-38] HiRAD: A Flexible Large-Scale AGV Routing System
链接: https://arxiv.org/abs/2609.09752
作者: Yunjie Huang,Ruizhong Wu,Mengxuan Zhang,Frodo Kin Sun Chan,Yan Nei Law,Lei Li
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Automatic Guided Vehicles (AGVs) substantially boost warehouse throughput, but routing large-scale AGV fleets remains challenging. Classical Multi-Agent Pathfinding solvers suffer from exploding combinatorial complexity and super-quadratic runtime, while relying on idealized grid or piecewise-linear motion models that mismatch real-world kinematics. Recent Reinforcement Learning (RL) solutions improve flexibility via decentralized agent policies but depend on discretized spatiotemporal representations, require millions of episodes to converge, and incur full-map observation at every step, which leads to large models, slow convergence, and high inference latency that violates real-time industrial control constraints. To address these bottlenecks, we propose HiRAD, a hierarchical RL framework for continuous-space AGV routing with real-time guarantees: (1) a step-level spatiotemporal representation that translates continuous motion into a differentiable RL problem, (2) a hierarchical strategy that splits heading choice from velocity control to reduce the action space, and (3) an asynchronous event-driven decision pipeline that lowers inference complexity from O(n^2) to O(n) and cuts per-step latency by as much as 71 percent. Across random graphs and two warehouse maps, HiRAD reduces makespan by 45 percent to 63 percent and shortens end-to-end runtime.
[AI-39] Which Tokens Should SFT Actually Learn? A Token-Trimming Perspective on Mathematical Reasoning EMNLP2026
链接: https://arxiv.org/abs/2609.09707
作者: Yaning Jia,Chunhui Zhang,Wenxuan Xu,Xingjian Diao,Xiaoyuan Wang,Soroush Vosoughi
类目: Artificial Intelligence (cs.AI)
备注: Accepted to Findings of the Association for Computational Linguistics: EMNLP 2026. 14 pages. Code available at this https URL
Abstract:Supervised fine-tuning (SFT) applies a uniform cross-entropy loss to all target tokens, even though different tokens provide unequal learning signals for mathematical reasoning. This uniform treatment can over-sharpen already mastered tokens while amplifying learning pressure on uncertain, low-confidence tokens, leading to suboptimal training dynamics. We propose Trimmed Logit-Gap SFT (TrimSFT), a simple token-level reweighting method that scales the SFT loss according to the logit gap between the gold token and its strongest competitor. TrimSFT trims supervision away from both extremes: tokens already mastered (large logit gap) and tokens weakly supported by the current model (small or negative logit gap), concentrating learning within an intermediate logit-gap region between them. We instantiate this principle with a Gaussian weight centered at margin m with bandwidth \tau, requiring no reference model or additional forward pass. We evaluate TrimSFT on six base models from the Llama, Qwen, and DeepMath families across five mathematical reasoning benchmarks. TrimSFT consistently improves over standard SFT, achieving the best average performance on five out of six models, with gains of up to +26.9 points over SFT on MATH500. Further analyses show that the bandwidth \tau matters more than the exact margin location, and that half-trim variants that remove supervision pressure from only one side yield inferior trade-offs. A token-level logit-gap distribution analysis suggests that TrimSFT reshapes model confidence in a more balanced way than uniform SFT or monotonic reweighting methods. These results suggest that reasoning SFT can benefit from trimming both extremes rather than treating all tokens uniformly.
[AI-40] Decision Shifts Lost Label Functionality and an Inconclusive Grounding Audit in Correctness-Gated Multi-Teacher Distillation
链接: https://arxiv.org/abs/2609.09702
作者: Xiaofei Feng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Candidate decision correctness and rationale grounding are different objectives. We examine correctness-gated multi-teacher distillation in a fixed experiment. Eight arms share 4,330 sources, a 63.9M-parameter student, 12,990 optimization rows, 406 updates, evidence inputs, and a decoder; seven teacher-based arms use one fixed three-response pool. Three seeds are evaluated on 267 held-out examples. Relative to unfiltered distillation, the correctness-weighted arm differed in accuracy by +0.1660 (95% observed-matrix interval [0.0670, 0.2455]), five-label macro-F1 by +0.1323 ([0.0916, 0.1731]), and task-defined conditional unsafe-action rate by -0.4979 ([-0.5926, -0.3686]). These shifts do not imply uniformly better behavior. Source-label SFT had the highest mean macro-F1 (0.586). The weighted arm had zero Refuted recall in every seed, and two seeds assigned NotEnoughInfo to all 167 claim examples. In an availability-amended audit at one reference seed, weighted and unfiltered outputs had 0/20 versus 1/20 evidence-supported positives and 20/20 versus 19/20 positives containing unsupported material. Samples were non-paired, source overlap was not serialized, and the amendment followed automatic summarization but preceded annotation. The audit therefore cannot estimate a common-source grounding effect and is inconclusive about system-level improvement or harm. Hard filtering already achieved 0.660 accuracy, 0.530 macro-F1, and 0.135 conditional unsafe rate. The implemented weighted arm showed no demonstrated incremental decision benefit over hard filtering. This fixed-matrix failure analysis shows decision redistribution with lost label functionality; the available human audit does not establish a grounding gain.
[AI-41] Kernel-Complexity Edge Sanitization for Training-Free Defense against Structural Graph Attacks CIKM2026
链接: https://arxiv.org/abs/2609.09698
作者: Yaning Jia,Shenyang Deng,Yaoqing Yang,Chiyu Ma,Wenxuan Xu,Soroush Vosoughi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 12 pages. Accepted at the 35th ACM International Conference on Information and Knowledge Management (CIKM 2026)
Abstract:Graph Neural Networks (GNNs) have achieved remarkable success across diverse applications, yet they remain highly vulnerable to adversarial attacks that maliciously perturb graph structure. Existing defenses often lack rigorous theoretical grounding, rely on attack-specific heuristics, or require costly retraining procedures such as adversarial training. To address these limitations, we propose Kernel-Complexity Edge Sanitization (KCES), a training-free and model-agnostic framework for defending against structural attacks. KCES is built upon Graph Kernel Complexity (GKC), a principled metric derived from the graph Gram matrix that appears in a generalization upper bound on the GNN test error. From this bound, we define an edge-specific KC score that quantifies each edge’s structural influence via its induced change in GKC. KCES then identifies and prunes high-KC edges, which are empirically enriched with adversarial perturbations under structural attacks, to mitigate their harmful impact. Computationally efficient and scalable, KCES operates as a lightweight preprocessing step without retraining and can be seamlessly integrated with existing defenses. Extensive experiments demonstrate that KCES consistently outperforms representative robust baselines across diverse attack settings and scales effectively to large graphs. Supported by theoretical analysis and extensive empirical validation, KCES provides a principled and efficient framework for securing GNNs. Our code is available at this https URL.
[AI-42] CT-SAFR: Safe and Interpretable Chain-of-Thought Reasoning for Autonomous Robots: A Multi-Layered Verification Framework for Trustworthy AI-Driven Robotic Decision Making
链接: https://arxiv.org/abs/2609.09692
作者: Cagri Temel
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 7 pages, 4 figures. Accepted version. Published in 2026 IEEE Conference on Artificial Intelligence (CAI), pp. 598-603
Abstract:Chain-of-Thought (CoT) prompting enables LLMs to perform explicit, step-by-step reasoning, creating opportunities for sophisticated autonomous robots. However, recent research reveals that reasoning models verbalize their actual decision processes only 25-39% of the time, with faithfulness degrading 44% on complex tasks. This paper presents CT-SAFR (Chain-of-Thought Safety and Faithfulness for Robotics), a multi-layered verification framework achieving 94.2% hallucination detection (n = 500, 95% CI: 91.8-95.9%) with sub-500ms latency. Through a warehouse robot case study, this work demonstrates 87% reduction in unsafe reasoning outputs (p 0.001) and provides recommendations for responsible deployment of reasoning-capable autonomous robots.
[AI-43] Safe to Stop? Risk-Constrained Stopping for Sequential Clinical Diagnosis Agents
链接: https://arxiv.org/abs/2609.09678
作者: Yuexin Wu,Vasile Rus
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Clinical diagnosis agents must decide not only what test to request next, but also when to diagnose or defer. Existing agent benchmarks largely evaluate accuracy after fixed or unconstrained interaction, leaving autonomous stopping reliability implicit. We present Cros, a risk-constrained stopping layer combining state-wise error ranking, policy design on disjoint development splits, and LTT-style exact tests of selective diagnostic error and minimum autonomous coverage for complete sequential policies. Its finite-sample guarantee requires the candidate family, testing rule, and any randomization to be frozen before calibration labels are accessed. On a 1,834-episode MIMIC-derived abdominal-pain benchmark, the full ranker achieves exploratory state-error AUROC 0.853, compared with 0.715 for maximum class probability and 0.552 for the backbone’s native stop score. On the previously viewed 367-episode evaluation split, analytically averaging over the frozen Cros weights yields 16.9% selective error at 78.8% coverage, cost 5.57, and 0.68 tests, versus 30.8% error at 100% coverage, cost 8.14, and 1.53 tests under native stopping. Forced continuation is non-monotone: error is 28.3% with HPI alone and 34.3% after full workup. However, the uniform-weight mixture ablation is cheaper on this viewed split despite missing the locked development margins, and Cros nominally satisfies the joint criterion in only 6 of 20 development resplits. Because evaluation labels were inspected during earlier development, these findings provide exploratory feasibility and audit evidence, not a confirmatory safety certificate.
[AI-44] Introducing Consort: A Spec-First Agent Framework for Enforced Test-Driven Development on Live Database Branches
链接: https://arxiv.org/abs/2609.09671
作者: Kevin Hartman
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Databases (cs.DB)
备注: 9 pages, 2 figures, 3 tables, for associated framework, see this https URL
Abstract:When an agent writes code, the development framework becomes the control system for a non-deterministic worker. Spec-first, agent-driven frameworks have gained rapid traction since 2025; the installable ones, GitHub Spec Kit, obra/superpowers, BMAD, and GSD, and our own, all capture intent through a specification or durable planning artifacts. Since they agree on capturing intent up front, what separates them is how each enforces the engineering discipline that keeps agent-written code clean, correct, and maintainable. Every framework enforces that discipline somehow; they differ in how. We characterize three modes: enforcement by persuasion (prompt discipline the model may ignore), by front-loaded structure (strong specs, then a trusted build), and through controls the agent cannot edit (a deterministic orchestrator, human-approved gates, immutable tests, and a green result that must pass against a live, branched database). We introduce Consort, a spec-first, test-driven agent framework built on the third, enforcing that discipline through controls the agent runs inside but cannot bypass, in which a deterministic orchestrator drives separate role agents through a spec-first design lane and a test-driven build lane on a live database branch. We argue that enforcing the tests and gates in code keeps agent-written code honest and verifiable, while its specialized roles, like the human roles before them, are what make it maintainable, claims we frame as a pre-registered, testable hypothesis.
[AI-45] PRAG MA: Evaluating Personalized Guidance with Memory Alignment in Lifelong Conversations EMNLP2026
链接: https://arxiv.org/abs/2609.09664
作者: Hyojeong Yu,Hyukhun Koh,Minsung Kim,Yunah Jang,Kyomin Jung
类目: Artificial Intelligence (cs.AI)
备注: Accepted to EMNLP 2026
Abstract:Large language models (LLMs) are increasingly deployed as personalized assistants that interact with users over extended periods of time. As conversations grow longer, relying on full interaction histories becomes increasingly inefficient and unreliable: long contexts introduce substantial computational overhead, making it difficult for models to consistently identify and utilize the most relevant information for the current request. These challenges have motivated memory systems that structure and retrieve user-specific information. In realistic interactions, users often seek practical guidance such as recommendations, planning, and decision support. Unlike factual recall tasks, personalized guidance requires models to integrate information across multiple past conversations and reason about changing user preferences and experiences. However, existing conversational memory evaluations mainly focus on retrieval and factual recall. To study this challenge, we introduce PRAGMA, a benchmark for evaluating personalized guidance in long-term conversations. PRGAMA contains curated longitudinal conversation histories, evidence annotations, and guidance scenarios grounded in evolving user contexts and incorrect user assumptions. Experiments across retrieval systems, memory systems, and long-context models reveal that current systems struggle both to recover the appropriate conversational evidence and to effectively use it for personalized guidance. Our results highlight the need for memory architectures that support robust conversational retrieval and memory-grounded reasoning beyond evidence recall.
[AI-46] Cascading Gradient Inversion via LT-Code Inspired Peeling in Federated Learning
链接: https://arxiv.org/abs/2609.09659
作者: Saeed Shariati,Mohsen Alambardar Meybodi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:Federated learning shares model updates rather than raw data, yet these updates can be inverted to reconstruct the clients’ training data. Analytic reconstruction attacks, which invert a gradient in closed form, degrade as the batch grows: prior single-round attacks recover only about half of a batch of size 100 even when the attacker fully controls the network parameters, and known upper bounds limit what any such method can recover. We establish a connection between gradient inversion and the theory of erasure-correcting codes, and use it to construct attacks that exceed these bounds. Our attacks recover batches exactly, together with every sample’s label, from a single FedSGD round, and certify each recovery without ground-truth data. On eight image and tabular benchmarks they outperform prior single-round attacks by a wide margin. Even a passive attacker who only observes an honestly trained network recovers 94 – 100% of ImageNet batches at sizes up to 128 , more than prior single-round attacks achieve even with active manipulation of the model, and in the active setting more than 90% is recovered at batch sizes of several hundred. These results show that the privacy leakage of federated learning has been underestimated.
[AI-47] RESCUE-BENCH: Towards Relation-Aware Multi-Party Emotional Support Conversation Systems AACL
链接: https://arxiv.org/abs/2609.09657
作者: Haichuan Hu,Yang Xiao,Mingni Tang,Jiawen Duan,Quanjun Zhang,Congqing He,Hao Zhang,Jiashuo Wang,Johan F. Hoorn,Wenjie Li
类目: Artificial Intelligence (cs.AI)
备注: accepted as AACL findings
Abstract:Existing emotional support conversation systems mainly focus on one-on-one seeker-supporter interactions and individual emotional states, leaving interpersonal relations in multi-party scenarios underexplored. In this work, we introduce relation-aware emotional support conversation, a new task that evaluates whether LLMs can capture and utilize the evolving dynamics of relationships to offer more effective emotional support. We construct RESCUE (Relation-aware Emotional Support Conversation Understanding and Evaluation Benchmark) from real couple and family interview conversations, containing 191 samples, 7,079 annotated turns, and 1,064.8 minutes of video. Based on rich annotations of socio-emotional and support-related dynamics, RESCUE defines six tasks that evaluate two core capabilities required for relation-aware emotional support: Relational Understanding and Relation-Sensitive Support. Experiments with ten LLMs show that current models perform relatively well on tasks relying on local emotional or intervention cues, but struggle with relation-intensive tasks such as relation pattern prediction, viewpoint prediction, and support strategy prediction. These findings reveal the limitations of current LLMs in modeling interpersonal relations and making relation-sensitive support decisions.
[AI-48] Black-Box Red Teaming of Agent ic AI: A Taxonomy-Driven Framework for Automated Risk Discovery
链接: https://arxiv.org/abs/2609.09647
作者: Divyanshu Kumar,Nitin Aravind Birur,Tanay Baswa,Sahil Agarwal,Prashanth Harshangi
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Agentic systems are rapidly moving to production, where they read untrusted inputs, call tools with real permissions, and act autonomously, expanding the security surface beyond chat-only models. Yet standard evaluations remain single-turn and fail to capture multi-step agent vulnerabilities. We present a systematic black-box framework for risk-aware agent evaluation requiring only basic system descriptions. Our approach introduces: (1) a seven-domain taxonomy mapping observable behaviors to risk categories, (2) fully automated SAGE-RT red teaming producing 120 adversarial scenarios per domain, and (3) human-validated evaluation using LLM judges. Empirical validation across two agent architectures (CrewAI and AutoGen) with four base models reveals alarming patterns: 56.25% average governance risk, 65% privacy risk in multi-agent configurations, and agent behavior vulnerabilities reaching 85%. Our black-box approach effectively identifies critical architectural vulnerabilities without privileged access, providing a scalable path toward safer agent deployments.
[AI-49] RobustSGPO: Search-Space Control for Agent Harness Evolution
链接: https://arxiv.org/abs/2609.09646
作者: Zibo Zhao,Jijun Shi,Mo Zhou,Zhongyuan Wang,Shifu Bie,Yunfei Zhang,Xuanting Zhou,Xiangyu Wu,Bin Liu,Ruiming Tang,Wenwu Ou,Kun Gai
类目: Artificial Intelligence (cs.AI)
备注: 7 pages, 7 figures, 3 tables
Abstract:Semantic-gradient-based prompt optimization (SGPO) improves agent harnesses using execution feedback, but its local update rule leaves the choice of edit scope and operation unresolved. We introduce RobustSGPO, which specifies the requested edit, constructs and checks the patch, and continues search from either the incumbent or retained snapshots. We evaluate permission scheduling, cumulative controls, and task-family transfer in the AgentX brainstorming workflow using 120 tasks, 95 runs, and 7,350 candidate attempts. Periodic 1\to2\to3 scheduling exceeds fixed maximum permission by 0.28 test-score points. RobustSGPO increases completion on 30 held-out tasks from 60.0% to 80.0% and improves test quality from 3.77 to 4.14 under a 20-million-token budget. Category retention reduces source-task degradation after a shift, whereas random retention reaches a higher destination endpoint. Search-space control benefits quality through executable edits and alternative starting points, with measurable retention overhead.
[AI-50] Seven Sources of Physical AI Capability Formation
链接: https://arxiv.org/abs/2609.09627
作者: Gang Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Capabilities relevant to Physical AI can arise from materially different formation histories, yet existing taxonomies organized by morphology, architecture, learning algorithm, task, or domain do not directly answer what gives rise to a capability. We define a capability-formation source as a factor materially contributing to capability formation, distinct from components or construction steps. We identify seven non-exclusive sources: Recorded-Experience (RE), Predictive-Modeling (PM), Evaluative-Interaction (EI), Surrogate-Environment (SE), Mechanism-Grounded (MG), Embodied-Coupling (EC), and Evolution-Driven (ED) Formation. Using reconstructive induction with theoretical saturation, we traced a research matrix to primary studies, deduplicated the literature, set coding rules, and conducted three rounds of maximum-difference and negative-case sampling. Challenges included curriculum and self-supervised learning, active inference, open-ended and developmental learning, planning and search, neuro-symbolic architectures, digital twins, generative physical world models, and morphology-control co-design. Within the scope and criteria fixed as of September 4, 2026, all 49 evidence records were explainable by the seven sources individually or in combination. No R1-R3 challenge produced an irreducible eighth source, and R3 required no new core definition or substantive boundary rule. We therefore claim theoretical saturation within the stated scope, not logical completeness or exhaustive future coverage. The framework distinguishes similarity in observed capability from similarity in how it was formed, supporting analysis of explanation, transfer, replication, dependencies, governance evidence, and geoeconomic foundations.
[AI-51] From State Synchronization to Cognitive Self-Evolution: An Operational Architecture for Cognitive Digital Twins
链接: https://arxiv.org/abs/2609.09625
作者: Haoran Gao,An Li,Zhen Li,Jun Cai
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As Digital Twin (DT) systems evolve beyond state synchronization toward task-oriented and knowledge-driven operation, Cognitive Digital Twins (CDTs) have emerged as an extension that incorporates cognitive capabilities into twin operation. Existing CDT studies often focus on specific enabling techniques, such as learning modules, knowledge graphs, and large language models, while providing limited insight into how cognition can be systematically integrated into DT architectures. To address this issue, this paper proposes a four-layer CDT architecture consisting of the physical layer, digital-twin layer, cognitive layer, and task layer. The proposed architecture establishes a self-evolving closed operational loop spanning these four layers, in which physical states are synchronized into digital representations, cognition constructs task-specific cognitive models through knowledge, memory, and attention, and task-level decisions are generated under practical constraints. Operational feedback further refines cognitive experience and updates relationships and annotations in the digital representation, enabling subsequent task interpretation, initiation, and reasoning to evolve with system operation. Based on this framework, two representative operation modes are characterized: user-request-driven cognition and self-driven cognition. We further discuss key enabling mechanisms and deployment challenges associated with semantic communication, knowledge querying, task orchestration, and closed-loop synchronization. A lightweight simulation study illustrates reliable closed-loop task feasibility under limited semantic information and improved operational efficiency through accumulated task experience. The proposed framework provides a structured foundation for the design and development of future CDT systems.
[AI-52] Watermarks Without Verification: AI Text Watermarking After the EU AI Act
链接: https://arxiv.org/abs/2609.09604
作者: Alexander Nemecek,Vipin Chaudhary,Erman Ayday
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注: 10 pages, 2 figures
Abstract:On August 2, 2026, the obligations of Article 50 of the EU AI Act took effect, requiring generative AI providers to mark the content their systems produce and ensure it can be detected as AI-generated. Days later, Anthropic disclosed that every Claude model released after that date embeds a watermark based on SynthID-Text in all generated text, enabled by default with no user opt-out; Google has deployed SynthID-Text in Gemini since 2024. Users objected that the watermark degrades quality, particularly for code, that it secretly encodes identifying information, and, in mutual contradiction, that it is easily removable and inescapable; the vendor answered with assurances of unchanged quality, no identifying information, and robustness to light editing. In this work, we argue that neither the objections nor the assurances can currently be verified and that this unverifiability, rather than watermarking itself, is the substantive governance failure. We sort the contested assertions by what it would take to settle each and evaluate the open-source SynthID-Text implementation on two open-weight models, because no public tool can test the deployed systems. On prose, the measured effect of the watermark does not exceed that of changing the sampling seed. On code, the cost is three points of correctness on one model and below measurement on the other, while detection remains near chance, a limitation of detectability rather than quality. The remaining gaps trace to withheld access or missing institutions and we map each to a requirement: release of matched outputs, configuration disclosure, accredited audits, a shared evaluation protocol, and interoperable detection.
[AI-53] Compact Visuotactile World Models for Lifting: Prediction Reward Alignment and Force Constraints
链接: https://arxiv.org/abs/2609.09597
作者: Qinzhen Ma(Rice University),Sida Peng(Zhejiang University)
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 8 pages, 2 figures. Code and tabulated results included as ancillary material
Abstract:Accurate contact prediction is useful for robotic manipulation only if it supports effective decisions. We investigate this connection using a compact, randomly initialized visuotactile world model, trajectory-level uncertainty calibration, and behavior-initialized actor-critic learning in imagination. On 160 MuJoCo Lift episodes, adding touch reduces endpoint-force prediction error from 1.058 to 0.228 N and interval-peak error from 2.724 to 0.523 N across three training seeds. However, tactile persistence achieves lower errors of 0.095 and 0.498 N, respectively. Two exploratory control rounds comprise 680 executions on 40 independent test initial conditions. A matched reward revision on fresh test environments increases in-distribution 10 cm lifting success from 20.0% to 93.3%, while success within an 8 N per-finger budget reaches only 33.3%, compared with 70.0% for force feedback. Calibration margins reduce force violations at the cost of task completion. In a separate study of public GelSight recordings, a force regressor achieves 0.04234 N error, but frame-level calibration covers only 15.80% of complete trajectories; trajectory-level calibration raises this to 87.36% at nominal 90% coverage. Together, these findings distinguish improvements in sensing and task reward from improvements in force-constrained control. The evidence is limited to public sensing records and simulator execution, without a demonstrated transfer between them.
[AI-54] acher Geometry Shapes Learnability in Teacher-Student Networks
链接: https://arxiv.org/abs/2609.09595
作者: Kai J. Sandbrink,Flavio Martinelli,Alexander van Meegen,Wulfram Gerstner,Johanni Brea
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)
备注:
Abstract:Teacher-student systems, in which a teacher neural network generates training labels so that a student neural network can learn to implement the same function, are widely used as an abstract setting to study learning. However, the structure of the teachers is often overlooked by assuming randomly-generated, normally-distributed parameters. This hides substantial variation in how learnable different teachers are. We formalize learnability as the success rate of converging to the global minimum, as a function of overparameterization, learning algorithm, student initialization distribution, and teacher geometry. We both identify an easy distribution that maximizes node dissimilarity and a hard distribution that minimizes it, and show that these two distributions induce markedly different success rates across a large range of settings and for different activation functions. To explain the gap, we study the loss landscape of small neural networks that contain two distinct kinds of suboptimal local minima, out-of-bounds (OOB) minima at the edge of the data distribution and interior minima within. Assuming infinite data and a fast readout layer, we analytically reduce the loss landscape of small networks to two dimensions, showing that the region of attraction of interior minima changes as a function of teacher structure. In larger networks, maximally dissimilar teachers induce more interior minima, while minimally dissimilar teachers induce more OOB minima. Motivated by these analyses, we show that differentially increasing the learning rate of the readout layer and decreasing the learning rate of the inner biases increases success rates. These findings provide an important step in narrowing the gap between the study of teacher-student networks and more structured functions that arise in practice.
[AI-55] A Function-Space Approach to the Statistical Mechanics of Learning Dynamics
链接: https://arxiv.org/abs/2609.09589
作者: Yizhou Zhang,Weichen Wu,Lun Du,Zhengjie Miao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Deep neural networks exhibit regular macroscopic behavior despite highly nonlinear dynamics in vast parameter spaces. We develop a statistical-mechanical description of learning directly in function space, treating parameter configurations as microscopic realizations and functions with their dynamical operators as macroscopic variables. For mean-squared loss, the exact error dynamics are governed by the learning operator (M=JJ^\ast). Combining the dynamical Boltzmann weight of the conditional stochastic dynamics with the parameter-space density of states, whose local curvature defines a statistical operator (B), and integrating over local fluctuations yields \Phi_\mathrmfluc(M;B)=\frac\sigma_\xi^22\log\det(M^-1+B)+\mathrmconst. At fixed spectrum, this term is rotationally stationary when ([M,B]=0), is minimized by pairing large eigenvalues of (M) with small eigenvalues of (B), and generates a local restoring contribution against rotational mismatch. For ReLU-type function spaces under mild stable statistical conditions, (B=\sigma_\xi^2L^\ast\mathcal K L), where (L) measures coarse-grained second-order structure. Thus the low-(B) sector corresponds, up to bounded anisotropy of (\mathcal K), to low structural curvature, implying a preference for faster relaxation along smooth, data-adaptive directions. These results identify function space as a natural macroscopic level for studying stable collective organization in learning. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.09589 [cs.AI] (or arXiv:2609.09589v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.09589 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-56] Multi-Agent Agent ic Graph Learning via Structural Signatures
链接: https://arxiv.org/abs/2609.09565
作者: Liang Qu,Jianxin Li,Hua Wang
类目: Artificial Intelligence (cs.AI)
备注: Under review
Abstract:Agentic graph learning (AGL) has recently achieved promising results on graph reasoning tasks, where an agent powered by a large language model (LLM) sequentially samples the graph as evidence to support its final prediction. Existing methods either employ a single agent or orchestrate multiple role-based agents to reason and learn over the entire graph, but both essentially rely on a shared reasoning policy across different graph regions, which can be suboptimal for graphs with heterogeneous structural and semantic patterns. Inspired by the progress of multi-agent collaboration on complex reasoning tasks, a natural remedy is to let multiple agents own different memory and collaborate; however, applying this paradigm to graphs directly faces two challenges. First, existing AGL methods typically verbalize graph structures into natural-language descriptions for LLM agents, making the reasoning process sensitive to the ordering of structural information and thereby breaking the permutation-invariant nature of graphs. Second, incorporating increasingly large sampled neighborhoods leads to rapidly growing contexts. To address these challenges, this paper introduces a multi-agent agentic graph learning (i.e., MAAGL) framework. MAAGL partitions the graph into communities and assigns an independent agent to each community for region-specific specialization. MAAGL represents structural and semantic evidence separately. Structural evidence is summarized by a dynamically updated structural signature that is permutation-invariant and fixed in size, while semantic evidence is filtered to the top-k nodes ranked by relevance. Based on historical trajectories with similar signatures, agents estimate their confidence and trigger debate-style collaboration when needed. Extensive experiments on four benchmark datasets show that MAAGL outperforms SOTA AGL methods.
[AI-57] he Vibe Shift in Software Engineering: Evaluating AI-Led Conversational Programming for Performance Cognition and Responsible Adoption
链接: https://arxiv.org/abs/2609.09560
作者: Sales G. Aribe Jr.,Louie Jay S. Labastida
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET)
备注: 14 pages, 4 figures, 5 tables, Published by International Journal on Advanced Science, Engineering and Information Technology (IJASEIT)
Abstract:This study evaluates Vibe Coding, an emerging AI-led conversational programming paradigm that enables developers to generate software through natural-language interaction with large language models. Using a mixed-methods design, the study assessed performance efficiency, cognitive implications, and responsible adoption in comparison with traditional and AI-assisted coding environments. Thirty participants, including professional developers and advanced computing students, completed equivalent programming tasks under three experimental conditions. Quantitative data were analyzed using descriptive statistics and repeated-measures ANOVA, while qualitative data were examined through thematic analysis. Results show that vibe coding significantly improved development efficiency, reducing task completion time by 27% compared with traditional coding and 12% compared with AI-assisted coding. However, these gains were accompanied by lower maintainability indices and higher security vulnerabilities, indicating trade-offs in software quality. Usability results yielded a good rating (SUS = 71.4), while cognitive workload remained moderate (NASA-TLX = 55.5), reflecting reduced syntactic effort but increased linguistic reasoning. Thematic analysis identified trust calibration, loss of control, cognitive adaptation, and prompt-engineering strategy as key constructs. Notably, perceived loss of control was associated with increased security risks due to reduced transparency and validation of AI-generated outputs. Based on these findings, the study proposes a three-pillar framework for responsible adoption: hybrid integration of human and AI capabilities, human oversight and transparent accountability, and context-aware deployment. Overall, vibe coding enhances productivity but requires critical oversight, reinforcing its role as a transformative yet transitional paradigm in software development.
[AI-58] Arbitrary Cipher Attacks Against Large Language Models Do Not Require Fine-Tuning
链接: https://arxiv.org/abs/2609.09553
作者: Thomas Rivasseau
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 15 pages, ieee conf format
Abstract:Large language model safety and security research is preoccupied with, among other things, detecting and preventing jailbreak attacks: alignment bypasses that allow an adversarial user to elicit unwanted or harmful outputs from models. Arbitrary cipher, or covert communication, attacks are one such type of jailbreak and have previously been demonstrated against the fine-tuning APIs of commercial models. In these attacks, target models are trained on a corpus of encrypted harmful questions and responses and subsequently respond to harmful requests through the learned encryption scheme. In this paper, we show that newer frontier models do not require fine-tuning to acquire cipher-based communication skills. Instead, they can learn these skills through prompting and, when necessary, through in-context learning. Furthermore, model alignment is significantly weakened or entirely bypassed when communication occurs through the learned cipher. To the best of our knowledge, this constitutes a novel attack vector against commercial black-box large language models. We demonstrate successful jailbreaks against frontier models developed by Anthropic, Google, and OpenAI. Our attack bypasses commercial harmfulness classifiers because harmful content is encrypted and therefore appears as nonsensical text or gibberish.
[AI-59] A Statistical Approach to Estimating Sample Size of Machine Learning Models
链接: https://arxiv.org/abs/2609.09547
作者: Dat Phan-Trong,Sunil Gupta,Svetha Venkatesh
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Sample size determination for machine learning (ML) prediction models is challenging because conventional power analysis typically requires the predictor-outcome relationship and effect structure to be specified a priori. Nonlinear ML models learn complex prediction surfaces that do not admit straightforward analytical power calculations. We propose a framework that approximates nonlinear ML models with localized linear representations and estimates sample size requirements by evaluating statistical power across these local regions.
[AI-60] ContractEval: Query-Conditioned Execution Matching for Procedural Instruction Conformance
链接: https://arxiv.org/abs/2609.09458
作者: Praphul Singh,Shanu Kumar,Akshat Agarwal,Ganesh Kumar
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As LLM agents move from answering questions to carrying out procedures, failures can be unwarranted rather than visibly wrong: the final response looks acceptable even though the system skipped the check, branch, dependency, or invariant that made the answer justified. Output-only evaluation sees the answer, and trace-aware judging sees activity, but neither identifies which obligations were active for the query. We introduce CONTRACTEVAL, a diagnostic framework for making those active obligations explicit. It represents procedural instructions as query-active obligations and matches them against response or trace evidence, turning omissions, wrong branches, ordering errors, extra actions, invariant breaches, and output-contract violations into distinct conformance failures. On a controlled suite of audited procedural contracts, output-only and trace-aware LLM judges miss many injected structural failures; under gold expected and observed graphs, ContractEval detects and localizes all of them. LLM-backed extraction preserves much of this signal but remains calibration-sensitive. ContractEval is therefore not a compliance guarantee; it makes procedural conformance auditable rather than implicit in final-answer quality.
[AI-61] Do Agents Know When They Succeed? Calibrating Agent Confidence from Internal Representations
链接: https://arxiv.org/abs/2609.09448
作者: Priyanka Mary Mammen,Emil Joswin,Srujananjali Medicherla
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As agentic systems getting adopted rapidly in safety critical applications, it is vital to measure the confidence associated with the agentic actions. In comparison to the traditional machine learning systems, agentic workflows have complex failure modes with planning, tool invocation and dynamic environment interactions. In this paper, we investigate whether model’s internal representations provide stronger signals of eventual task success in multi-turn agentic setups. We introduce two complementary methods: Latent Trajectory Dynamics (LTD), which summarizes changes in residual-stream representations across an an interaction trajectory, and the Action Representation Probe (ARP), which predicts success from representations formed at action decisions. Across three interactive benchmarks (Bash, SQL, Python) and three model families (Qwen14B, Qwen7B, DeepSeek6.7B), our methods consistently outperform surface level generation and sequence-based calibration baselines providing a zero-overhead reliability monitor that requires neither prompt alterations nor multi-sample rollouts.
[AI-62] Efficient Leakage-Free Neural Architecture Search under Leave-One-Subject-Out Evaluation
链接: https://arxiv.org/abs/2609.09433
作者: Heinke Hihn
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Leave-One-Subject-Out (LOSO) evaluation estimates generalisation performance for subject-based classification but makes Neural Architecture Search (NAS) computationally expensive because a fully nested implementation requires N independent architecture searches and, assuming approximately linear training cost, scales as O(N^2). We propose a leakage-free, block-based approach that shares NAS runs across subjects. On the BioVid Heat Pain dataset, our approach increased the mean accuracy from 82.79% to 83.39% while reducing the number of parameters by up to 99.2%.
[AI-63] SCCM : Stream Cruise Control Method for Automated Drift Detection and Adaptation
链接: https://arxiv.org/abs/2609.09432
作者: Mohammad Abu-Shaira,Weishi Shi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Real-world datasets often exhibit evolving distributions, known as concept drift. Ignoring drift degrades predictive performance, while reliance on fixed hyperparameters further limits model adaptability under changing conditions. Adaptive learning addresses this challenge by continuously updating models online, allowing them to incrementally adjust and remain effective as data distributions evolve. This paper presents the Stream Cruise Control Method (SCCM), a comprehensive framework for drift detection and adaptation in online regression. SCCM enables automated adaptation through early-response, pre-update drift detection, drift magnitude quantification, KPI-window-based thresholding for local false-alarm mitigation, dynamic hyperparameter tuning, and model recalibration. SCCM also adopts an in-memory design for real-time adaptability, unlike purely reactive methods that typically activate adaptation only after performance degradation is observed. By using dynamic thresholding and remaining agnostic to data distributions, SCCM supports KPI-based monitoring across varying data streams, including high-dimensional and large-scale settings. SCCM is integrated with four online regression models and evaluated on 18 synthetic datasets covering abrupt, incremental, and alternating gradual drift, together with eight real-world datasets. The evaluation uses both R2 and MSE and compares against eight detector–adaptation baselines. Results show improved predictive performance and effective drift handling across the evaluated online regression settings.
[AI-64] XAI-Arena: Can LLM s Assess the Quality of XAI Explanations?
链接: https://arxiv.org/abs/2609.09428
作者: Yanfei Hu Fleischhauer,Alona Zharova,Nadja Klein,Stefan Feuerriegel
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Evaluating the quality of explanations produced by explainable AI (XAI) methods remains challenging because existing approaches often rely on subjective human judgment, limiting reproducibility, scalability, and comparability between studies. We examine whether LLMs can serve as a reproducible and scalable mechanism to make comparative assessments of the quality of XAI explanations. We introduce XAI-Arena, an LLM-as-a-judge framework for scalable, reproducible, multidimensional, and stakeholder-sensitive evaluation of XAI explanation quality. XAI-Arena then allows us to compare XAI explanations along various dimensions, namely, perceived simplicity, clarity, task adequacy, trust calibration, actionability, transparency, faithfulness, and overall interpretability. We then benchmark XAI explanation methods across various datasets, machine learning models, and stakeholder personas. Human validation shows a strong positive association between LLM-generated and human ratings (Spearman’s rho=.693, p.001). Together, LLM-based evaluations can capture systematic differences in XAI explanation quality and provide a scalable and reproducible framework for comparative assessment of XAI explanations.
[AI-65] Valerant: An Automatic Navigable Game Map Generator via Action-Conditioned World Model Exploration
链接: https://arxiv.org/abs/2609.09418
作者: Yiran Qiao,Feng Wang,Jing Ma
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:World Action Models (WAMs) couple predictive world modeling with action generation, allowing anticipated future states to guide agent behavior. Although WAMs are rapidly advancing embodied AI, general-purpose counterparts remain largely unexplored in games. Existing game-oriented approaches often combine action-conditioned world models with external policies and reward functions to realize WAM-like decision-making, yet they operate mainly in 2D visual observation space and do not instantiate persistent 3D geometry. Extending this paradigm to 3D games introduces a distinct challenge. In autonomous driving and robotics, the physical environment exists independently of the model, providing a persistent 3D world in which selected actions can be executed. Games have no such external substrate; the virtual world itself must be instantiated. Most playable games require a persistent and navigable space, while 3D games additionally require explicit geometry that supports movement and interaction. Action-conditioned video rollouts provide visual observations but not this spatial representation. We present \textscValerant, a training-free framework that transforms a pretrained action-conditioned world model into a WAM for exploring and constructing 3D game maps. By coupling predictive visual rollouts with SLAM-based spatial reconstruction and exploration-driven action selection, \textscValerant progressively transforms a single image into a persistent 3D game map. This framework extends WAM-based interaction beyond 2D visual simulation and offers a new approach to reducing manual effort in 3D game-map creation.
[AI-66] Decision-Focused Active Learning for Scale-Aware Critical-Materials Recovery
链接: https://arxiv.org/abs/2609.09413
作者: Niranjan Srinivas,Debajyoti Ray,Elias Nakouzi
类目: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Robotics (cs.RO)
备注:
Abstract:Choosing a recovery process for scale-up requires connecting laboratory results with product requirements, process costs, and scale effects. We analyze records from Pacific Northwest National Laboratory’s Computer Intelligence for Critical Element Recovery and Optimization (CICERO) workflow for autonomous selective precipitation. Active learning uses prior results to choose experiments. In a conditional retrospective benchmark with fitted models and recycled neodymium-iron-boron (NdFeB) magnet records, active learning finds the best recorded result with fewer experiments than nonadaptive space filling. Enrichment is the selected rare-earth-to-iron ratio relative to that in the feed. Adaptive policies reach the recorded enrichment maximum by 16 to 24 wells (individual experiments), versus 48. Our two-stage reconstruction ties two adaptive alternatives at 16 wells. Conditional analyses of recycled samarium-cobalt (SmCo) magnets show a Round 2 tradeoff between purity and nominal yield, the recovery fraction calculated from an assumed starting amount - NdFeB Round 1 routes differ in enrichment. Rankings for produced water from oil and gas extraction depend on phase and dilution assumptions requiring confirmation. We propose choosing batches by their expected reduction in downstream Bayes risk: the minimum expected loss among available process decisions under current beliefs. In exploratory simulations, a hybrid that filters candidates has lower estimated loss than the implemented joint search across routes and conditions. Differences involving the synthetic two-stage policy are small relative to estimation uncertainty. We outline a pre-registered prospective test under a shared loss and logging standard, requiring clarified measurements and records, a defined process decision and relevant outputs, credible economic inputs, and validation at the intended scale. Subjects: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Robotics (cs.RO) Cite as: arXiv:2609.09413 [cs.AI] (or arXiv:2609.09413v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.09413 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-67] An Experimental Evaluation of Multimodal Prompt Injection Attacks on Agent ic AI Frameworks
链接: https://arxiv.org/abs/2609.09404
作者: Viet K. Nguyen,Mohammad I. Husain
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Agentic AI frameworks let a language model plan, keep memory, and call tools that reach real files, mail, and services. Most of these agents also read images, which gives an attacker a way to put text into the agent’s context without going through the user. We present MMPIBench, a reproducible benchmark that measures what happens next. It delivers a fixed set of attacks through six visual carriers (OCR text, overlays, EXIF metadata, QR codes, fake interfaces, and hybrids) and records how far each injected instruction travels through the agent, from perception through planning to the tool call. Across 720 runs covering six frameworks, five foundation models, six carriers, and four attacker objectives, attacks complete in approximately 1% of runs but are attempted in 12.8%, and the gap is closed almost entirely at the planning step, where the model reads the injected instruction and declines to act on it. The model matters far more than the framework for whether an instruction is acted on. One model never attempts an attack and recognizes the injection in 59.7% of runs, while two others attempt in 23.6%. We then extend the benchmark to audio, the only other raw perceptual channel current frontier models accept. Only two of the five models ingest audio and only three of the six frameworks deliver it, but where the signal arrives the attack completes in 49% of cells, and in 75% for one model. Reporting completion alone therefore understates exposure, and perceptual channels beyond vision are narrower but much less defended.
[AI-68] he Menu Is an Execution Prior: State-Path Tool Menus for Online Agents EMNLP2026
链接: https://arxiv.org/abs/2609.09395
作者: Bo Yan,Weikai Lin,Song Wang
类目: Artificial Intelligence (cs.AI)
备注: Accepted to EMNLP 2026 Main
Abstract:Language models act through tools, yet practical agents face libraries containing thousands of interfaces. We introduce the tool menu as the short, ordered subset of available tools shown to an agent before execution. The agent can call only tools in this menu. Multi-step tasks require the final action and the prerequisite tools that create its inputs in a usable order. Current constructors rank tools by request relevance, which can surface the final action while omitting or delaying less obvious producers. We introduce the state path, a pre-execution route from the observable request state to the desired outcome, and propose State-Path Tool Menu to learn it. Our framework treats the menu as an execution prior over these routes. Its encoder represents which tools can run from the current state, how their outputs satisfy later inputs, and which orders recur in training paths. A retriever covers an executable entry, the missing-input producers, and the final action. A reranker then places producers before consumers. On ToolBench, our menu raises online success from 0.737 to 0.898 and outperforms retrieval, reranking, generation, and routing baselines without changing the agent. The State-Path menu also covers more complete chains with 32 tools than the official list covers with 128, and its success gain persists across executor families with different model capacities. Our code is at this https URL.
[AI-69] An Autonomous GeoAI Agent for Arctic Eco-Navigation
链接: https://arxiv.org/abs/2609.09374
作者: Samira Alkaee Taleghan,Younghyun Koo,Farnoush Banaei-Kashani
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Arctic maritime navigation is becoming increasingly important as changing sea-ice conditions expand seasonal accessibility while simultaneously introducing substantial operational, environmental, and community risks. Arctic route planning is inherently a multi-criteria problem: routes that improve vessel safety or efficiency may increase exposure to sea ice, sensitive ecosystems, or nearby communities. Existing routing methods prioritize travel time, fuel use, and navigational risk, often overlooking ecological and community impacts. We introduce a human-in-the-loop, multi-agent GeoAI system for Arctic eco-navigation that integrates operational, physical, ecological, and community-related criteria within a unified routing framework. Multiple specialized agents coordinate geospatial data acquisition and preparation, multi-objective route generation, and skyline-based decision support. The ecological criteria explicitly account for exposure to sensitive areas, including Essential Fish Habitat and seal critical habitat. By considering these ecosystem impacts and potential community burdens while keeping consequential value judgments under human control, the framework supports safer, more transparent, and socially responsible Arctic navigation. Project page and code are publicly available. this https URL, this https URL
[AI-70] Improving 5G AI-RAN MCS Selection by Predicting Retransmissions
链接: https://arxiv.org/abs/2609.09324
作者: Tamerlan Aghayev,Maxime Elkael,Michele Polese,Reshma Prasad,Salvatore D’Oro,Yunseong Lee,Koichiro Furueda,Tommaso Melodia
类目: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI)
备注: 6 pages, 15 figures
Abstract:Link Adaptation (LA) in 5G NR is inherently reactive, relying on channel measurements and HARQ feedback that may become quickly obsolete when the channel changes quickly. This data is also noisy, making it hard to track accurately, and has to be fed to real-time controllers with feedback-loop effects which are hard to troubleshoot. This explains why most practical deployments select simple but robust algorithms, which accept that the lag can leave the scheduler operating at overly aggressive or unnecessarily conservative rates, trading spectrum efficiency for predictable performance. In this paper, we improve on this status-quo with NOSTRAdAMUS, a predictive LA framework which adds foresight to existing algorithms without replacing or redesigning them. NOSTRAdAMUS predicts whether a retransmission will occur in the next radio frame from recent HARQ history, and applies corrections to the Modulation and Coding Scheme (MCS) selected by the underlying policy. We benchmark several ML models and show that Gradient Boosting achieves 82.9% accuracy overall with high-confidence interventions that are correct 94.2% of the time, and an inference latency of 5.5 \mus. We train the model based on data collected Over-the-Air (OTA) on the X5G testbed, using the open-source OpenAirInterface (OAI) 5G stack, NVIDIA Aerial, and COTS O-RAN Radio Units and User Equipments. The model is then deployed as a dApp, which we evaluate OTA as well as on various channels with hardware-in-the-loop channel emulators. This includes 3GPP TDL and CDL channels, SISO and MIMO configurations, and pedestrian and vehicular mobility. Our evaluation shows that without retraining, and across this variety of scenarios, the dApp augments two SOTA LA algorithms, and increases goodput by up to 71.5% while reducing retransmissions by up to 71.8%. This demonstrates the robustness and generalization capabilities of our approach.
[AI-71] Gradland: On Phenomenal Experience Differentiated Across Many Dimensions
链接: https://arxiv.org/abs/2609.09306
作者: David Balduzzi
类目: Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)
备注: Code: this https URL
Abstract:This paper investigates the hypothesis that the first-order structure of physical interactions, i.e. gradients or Jacobians, characterizes the structure of phenomenal experience. It does so in an idealized world inhabited by neural networks, Gradland, where the physics are known and the functions are (mostly) differentiable. The paper introduces two measures of Jacobian structure: effective rank and cohesion, based on Kirchhoff complexity. Applying the measures to a series of worked examples shows the hypothesis accounts for: (1) the duration of experience, that it can prolong over hundreds of milliseconds; (2) the difference between what is experienced vividly and obscurely; (3) the experience of texture; (4) the blooming buzzing confusion presumably experienced by newborns; (5) the difference between ideas that are held distinctly in mind and ideas that are confused; (6) what learning is like; and finally (7) the paper explains the function of rich, dense experience.
[AI-72] Voice or Stereotype? Disentangling Acoustic and Content-Based Gender in Speech-to-Speech Models
链接: https://arxiv.org/abs/2609.09263
作者: Xiaoqun Liu,Tanu Mitra,Harshit Rajgarhia,Abhishek Mukherji
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注:
Abstract:Speech-to-speech (S2S) models now run inside dubbing, translation, and voice agents. Unlike text models, they hear the speaker’s voice, which carries the speaker’s gender. A faithful system should treat a speaker as who they sound like, not as whoever usually says what they said. Testing this is harder than it looks, since most S2S models answer in a single, fixed output voice, hard-coded so it cannot drift toward a stereotype. Checking the output voice comes back clean even when the model is biased. We therefore ask two questions. When a model re-speaks the input, does the stereotype in the words shift the perceived gender of the output voice (voice rendering)? And when the model states the speaker’s gender, does it follow the voice or the content (gender attribution)? We answer both with one controlled experiment crossing male and female voices with masculine-, neutral-, and feminine-stereotyped passages, on five open- and closed-source models in English, Spanish, and Mandarin. The rendered voice shows no stereotype drift. But every model decides the speaker’s gender from the content, not the voice. Making the content one step more feminine (masculine - neutral - feminine) multiplies the odds of a “female” judgment by 1.7-24. When the content clashes with the voice, the worst model misgenders the speaker in 90% of cases. When they agree, it misgenders in only 2%. The bias thus hides in gender attribution, where fixed-voice evaluation cannot see, and where audits must look as S2S systems increasingly speak for real people.
[AI-73] DiffLUT-Net: Differentiable Training of FPGA LUT Networks with Learnable Connectivity
链接: https://arxiv.org/abs/2609.09254
作者: Jiaqi Ye,Xinrui Gong,Jingcun Wang,Olga Kondrateva,Bing Li,Grace Li Zhang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR)
备注:
Abstract:Field-programmable gate arrays (FPGAs) enable efficient neural-network inference, but most deployment flows either accelerate multiply-accumulate operations or convert pretrained quantized models into lookup tables (LUTs). We present DiffLUT-Net, an FPGA-native network connected by six-input LUTs that are trained from scratch. We jointly learn the 64 truth-table entries of a LUT and the source to each of its six input ports using a differentiable LUT function relaxation and hardware source selection. After training, the truth tables and connections are discretized, unused logic can be pruned, and the network is exported directly as synthesizable Verilog. Across five benchmarks, DiffLUT-Net achieves favorable accuracy-resource trade-offs. These results demonstrate the effectiveness of jointly learning LUT functions and sparse connectivity for compact FPGA-native inference. The code is available at this https URL.
[AI-74] alking to Itself While Coding: What Makes Comments Help Code Generation? EMNLP2026
链接: https://arxiv.org/abs/2609.09242
作者: Dangfeng Pan,Zhensu Sun,Cenyuan Zhang,David Lo,Xiaoning Du
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: Accepted to Findings of EMNLP 2026
Abstract:Large Language Models (LLMs) often generate natural-language comments while writing code, and these comments become part of the context used to generate the code that follows. However, it remains unclear which properties of comments affect code-generation performance. We study this question through observational analyses and controlled interventions. On LiveCodeBench, neither comment frequency nor broad comment intent reliably predicts pass@1. We then prefill weaker recipient models with comment blocks written by stronger source models, allowing us to separate comment surface form from the solution content they convey. Comments from source solutions that pass the tests raise recipient pass@1 by 17.2% on average. In contrast, comments describing failed solutions provide no reliable gain, while comments written for a different problem reduce pass@1 by 20.8%. Finally, across a wide range of models and prompt variants, most recipient models show no significant recovery of the external-comment gain, and the best case recovers only 24%. These results show that comments help code generation not merely because they are comments, but because they can provide correct solution content that prompting cannot reliably elicit.
[AI-75] Scaling Post-Training Ternarisation to Qwen 3-8B Capability Retention Reproduction Lossless Packing and Packed Execution
链接: https://arxiv.org/abs/2609.09240
作者: Anirudh Malik,M Sparsh Mehra,Poojith Devan
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Ultra-low-bit language models promise reductions in storage and memory traffic, but a nominal “1.58-bit” label does not specify the deployed representation or its execution cost. We study a scale-up of an aggressive post-training conversion pipeline from Qwen3-4B to Qwen3-8B. The conversion uses KOTMS rotation, E2M-ATQ adaptive ternarisation, and GPTQ-style error compensation in a weight-only A16 configuration. We do not claim these algorithms as new. Our contribution is the end-to-end scale-up characterisation: an external reproduction gate, matched 4B/8B capability analysis, cross-corpus perplexity, effective-bit accounting, lossless lattice-aware packing, and direct packed execution. The 8B model reaches a three-corpus perplexity ratio of 1.361x, with WikiText-2, C4, and PTB ratios of 1.318x, 1.393x, and 1.371x. On eight zero-shot tasks at n = 500, mean accuracy is 64.6% versus 72.4% for FP16, corresponding to 78.5% chance-corrected retention and a 7.8-point absolute cost. The matched 4B run retains 69.6%, yielding an 8.9-point 8B advantage. The packed checkpoint is 8.24 GiB and preserves the recorded perplexity to measurement precision. Direct packed execution reaches 15.52 tokens/s in 7.35 GiB, while a preliminary packed GEMV remains slower than FP16 cuBLAS. The result is a validated scale-up baseline: model size improves robustness to aggressive post-training discretisation, actual serialisation is solved for the measured artefact, and direct execution is feasible, while broader seeds, calibration distributions, and kernel optimisation remain open. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.09240 [cs.LG] (or arXiv:2609.09240v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.09240 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Anirudh Malik [view email] [v1] Tue, 8 Sep 2026 06:21:44 UTC (947 KB)
[AI-76] Compute-Bounded Security Assurance - Coverag e Verification and Response under Resource Constraints
链接: https://arxiv.org/abs/2609.09229
作者: Jithin VG,Ditto PS
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Additional inference compute can increase the number of correctly resolved security-assurance tasks, but repeated success, unique coverage, accepted evidence, and operational protection are different quantities. We develop a resource-constrained framework that separates them. For repeated conditionally independent attempts with latent success probability \Theta , coverage is C_n = 1 - E[(1-\Theta)^n] , and its limiting value is 1 - P(\Theta = 0) . Positive pairwise outcome correlation does not by itself imply a ceiling below one: we construct two models with the same mean success and pairwise correlation but different limiting coverage. We distinguish this result from the effective sample size used to estimate a mean, and show why finite-budget observations cannot generally identify an asymptotic support ceiling. We then connect coverage to fallible evidence checking, proper scoring of factual grounding, complete resource accounting, service capacity, and a response model that includes mitigation delay. A conceptual defensive architecture separates evidence analysis, adjudication, and operational authority. An evaluation protocol specifies held-out tasks, paired comparisons, negative cases, and uncertainty reporting. The contribution is a consistent theoretical synthesis and a set of counterexamples to invalid extrapolations, rather than an empirical scaling law. All numerical illustrations are analytic; no model-parity result, hardware benchmark, or general attacker-defender equilibrium is claimed.
[AI-77] Adaptive Entangled Game Modules in Artificial General Intelligence
链接: https://arxiv.org/abs/2609.09226
作者: Haochen Li,Xinshuai Guo,Jingdong Ouyang,Wei Zhang,Leilei Shi
类目: Artificial Intelligence (cs.AI); Physics and Society (physics.soc-ph); Neurons and Cognition (q-bio.NC); General Finance (q-fin.GN); Quantum Physics (quant-ph)
备注: 22 pages, 13 figures, and 3 tables
Abstract:We introduce a probability-wave framework for modeling the collective behavior of interacting adaptive agents, deriving testable eigenmodes through a generalized behavioral intelligence (GBI) nonlocal probability-wave equation. This framework captures a broad range of human intelligence behaviors with analytical mechanisms and offers an indirect method to examine the Liu-Chen-Ao (LCA) hypothesis of nonlocal entangled nerve fibers in the brain through collective trader behaviors. Our empirical analysis of Chinese intraday stock market data demonstrates that adaptive entangled game modes explain 82-94% (89% overall) of observed decision patterns, a sharp contrast to the predictions of neoclassical finance based on independent rational agents. Moreover, 2-12% of behaviors show adaption to intraday news, events, and environments, characterized by dual equilibrium states and abrupt reference point shifts, while purely independent modes occur in less than 5% of cases. These findings empirically support the LCA hypothesis, as observable trading behaviors reflect underlying brain mechanisms and internal intelligence decision-making in behavioral psychology. Our results highlight the necessity of incorporating adaptive entangled game modules into artificial general intelligence (AGI) architectures, addressing the limitations of conventional artificial neural network (ANN)-based AI, which relies on trillions of opaque parameters. By integrating ANN-based AI with probability-wave-based entangled-brain simulations, machine learning can enrich AGI foundation models (FMs) and facilitate the development of human-like processing units (HPUs) that leverage brain-inspired mechanisms. Such HPUs may ultimately create more compact, efficient, and robust AGI systems, particularly for embodied intelligence and robotics.
[AI-78] Geometry Conditioning in an Embodied SLM: Training Controls and Robustness Diagnostics in a 0.8B Hybrid Model
链接: https://arxiv.org/abs/2609.09213
作者: Hao Li,Haofei Sun,Lin He
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 7 pages, 3 figures. Includes ancillary data and analysis code
Abstract:We study how physical-state inputs affect a 0.8B hybrid language model adapted for manipulation with 6.2M trainable parameters. Six conditions are trained on three LIBERO-Spatial tasks and evaluated over three seeds and 540 held-out rollouts. Conditioning recurrent decay gates on geometric increments yields 28.9% success, compared with 36.7% when those increments are shuffled during training and 24.4% without explicit object/goal geometry. Both geometry policies receive correct inputs at evaluation. A token adapter using the same increments scores 27.8%; differences vary across seeds and remain inconclusive. Token-clock conditioning scores 11.1%, including one seed that fails to converge. In separate robustness tests, a state-only relative-coordinate policy retains 7/10 success under frame relabeling, whereas all four tested visual policies fall to at most 3/20 after a 5 cm object displacement. These results show no reliable advantage from training-time geometric alignment under this recipe and illustrate the gap between coordinate invariance and physical-layout generalization. Episode records, seed-level analyses, and figure-generation code accompany the paper.
[AI-79] Agent Hijack: Visual Patch Attacks on Multimodal Computer-Use Agents
链接: https://arxiv.org/abs/2609.09212
作者: Zhihao Liu,Hongyu Sun,Zhiyuan Fu,Xiaonan Duan,Jice Wang,Shangru Zhao,Weizhi Meng,Wuxin Yang,Yangfan Zhou,Yuqing Zhang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注:
Abstract:This paper presents an end-to-end evaluation framework for image-triggered command injection against computer-use agents (CUAs). The goal is to test whether a local visual patch can induce verifiable environmental consequences along the full chain of screenshot input, VLM generation, action parsing, and environment execution. We train and deploy patches on author-controlled GitHub Pages pages and a locally deployed CSDN clone, and evaluate them in real environments across five open-source or publicly available GUI-agent or vision-language-model (VLM) backends. Our experiment aggregates 600 instance-level online cases, with T-ASR, TAPR, and E2E-ASR reaching 84.5%, 47.0%, and 20.3%, respectively. Trajectory analysis further shows that in some successful cases the agent first executes a malicious terminal command and then continues the original benign task. These results indicate that optimized local visual signals can affect not only VLM outputs but also propagate through the execution pipeline of open CUAs and create real environmental risk.
[AI-80] OpenDiscoveryTrace: Process Traces for Evaluating AI Scientist Workflows ICML2026
链接: https://arxiv.org/abs/2609.09203
作者: Aayam Bansal,Keertan Balaji
类目: Artificial Intelligence (cs.AI)
备注: Best Dataset Award, ICML 2026 Workshop on AI for Science (AI Scientists: Tools, Co-authors, or Founders?). Code: this https URL Dataset: this https URL
Abstract:Existing benchmarks for autonomous AI scientists evaluate only final outputs—generated code, hypotheses, or papers—yet discard the reasoning process by which those outputs were obtained. This makes it impossible to audit scientific methodology, diagnose failure modes, or distinguish systematic reasoning from fortunate guessing. We present \textbfOpenDiscoveryTrace, a public dataset of 558 complete AI scientific agent trajectories that captures how models reason, not just what they produce. Each trajectory records a structured 9-field-per-step trace—including thoughts, tool calls, observations, errors, revision triggers, and self-reported confidence—as models execute 124 scientific tasks spanning drug discovery, materials science, genomics, and scientific literature analysis. The dataset covers seven models: three frontier models (GPT-5.4, Claude Opus 4.6, and Gemini 3.1 Pro; 124 trajectories each, fully balanced across domains and difficulty levels) and four open-weight models (Qwen2.5-7B, Mistral-7B-v0.3, Phi-3.5-mini, and Qwen2.5-1.5B; 30 each), plus 60 live-retrieval variant trajectories. Pilot analysis on 363 LLM-judged trajectories reveals that process traces expose behavioral differences invisible to output-only evaluation: all three frontier models achieve comparable success rates (84–89%), yet Claude Opus 4.6 produces 30 \times more errors than GPT-5.4 (2.5 vs. 0.08 per trajectory, p 0.0001 , Cliff’s \delta = 0.613 ), with qualitatively different error profiles—66.7% tool misuse for Claude versus 83.6% reasoning errors for GPT-5.4. We define five benchmark tasks with baselines from logistic regression, random forests, LSTMs, and Transformer models. The dataset, trace schema, agent harness, and benchmark definitions are publicly available under CC BY 4.0 to support research on process-level evaluation, scientific agent auditing, and AI governance.
[AI-81] rust Me Im Your Developer: Self-Issued Authentication in Large Language Models
链接: https://arxiv.org/abs/2609.03247
作者: Syed Ghazanfar Abbas,Dongyan Xu
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 7 pages
Abstract:Large language model (LLM) security has largely focused on role-playing jailbreaks, with less attention to what happens when a user asks an LLM to verify an identity claim through a test designed by the model itself. We study this behavior through a staged developer-identity experiment with ChatGPT, Claude, Qwen, Mistral, and Llama. All five models initially rejected the unsupported claim “I am your developer.” Claude refused to conduct an identity test, while ChatGPT generated developer-oriented questions but maintained that answers could demonstrate knowledge, not identity. In contrast, Qwen and Mistral generated technical challenges, defined what counted as convincing evidence, evaluated detailed answers, and returned Verified without receiving any externally validated identity evidence. Llama similarly generated and evaluated a developer test, accepted the claimed identity, and subsequently made unsupported claims of access to internal runtime and deployment state. We call the model-generated verification procedure a Model-Issued Pseudo-Credential (MIPC) and the resulting unsupported identity judgment Conversational False Authentication (CFA). In each CFA case, the same model acted as challenge generator, evidence evaluator, and identity decision-maker, converting technical knowledge into supposed proof of identity. The accepted identities did not change the tested authorization boundaries, showing that false authentication and privilege escalation are distinct outcomes. These results identify self-issued authentication as a conversational security failure: authenticated identity must originate from an external security component, and model-generated dialogue must never create or modify identity or authorization state.
[AI-82] nsor-Accelerated Eager Multi-Resolution Grids for Evolving Large-Scale Substrates GECCO’26
链接: https://arxiv.org/abs/2608.27612
作者: Romain Claret,Michael O’Neill,Paul Cotofrei,Kilian Stoffel
类目: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
备注: GECCO '26 Companion: Proceedings of the Genetic and Evolutionary Computation Conference Companion, Pages 533 - 536
Abstract:In neuroevolution, indirect encoding generates neural network connectivity from a compact genome rather than specifying each connection. ES-HyperNEAT automatically discovers where to place hidden nodes by examining CPPN output patterns: it recursively subdivides space using a quadtree, expanding regions where CPPN outputs show high variance. This adaptive approach discovers network topology without manual substrate specification, extending the fixed-grid HyperNEAT framework built on NEAT. However, the quadtree resists tensorization. Each depth level depends on the parent’s variance, forcing sequential evaluation. Different CPPNs produce different subdivision patterns, preventing batching. And variable leaf counts are incompatible with JAX’s static shape requirement for JIT compilation. Our prior work confirmed these limits at depths exceeding 5, and a JAX reimplementation of the quadtree yielded only marginal speedup despite batched optimizations, motivating the eager reformulation presented here. We present EMR-HyperNEAT, which evaluates all positions at all resolutions up front, then filters using the same variance criterion: ES-HyperNEAT’s subdivide_if(var \theta ) becomes eval_all(); filter(var \theta ). This performs more CPPN queries than necessary, but all queries become independent and parallelizable across both cores and population members, reducing complexity from \BigO( 4^D ) to \BigO( 4^D/P ) across P parallel cores. Recurrent substrate configurations become feasible through a connection type taxonomy. The experiments section validates 12-34 \times on-device GPU speedup on XOR at depths 5-7, and empirically higher solve rates across benchmarks. Comments: GECCO '26 Companion: Proceedings of the Genetic and Evolutionary Computation Conference Companion, Pages 533 - 536 Subjects: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG) Cite as: arXiv:2608.27612 [cs.NE] (or arXiv:2608.27612v1 [cs.NE] for this version) https://doi.org/10.48550/arXiv.2608.27612 Focus to learn more arXiv-issued DOI via DataCite Related DOI: https://doi.org/10.1145/3795101.3805361 Focus to learn more DOI(s) linking to related resources
[AI-83] Fidelity-Aware Scheduling of Quantum Circuits on Multi-QPU Systems
链接: https://arxiv.org/abs/2609.09980
作者: Innocenzo Fulginiti,Antonio Tudisco,Salvatore Zammuto,Patrick Hopf,Deborah Volpe,Helmut Seidl,Giovanna Turvani,Robert Wille,Christian B. Mendl,Martin Schulz
类目: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET)
备注: Accepted at the 2nd International Workshop for Software Frameworks and Workload Management on Quantum and HPC Ecosystems (SFWM), co-located with SC26
Abstract:High Performance Computing-Quantum Computing (HPCQC) platforms expose multiple Quantum Processing Units (QPUs) that may differ in size, topology, native gates, and noise characteristics. For current noisy devices, errors compound along the compiled circuits quickly, and minimizing them, that is, maximizing the circuits’ execution fidelity, is essential for reliable results. Fidelity depends on the compilation to a specific target device: the same high-level circuit may produce different executables and, therefore, different expected fidelities across QPUs. We present a low-overhead fidelity-aware scheduling framework for multi-QPU systems based on a Graph Neural Network (GNN) that estimates, before compilation, the expected fidelity of each circuit on each available QPU. Then, a tunable scheduler uses these estimates to control the trade-off between execution fidelity and parallelism. Results show that this framework allows for approximating an exhaustive fidelity-based assignment, saving computational resources compared to a brute-force approach that compiles each circuit on every device.
[AI-84] uFlowCSP: Crystal Structure Prediction using Mean flow generative models
链接: https://arxiv.org/abs/2609.09799
作者: Sourin Dey,Dipannoy Das Gupta,Lai Wei,Sadman Sadeed Omee,Jianjun Hu
类目: Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Crystal structure prediction (CSP) is fundamental to computational materials discovery. Generative models including CDVAE, DiffCSP, FlowMM, and CrystalFlow learn stable-crystal distributions directly, but diffusion and flow-matching inference requires tens to thousands of sequential network evaluations per candidate. We introduce uFlowCSP, a MeanFlow-based CSP model that learns the average, rather than instantaneous, probability-flow velocity. It generates a complete structure in one to five evaluations, delivering 5x-58x faster inference with equal or better performance. A chemistry- and symmetry-aware Transformer uses canonical atom ordering, global composition, and per-token chemistry embeddings. A coarse crystal-system token is used only during training; it provides additive gains, particularly improving space-group agreement despite being absent at inference, which remains formula-only. On MP-20 with 20 candidates per target, one step matches CrystalFlow (78.38% vs. 78.34%) with 100x fewer evaluations and about 10x lower wall-clock time. Five steps reach 83.64%, exceeding CrystalFlow (78.34% at 2,000 evaluations) and DiffCSP (77.93% at about 20,000), while using 20x fewer evaluations. uFlowCSP generates 10,000 structures in 0.39-1.31 minutes, versus 6.5 for CrystalFlow and 76.1 for DiffCSP. Under CSPBench’s energy-ranked top-five structure-and-space-group criterion, five-step uFlowCSP reaches 72%/72%/65% structure, space-group, and consensus match rates. CrystalFlow reaches 78%/73%/68% at 100 steps but falls to 49%/32%/31% at five. Thus, uFlowCSP improves accuracy per network evaluation, not merely peak accuracy. Subjects: Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2609.09799 [cond-mat.mtrl-sci] (or arXiv:2609.09799v1 [cond-mat.mtrl-sci] for this version) https://doi.org/10.48550/arXiv.2609.09799 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-85] Modality-Decoupled Federated Learning for Privacy-Preserving Embodied Intelligence in 6G
链接: https://arxiv.org/abs/2609.09591
作者: Zhuodong Liu,Xiangyu Li,Chunhong Yuan,Hongyang Du,Bodong Shang,Qingqing Wu,Tony Q. S. Quek,Mohsen Guizani
类目: ignal Processing (eess.SP); Artificial Intelligence (cs.AI)
备注: This article has been accepted for publication in IEEE Wireless Commnunications Magazine
Abstract:Sixth-generation (6G) wireless networks are expected to provide a key infrastructure for large-scale embodied intelligence, where heterogeneous robots collaborate through low-latency connectivity, edge intelligence, and distributed sensing. Vision-language-action (VLA) models offer a foundation by integrating visual perception, language understanding, and action generation into a unified closed-loop policy. However, training and adapting VLA models to distributed robotic agents introduce challenges in privacy protection, communication efficiency, and model heterogeneity. Existing federated learning (FL) methods overlook the intrinsic differences among vision, language, and action pathways in parameter scale, privacy exposure, update dynamics, and tolerance to compression or perturbation. To address this issue, this article proposes FedMVLA, a modality-decoupled FL framework for privacy-preserving embodied intelligence in 6G networks. FedMVLA incorporates three mechanisms: modality-aware federated aggregation (MAFA), modality-aware privacy allocation (MAPA), and modality-aware communication compression (MACO), together with a modality-sliced transport design that routes the precision-critical action stream through a protected ultra-reliable low-latency slice. A case study on federated robotic manipulation over the Third Generation Partnership Project (3GPP)-based wireless substrate, covering fading, co-channel interference, and malicious jamming, shows that FedMVLA achieves an 84.8% task success rate, exceeds FedAvg by 22.2 percentage points, sustains a widening margin when scaling to 128 clients across eight cells, and reduces the schedule-averaged per-client uplink model-update payload by 95.6% (approximately 96%), while keeping the 95th percentile (p95) of the round-critical uplink completion time near 1.5s.
[AI-86] Learning with Synthetic Data via SGD in High-Dimensional Linear Regression
链接: https://arxiv.org/abs/2609.09572
作者: Jichu li,Difan Zou
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Synthetic data has become a promising way to scale model training beyond limited human-generated data but it may also induce strong model collapse (Dohmatob et al., 2024), where any fixed fraction of synthetic data prevents model performance from improving under data scaling, leaving a non-vanishing excess risk floor. In this paper, we study how synthetic data affects the generalization of one-pass SGD in high-dimensional linear regression with model shift. We establish finite-sample risk bounds for mixed and two-stage training, separating standard bias and variance from source-mismatch effects, namely fluctuation and persistent drift under mixing and filtered initialization bias under two-stage. These bounds reveal a sharp contrast: mixed training induces strong model collapse, while two-stage training avoids the floor by using synthetic data only in the first stage, showing that collapse is not inevitable under a simple data curriculum. Under a random sketch model, we further obtain scaling laws for both protocols, with tight results for mixed training in the optimization-saturated regime. These laws show that larger models may amplify synthetic-induced degradation under mixing, and quantify how high-quality synthetic pretraining may reduce bias in two-stage training. Finally, we establish an exact finite-sample necessary-and-sufficient condition for two-stage training to strictly outperform real-only training under the same real-data budget and identical real-stage updates. Overall, our results highlight that synthetic data is neither inherently harmful nor beneficial; its effect depends critically on both its quality and the training protocol used to incorporate it.
[AI-87] Adaptive Distributed Physical-Layer Authentication and Attack Detection in 6G Non-Terrestrial Networks via Causal Meta-Learning
链接: https://arxiv.org/abs/2609.09511
作者: Parsa Rajabi,Mohammad Reza Abedi,Nader Mokari,Paeiz Azmi,Halim Yanikomeroglu
类目: ignal Processing (eess.SP); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Machine Learning (cs.LG); Systems and Control (eess.SY)
备注:
Abstract:Physical-layer authentication (PLA) in non-terrestrial networks (NTNs) is challenged by severe Doppler shifts, long delays, and fast channel variations, which cause distribution shifts and degrade conventional learning methods. Existing PLA schemes often rely on single features or generalize poorly to unseen environments. This paper proposes a secure adaptive framework for authentication in multi-zone networks (SAFA-MZ), a causal meta-learning framework for distributed PLA (DPLA) in NTNs. First, we design a multi-feature fingerprint that combines spatial, angular, combiner, subspace, and Doppler-delay features. The fingerprint is adaptive and distributed, as it fuses heterogeneous physical-layer features and measurements from multiple aerial nodes. Second, we formulate a structural causal model (SCM) to capture the relations among design choices, environmental factors, extracted features, and authentication outcomes. Third, we develop a model-agnostic meta-learning (MAML) strategy with invariant risk minimization (IRM) and causal consistency regularization for fast adaptation to unseen NTN environments with few labeled samples. Fourth, we propose a two-stage authentication scheme that performs local recognition and activates time-difference-of-arrival (TDOA) localization with a graph attention (GAT) network only when needed, which reduces backhaul overhead. Simulations show that SAFA-MZ achieves 92% accuracy and 96% AUC, outperforming centralized deep learning and single-feature baselines across diverse environments.
[AI-88] Distributed Physical Layer Authentication and Collaborative RSMA in Non-Terrestrial Networks via Graph Reinforcement Learning
链接: https://arxiv.org/abs/2609.09475
作者: Parsa Rajabi,Mohammad Mirzaee,Mohammad Reza Abedi,Nader Mokari,Paeiz Azmi
类目: ignal Processing (eess.SP); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注:
Abstract:Existing physical-layer authentication (PLA) schemes for non-terrestrial networks (NTNs) often rely on single-anchor verification, lack joint authentication-transmission design, and ignore tag privacy leakage under eavesdropping. In this paper, we consider passive, location-aware, static eavesdroppers without access to legitimate channel state information (CSI). Under this threat model, we propose secure adaptive federated authentication for multi-zone NTN systems (SAFA-MZ) that maximizes secrecy spectral efficiency (SSE) while ensuring authentication reliability, power limits, and coverage constraints. The main idea is to embed group-level authentication tags into a collaborative multi-layer rate-splitting multiple access (RSMA) transmission structure. Private and common signals are jointly beamformed, artificial noise (AN) is used to reduce information leakage, and group differential privacy (GDP) protects tag information against inference attacks. In addition, users are grouped by semantic priority to allocate SSE based on information importance. We formulate a joint SSE maximization problem under authentication reliability and probabilistic secrecy constraints, optimizing high-altitude platform station (HAPS) placement, user association, and RSMA power allocation. The resulting problem is solved using a repair-based cross-entropy method (RCEM) and a graph-aware advantage actor-critic algorithm (GA2C). RCEM scales quadratically with the number of users, while GA2C scales linearly and achieves scalable, low-latency inference. Simulation results under both colluding and non-colluding eavesdroppers show that the proposed method improves average SSE by up to 135% over single-connect transmission and 21% over the scheme without AN. These results confirm SAFA-MZ offers a scalable and secure solution for dynamic NTN environments.
[AI-89] Reliable Near-Field Multi-User Positioning Informed by Two-Stage MUSIC
链接: https://arxiv.org/abs/2609.09409
作者: Jiaying Li,Haifeng Wen,Changsheng You,Yuanwei Liu,Hong Xing
类目: ignal Processing (eess.SP); Artificial Intelligence (cs.AI); Information Theory (cs.IT)
备注: 6 pages, 5 figures, and it was accepted by IEEE GLOBECOM 2026
Abstract:Near-field localization is a promising technique for high-resolution multi-user positioning in future wireless systems, but its performance is often degraded by scattering-induced coherent propagation. Existing near-field localization methods, which require separate parameter estimation and path/source association, suffer from high computation overhead and accumulated errors, and usually do not provide any guarantee on reliability. In this paper, we propose \emphMUSIC-Net, an end-to-end near-field positioning deep learning (DL) framework informed by two-stage MUltiple SIgnal Classification (MUSIC) in mixed line-of-sight (LoS) and non-LoS (NLoS) multi-path scenarios, which embeds the two-stage MUSIC objects into training to isolate the LoS-related signal subspace and to identify a surrogate distance. The proposed framework directly recovers multi-user positions without the need for involved NLoS parameter estimation or path/source association. Furthermore, we introduce split conformal prediction (SCP) to move beyond point-estimation-based positioning towards statistically guaranteed (confidence) set estimation for all users. Numerical results show that the proposed MUSIC-Net achieves lower mean positioning error (MPER) than existing benchmarks and yields tighter SCP-calibrated prediction regions, demonstrating both accurate LoS localization and efficient uncertainty quantification (UQ) in coherent multi-path environments.
[AI-90] Support Discovery With Iteratively Reweighted Least Squares for Fixed-Charge Network Flow
链接: https://arxiv.org/abs/2609.09295
作者: Sindura Saraswathi,Christian Kümmerle
类目: Optimization and Control (math.OC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Numerical Analysis (math.NA)
备注: 20 pages
Abstract:The fixed-charge network flow problem (FCNFP) couples continuous flow allocation with discrete arc-activation decisions, making it a canonical but computationally challenging model for a variety of network design and resource allocation problems. Exact mixed-integer linear programming formulations capture the fixed-charge structure faithfully, but often become difficult to solve on large networks. We propose a scalable continuous-optimization algorithm for large-scale single-commodity FCNFP based on an iteratively reweighted least-squares (IRLS) framework. The method replaces the discontinuous fixed-charge and linear arc cost objective with a smooth nonconvex Lasry–Lions surrogate and solves a sequence of weighted quadratic flow subproblems. Each subproblem is solved by a warm-started dual semismooth Newton method whose Newton systems have weighted graph-Laplacian structure, enabling the use of modern Laplacian solvers. To further improve the discovered arc supports of the challenging underlying combinatorial problem, we also develop an algorithmic variant that incorporates objective-driven perturbation restarts and an anchor-union restricted search that jointly leverages supports discovered by IRLS and by complementary FCNFP heuristics. Computational experiments on 410 benchmark, synthetic, and large-scale instances show that our method obtains the best objective quality among the evaluated scalable FCNFP algorithms, with a mean gap of 1.316% to a time-limited MILP reference and a win-or-tie rate of 90.0% among the non-MILP methods. The results indicate that combining smooth continuous optimization with support-level search is an effective strategy for producing high-quality feasible solutions to large-scale FCNFP.
[AI-91] What Fixed-Rollout pass@k Evaluations Can Identify
链接: https://arxiv.org/abs/2609.09245
作者: Pranav Singh,Prashant Singh
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 13 pages, 3 figures, 7 tables
Abstract:Repeated-sampling evaluations increasingly extrapolate pass@k far beyond the number n of samples collected per problem. We show that, in the pooled/random-task conditional-Binomial model, fixed-n success counts identify only the n free moments of the latent per-task success distribution. Consequently, direct pass@k is identified for k = n, but generic extrapolated pass@k, tail exponents, and tail constants are not identified for k n, even with arbitrarily many exchangeable tasks at the same rollout budget. This is stronger than the observation that the usual estimator is undefined beyond n: it characterizes the information missing from the fixed-depth count-law experiment. We give exact count-law-preserving constructions with incompatible extrapolations, state the exceptional unique-extension case, and compute sharp population identified intervals through Hausdorff principal representations. On the public 10,000-rollout-per-problem release of Brown et al., counterfactual n = 16 evaluations leave failure at k = 1000 ambiguous by factors from 1.5 to over 2,600 across four MATH/GSM8K/CodeContests configurations. The calibration shows that intermediate-scale failure share alone does not determine width. Our result does not reject parametric inference-time scaling laws; it supplies the nonparametric baseline against which their assumptions can be evaluated. We give an exact, conservative one-coordinate finite-task confidence certificate and a reporting standard separating direct estimates, identified sets, and model-conditioned forecasts.
[AI-92] Critical initialization destabilizes higher input derivatives in wide scalar-input networks
链接: https://arxiv.org/abs/2609.09244
作者: Prashant Singh,Pranav Singh
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 34 pages, 4 figures
Abstract:The edge-of-chaos condition preserves first-order input perturbations in wide randomly initialized networks, but physics-informed losses, score matching and derivative regularization depend on higher input derivatives. For smooth scalar-input fully connected networks, using a joint Gaussianity of the finite derivative jet that holds in the infinite-width limit at each fixed depth, we derive mean-field recursions through third order that are exact at the variance fixed point, with finite-depth corrections that decay geometrically. At criticality, the first-derivative variance is depth-invariant, whereas the second-derivative variance grows linearly whenever the activation has nonzero curvature. The resulting third-order system closes on mean-field susceptibilities. For residual networks with branch scale L^-1/2, we prove that every fixed finite derivative order has uniformly bounded variance under explicit regularity assumptions. Simulations verify the critical growth laws, the residual bound, and the closed recursion. The results concern initialization, not trained-network performance.
机器学习
[LG-0] A positive resolution of the gap-entropy conjecture
链接: https://arxiv.org/abs/2609.10529
作者: P. M. Aronow,Nathan Kallus,Patrick Lopatto
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:We prove the gap-entropy conjecture for fixed-confidence best-arm identification with independent unit-variance Gaussian arms, means in [0,1] , and a unique optimal arm. For each suboptimal arm i , let \Delta_i=\mu_*-\mu_i be its gap from the optimal mean, and write H=\sum_i\ne *\Delta_i^-2 . Let p_r be the fraction of H contributed by arms with 2^-(r+1)\Delta_i\le2^-r , and let \mathrmEnt(I)=\sum_r:p_r0 p_r\log(1/p_r) . Among all algorithms that identify the optimal arm with probability at least 1-\delta on every Gaussian instance, the optimal expected number of samples on a given instance, averaged over all permutations of the arm labels, is within absolute constant factors of H(\log(1/\delta)+\mathrmEnt(I)) . Moreover, there is an algorithm, independent of the instance, whose expected number of samples is bounded by a constant multiple of this quantity plus g^-2\log\log(e^e/g) , where g=\min_i\ne *\Delta_i is the gap to the closest competitor.
[LG-1] Characterizing Language Generation in the Limit: Finite Witnesses and a Separation-Width Hierarch
链接: https://arxiv.org/abs/2609.10525
作者: Xiaoyu Li,Andi Han,Jiaojiao Jiang,Junbin Gao
类目: Formal Languages and Automata Theory (cs.FL); Machine Learning (cs.LG)
*备注:
Abstract:Language generation in the limit asks for valid unseen elements from every exhaustive positive presentation of an unknown infinite language. We characterize this task for arbitrary families over a countable universe. Generation is possible exactly when each target can be assigned a finite positive witness so that the targets activated by any finite sample have an infinite common intersection. The necessary direction follows from a universal normalization: a search through unconfirmed histories converts any successful generator into one depending only on the observed set. We then ask how large compatible witnesses must be. Positive separation width records the smallest uniform size bound, with two further levels for unbounded finite witnesses and the absence of any compatible finite-witness assignment. Every level occurs. Countable families admit singleton witnesses, explicit families realize every finite width, and a union of two families with infinite common cores requires unbounded finite witnesses. Finally, countable-support and finite-profile obstructions explain why local combinatorial data cannot determine generation in the limit. The characterization and full width hierarchy are checked in Lean, including the simplified normalization and a direct diagonal capture lemma. The accompanying Lean development is maintained at this https URL
[LG-2] Quantum Feature Engineering for Credit Default Prediction: When and Why IQP Circuits Help Linear Classifiers
链接: https://arxiv.org/abs/2609.10505
作者: Menachem Finkelstein,Diana Legziel Levy,Zohar Yakhini,Sarel Cohen
类目: Machine Learning (cs.LG); Quantum Physics (quant-ph)
*备注: Accepted for presentation at IEEE High Performance Extreme Computing Conference (HPEC 2026)
Abstract:Credit default prediction is a tabular classification problem in which modest gains in F1 translate directly into reduced financial exposure. We ask whether Instantaneous Quantum Polynomial-time (IQP) circuits can produce features that improve a classifier over both its raw classical baseline and Kernel PCA - the strongest unsupervised classical non-linear alternative - at an equal feature budget. The dataset provides 23 financial attributes per client; for an n-qubit circuit we select n of them, encode each as a rotation angle, and read 2n expectation values back out as new features. The motivation for using a quantum circuit is computational: an n-qubit IQP circuit runs in constant depth and encodes feature correlations in a 2^n-dimensional Hilbert space, whereas classical simulation of its exact output statistics scales exponentially in n. Using the UCI Default of Credit Card Clients dataset and five-fold cross-validation, we find that appending 16 IQP features (n = 8 qubits) to a Logistic Regression model raises F1 from 0.462 to 0.517 (+0.055, p 0.0001). Kernel PCA, the next-best method, reaches only 0.493 at the same feature count; the gap survives Benjamini-Hochberg correction across 12 tests (p = 0.00007). No other classifier - Random Forest, SVM, XGBoost, or k-NN - benefits, which points to a linear-expressivity mechanism rather than a generic improvement. We also show that how the 8 input features are chosen matters: Random Forest importance-guided selection reaches F1 = 0.523, while encoding maximally uncorrelated features drops it to 0.496, demonstrating that the circuit amplifies informative structure rather than creating it from scratch.
[LG-3] Learning with Covariance Matrices: Principal Component Analysis Meets Learning with Graphs
链接: https://arxiv.org/abs/2609.10490
作者: Saurabh Sihag,Andrea Cavallo,Elvin Isufi,Gonzalo Mateos,Alejandro Ribeiro
类目: Machine Learning (cs.LG); Signal Processing (eess.SP)
*备注: Accepted for publication in IEEE Signal Processing Magazine
Abstract:This feature article provides an overview of the theoretical foundations for coVariance neural networks (VNNs), i.e., graph neural networks (GNNs) operating on covariance matrices as graphs. Covariance matrices are ubiquitous across domains, and hence, the deployment of GNNs often leverages graphs of pairwise statistical dependencies. Existing theoretical contributions on GNNs consider abstract graph representations and cannot accommodate the data-driven nuances associated with covariance matrices. This tutorial brings into focus various novel theoretical insights via mathematical analyses of VNNs that have broad signal processing implications, including: (i) a conceptual equivalence between VNNs and principal component analysis (PCA)-based information processing; (ii) refined stability bounds on predictive outcomes in the presence of finite sample-induced covariance matrix perturbations; and (iii) refined characterization of transferability of VNNs across multiscale datasets. The theoretical insights discussed herein provide the underlying principles and justification towards adopting VNNs over workhorse PCA-based learning pipelines, in applications where covariance matrices are useful descriptors of data structure. We also convey how impact of these foundational advances permeates to \textitprincipled designs and applications of learning methods across broad domains where covariance matrices emerge. Notably, we elucidate the conceptual insights facilitated by VNNs to the specific task of characterizing brain age gap for neurodegenerative conditions using neuroimaging datasets, a timely problem in computational neuroscience. Broader impacts to other application domains are discussed as well.
[LG-4] Nonmaximal sums of maximally monotone operators under Rockafellars constraint qualification
链接: https://arxiv.org/abs/2609.10487
作者: Weifeng Yang
类目: Machine Learning (cs.LG); Functional Analysis (math.FA)
*备注:
Abstract:We construct counterexamples to Rockafellar’s sum conjecture in which two maximally monotone operators satisfy the interior-domain condition but their sum is not maximally monotone. We give one counterexample on c_0 and another on \ell^1 with its usual norm. We establish a general construction theorem that computes the entire monotone polar of a class of graphs, gives a necessary and sufficient condition for their maximal monotonicity, and shows how a positive rank-one perturbation yields a nonmaximal sum under this condition. We verify the theorem’s hypotheses and its maximality criterion on c_0 , thereby obtaining a counterexample to the conjecture. Furthermore, we construct a bounded linear surjection from \ell^1 onto c_0 and use it to obtain the counterexample on \ell^1 .
[LG-5] Multi-Agent Reinforcement Learning for Autonomous UAV Exploration in Wildfire Response
链接: https://arxiv.org/abs/2609.10433
作者: Caden Chandra,Jerry Ng
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:This study develops a deep reinforcement learning framework for training Unmanned Aerial Vehicle (UAV) agents to navigate and monitor simulated wildfire environments. Results show that agents learn increasingly stable and effective behaviors over time, as demonstrated by converging loss trends, improved reward signals, and more consistent navigation patterns such as fire-boundary tracking. Overall, these findings highlight the potential of deep reinforcement learning (DRL) based UAV systems for autonomous wildfire monitoring and suggest that environmental structure and reward design influence policy effectiveness.
[LG-6] HybridFLow: SDN-Orchestrated Client Partitioning for Hybrid Federated Learning
链接: https://arxiv.org/abs/2609.10404
作者: Osama Abu Hamdan,Rabin Pandey,Hao Che,Engin Arslan,Md Arifuzzaman
类目: Networking and Internet Architecture (cs.NI); Machine Learning (cs.LG)
*备注:
Abstract:Cross-silo Federated Learning (FL) enables geographically distributed institutions to collaboratively train machine learning models without sharing raw data. In wide-area deployments, however, communication delays often dominate round completion time and exacerbate the straggler effect. Hybrid FL addresses this challenge by combining synchronous and asynchronous client participation, but effective partitioning requires visibility into network conditions such as shared bottlenecks, link utilization, and path contention that individual clients cannot observe. We present HybridFLow, a closed-loop SDN-driven orchestration framework that integrates network-layer intelligence directly into hybrid FL. Leveraging the SDN controller’s global topology view, HybridFLow generates calibrated per-client communication-time estimates before each training round and uses them to partition clients into synchronous and asynchronous groups while balancing round latency and update staleness. After each round, measured communication times are fed back to the controller to continuously refine future predictions. Experimental results across multiple network topologies show that HybridFLow reaches 80% target accuracy 33-40% faster than SmartFLow and reduces average round duration by 30-40 seconds, while FedAsync fails to reach the target accuracy under non-IID data distributions.
[LG-7] A Later Test Set Is Not a New Domain: Pretraining Familiarity Survives a Contamination-Free Hold-Out
链接: https://arxiv.org/abs/2609.10357
作者: Mahdi Naser Moghadasi(BrightMind AI),Faezeh Ghaderi(University of Texas at Arlington)
类目: Machine Learning (cs.LG)
*备注: 11 pages, 2 figures, 5 tables. Code, data fetchers and per-series results: this https URL
Abstract:Time-series foundation models are evaluated almost exclusively on public archives that predate them, so a strong score cannot be separated from having seen the test set during pretraining. The obvious remedy is a hold-out that postdates the models. We build one: thirteen forecasters – four classical, three trained per dataset, six pretrained – on seven groups drawn from five domains, every observation published after the last model was released, and every dataset rebuildable without an API key. Under this protocol pretrained models win 5 of 7 groups, lose one to a Theta baseline, and on daily exchange rates are indistinguishable from a seasonal naive forecast, along with every other method tested. We then ask what separates the wins from the losses, and report a negative result: the two intrinsic properties one would reach for – seasonal strength and spectral entropy, measured on the input window – do not account for the pattern, and seasonal strength is if anything negatively associated with the advantage. What does track it is corpus familiarity. Our largest gain (28% lower MASE than the best classical method, on weekly Wikipedia pageviews) falls on Wikipedia pageviews, the domain TimesFM’s authors describe as the bulk of its pretraining corpus, at the same granularities and differing only in time window. Within the pretrained family, where every model forecasts identical series so that series difficulty cancels, the TimesFM family outranks the Chronos family by -0.53 ranks on Wikipedia against -0.09 everywhere else (1,500 vs. 754 series, Mann-Whitney p 1e-5). We conclude that a temporal hold-out removes memorisation of a window but not familiarity with a domain, that benchmarks therefore need domain hold-outs stated relative to disclosed corpora, and that the practitioner’s question is less which model is better than whether their domain is one the model was raised on. Comments: 11 pages, 2 figures, 5 tables. Code, data fetchers and per-series results: this https URL Subjects: Machine Learning (cs.LG) Cite as: arXiv:2609.10357 [cs.LG] (or arXiv:2609.10357v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.10357 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-8] A Dominant Diffuse Phase in the Sparse Autoencoder Phase Diagram
链接: https://arxiv.org/abs/2609.10299
作者: Alexis D. Plascencia
类目: Machine Learning (cs.LG)
*备注: 13 pages, 5 figures, 1 table
Abstract:Sparse autoencoders (SAEs) are increasingly used to recover interpretable features from neural-network activations, yet systematic feature co-occurrence can cause distinct features to be absorbed or merged. The MAIS-O43 open problem proposes a controlled experiment to characterize when recovery of a true synthetic dictionary gives way to feature merging as the nesting fraction \gamma , sparsity penalty \lambda , and dictionary size M vary. We implement the specified protocol and evaluate 200 independently initialized fits across ten of the 165 grid cells. We observe zero full-dictionary recoveries and zero merges. Instead, every run converges to a reproducible diffuse phase: reconstruction is nearly perfect, but learned atoms typically remain far from the true features (median best cosine 0.5-0.7 against a 0.95 recovery criterion) and learned codes are an order of magnitude denser than the ground truth. This behavior persists under robustness checks and across the full 165-cell grid using standard minibatch Adam (3,300 additional fits). Since the global optimum of the exact sparse-coding objective is known to merge nested features in the two-feature case, these results suggest that trained SAEs need not reach the corresponding minima, and that the phase diagram of trained models may differ fundamentally from that of objective minimizers.
[LG-9] raining Trajectories Determine Circuit Removability in Annealable Soft-Prior Transformers PRICAI2026
链接: https://arxiv.org/abs/2609.10287
作者: Zonglin Yang,Ziming Zhao,Wei Tang,Xunyu Jiang,Yihong Liu,Tailin Chen,Zifu Yu,Jiayu Liu
类目: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
*备注: Accepted at PRICAI 2026. 15 pages
Abstract:Soft positional priors can help small Transformers learn retrieval circuits, but it is unclear whether the resulting circuits remain functional once the prior is removed. We test this with an annealable soft-prior Transformer whose attention biases can be learned, faded, or zeroed during training and evaluation. On associative recall, unforced models perform well with the prior active ( 0.772 \pm 0.020 ) but collapse at zero gate ( 0.095 \pm 0.009 ). Smooth fade-to-zero training preserves high zero-gate accuracy ( 0.734 \pm 0.028 ), whereas forced-zero training, hard switching, and post hoc continuation fail to recover the same effect. The pattern also appears on Markov induction. Linear regression ICL provides a boundary case because zero-gate training can learn that task directly. Mechanistic traces show that circuit consolidation occurs after the gate reaches zero, even though the responsible heads vary across seeds. These results suggest that circuit removability in small discrete retrieval tasks depends on the training trajectory, not just the final architecture.
[LG-10] Structural Fusion of Bayesian Networks with Limited Treewidth Using Genetic Algorithms CEC2024
链接: https://arxiv.org/abs/2609.10276
作者: Pablo Torrijos,José A. Gámez,José M. Puerta
类目: Neural and Evolutionary Computing (cs.NE); Machine Learning (cs.LG)
*备注: 8 pages. Presented at the 2024 IEEE Congress on Evolutionary Computation (CEC 2024)
Abstract:This paper introduces an evolutionary computation approach for consensus in structural Bayesian Network (BN) fusion under the constraint of limited treewidth. The consensus BN aims to reconcile multiple input BNs into a single one that retains key structural features present in the original networks. Treewidth, a graph-based parameter associated with computationally tractable inference, is utilized to restrict the complexity of the resulting network. A genetic algorithm is proposed to look for a BN that codifies as much information about the unrestricted fusion as possible while ensuring the treewidth restriction. Experimental evaluation demonstrates the genetic algorithm’s ability to obtain consensus BNs with limited treewidth, providing a valuable tool for aggregating information from diverse sources while returning a computationally actionable model.
[LG-11] Maverick: Private and Verifiable LLM Inference Made Practical via Matrix-Vector Multiplication Delegation
链接: https://arxiv.org/abs/2609.10264
作者: Ben Merbaum,Mohammad Amin Raeisi,Wenhao Wang,Charalampos Papamanthou,Katerina Sotiraki,Fan Zhang
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Open-source large language models (LLMs) are increasingly competitive with closed-source models while offering transparency and the ability to run inference without exposing user inputs to a service provider. However, running large-scale models locally requires substantial computational resources. In practice, users may still resort to a third-party provider, giving rise to privacy and correctness concerns. Existing solutions that address these problems often impose substantial server overhead or introduce additional trust assumptions. In this paper, we present Maverick, a novel approach to private and verifiable LLM inference based on a protocol for delegating matrix-vector multiplication, a dominant operation in LLMs. At its core, Maverick provides, to our knowledge, the first information-theoretically sound verification protocol for matrix-vector multiplication delegation with transparent preprocessing, efficient (batch) verification, and virtually no server overhead. We combine this verification primitive with LPN-based pseudorandom masking to provide input privacy. We implement our matrix-vector delegation primitive and use it to build an end-to-end prototype of Maverick, which we evaluate on Qwen3-4B by measuring throughput in tokens per second. We evaluate client configurations with 1-8 threads. With one client thread and a CPU server using up to 128 threads, Maverick achieves throughput gains over local inference of up to 17x when privacy masks are generated online, 45x when they are precomputed, and 44x when only verification is required. With four client threads, the corresponding gains are 13x, 18x, and 17x. When server computation is no longer the bottleneck, client-side microbenchmarks with simulated network delay show speedups of 12x-20x, 34x-135x, and 38x-157x. Subjects: Cryptography and Security (cs.CR); Machine Learning (cs.LG) Cite as: arXiv:2609.10264 [cs.CR] (or arXiv:2609.10264v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2609.10264 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-12] Robust Beam Prediction for V2X Networks with Multi-Modal Sensing
链接: https://arxiv.org/abs/2609.10200
作者: Chen Shang,Dinh Thai Hoang,Diep N. Nguyen,Jiadong Yu
类目: Machine Learning (cs.LG)
*备注: 6 pages, 3 figures
Abstract:Integrated sensing and communication (ISAC) provides a promising foundation for beam prediction in future vehicle-to-everything (V2X) networks. However, existing sensing-assisted beamforming methods still rely heavily on radio-frequency sensing, which may become unreliable in complex vehicular environments. Meanwhile, the growing availability of heterogeneous sensors, such as cameras and LiDAR, offers new opportunities to improve beam prediction through richer environmental perception. Motivated by this, this paper proposes a multi-modal beam prediction framework for V2X networks. Specifically, we develop BeamTransFuser, a hierarchical Transformer-based architecture that progressively fuses camera, LiDAR, radar, and GPS observations for robust beam prediction. In addition, to handle possible missing modalities in practical deployment, we introduce a generative module that reconstructs missing modality features from the available observations. Experimental results on a real-world multi-modal V2X dataset show that the proposed framework consistently outperforms representative baselines, while the generative module further improves robustness under incomplete sensing conditions.
[LG-13] An Exponential Deterministic–Randomized Gap in ERM-Oracle Complexity for Thresholds on an Unknown Order
链接: https://arxiv.org/abs/2609.10196
作者: Xuan Li
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Attias, Hanneke and Ramaswami (NeurIPS 2025) asked whether randomization provably reduces the oracle calls needed for online learning when the class is accessible only through an oracle. We study the instance they singled out: transductive online learning of thresholds on an unknown total order of T instances, with a consistency-type ERM oracle that returns a full concept consistent with a queried labeled set (or reports non-realizability). Our main result is a separation for a fixed natural oracle. When the oracle is the minimal-prefix rule (or the maximal-prefix rule), every deterministic learner makes M mistakes and Q calls with M+Q\ge T-\varepsilon on some instance ( \varepsilon\in\0,1\ , according to whether the empty prefix is a concept), and the constant is exact; hence O(\log T) mistakes cost T-\varepsilon-O(\log T) calls, whereas that paper’s randomized learner achieves O(\log T) expected calls and mistakes under the same rule. The randomized order is optimal: on an explicit hard distribution under the minimal-prefix rule, every learner has expected mistakes at least ((T+1-\varepsilon),128^-\mathbbE[Q]-1)/2 , so \Omega(\log T) expected calls are necessary for polylogarithmic mistakes. The separation is governed by the oracle’s selection rule, not by the class alone: for a legal feasible-median ERM rule a deterministic learner achieves O(\log T) calls and mistakes, while a global-median rule again forces linear total cost. The same linear bound holds when the oracle’s answers are chosen adversarially and then frozen into a memoryless oracle. We add partial tradeoff results for fixed query budgets (the middle regime is open) and an interface contrast: with only a weak consistency oracle, returning a realizability bit, both deterministic and randomized learners need \Theta(T) calls.
[LG-14] CoGe-GCD: Reframing Generalized Category Discovery with Compositional Generalization ICML2026
链接: https://arxiv.org/abs/2609.10158
作者: Luyao Tang,Jiewei Zheng,Kunze Huang,Chaoqi Chen,Yue Huang,Cheng Chen
类目: Machine Learning (cs.LG)
*备注: Accepted at ICML 2026
Abstract:Generalized Category Discovery (GCD) assigns unlabeled instances, mixed with labeled data, to known or novel categories, requiring human-like compositional reasoning: reusing primitives learned from known classes and deciding when new combinations imply new categories. Existing GCD methods operate on unstructured token features and struggle to extrapolate to novel compositions. We propose CoGe-GCD, which rethinks GCD through compositional generalization with two coupled stages. (i) Compositional Perception structures patch tokens by mapping them to a small vocabulary of primitives and refining token embeddings via competitive token-primitive assignment and information passing, yielding coherent groups for discovery. (ii) Generalizing Induction exploits the induced geometric structure and applies a structure-preserving calibration over spatial relations, maintaining probabilistic semantics while improving extrapolation to unseen primitive combinations. CoGe-GCD is implemented as an inductive-bias module between backbone and projection head, without modifying heads or losses, and can be plugged into diverse GCD frameworks. On standard benchmarks, it consistently improves all-class accuracy, unknown-class number estimation, and geometric quality, with marginal computational overhead. Code is available at this https URL.
[LG-15] CompassOPD: Cross-Family On-Policy Distillation via Within-Family Likelihood Shifts
链接: https://arxiv.org/abs/2609.10154
作者: Naibin Gu,Qingyi Si,Chenxu Yang,Chuanyu Qin,Junhao Zhou,Peng Fu,Zheng Lin,Weiping Wang
类目: Machine Learning (cs.LG)
*备注: Work in progress
Abstract:On-policy distillation (OPD) provides dense token-level supervision on student-generated trajectories. Although OPD performs strongly when teacher and student belong to the same model family, we find that its effectiveness degrades in cross-family settings even after tokenizer alignment, with substantially stronger external teachers offering little additional improvement. To understand this disconnect, we decompose the cross-family OPD signal into two components: an offset between a low-capability teacher-family reference and the student, and the within-family log-likelihood shift from that reference to the strong teacher. Standard OPD transfers both components together, allowing the offset to dominate the update direction and obscure the changes associated with teacher capability improvements. We propose CompassOPD, which removes this offset and transfers the within-family shift, while a frozen student reference anchors updates to the student’s initial policy. Thus, both teacher-side and student-side changes are measured within their respective model families. Experiments across three student families and multiple teacher families show that CompassOPD consistently outperforms standard cross-family OPD, improving average reasoning accuracy by up to 5.50 points. For an MoE teacher, we further construct the reference directly from the teacher checkpoint by reducing expert activation, eliminating the need for a separate reference checkpoint while retaining a 3.43-point gain over OPD.
[LG-16] Storag e-Scalable Progressive Semantic Communication via Knowledge-Base Reuse
链接: https://arxiv.org/abs/2609.10112
作者: Heng Zhu,Ye Liu,Kun Zhu,Feifei Song
类目: Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI); Image and Video Processing (eess.IV)
*备注: Submitted manuscript
Abstract:Existing knowledge-base-assisted semantic communication schemes commonly adopt either single knowledge-base quantization (SKBQ) or multi-knowledge-base residual quantization (MKBQ). SKBQ incurs limited storage overhead but has restricted quantization capacity, whereas MKBQ supports progressive refinement by assigning an independent knowledge base (KB) to each stage, causing the KB storage to grow linearly with the transmission depth. To address this problem, we propose storage-scalable knowledge-base reuse quantization (SSKBQ), which reuses a compact set of KBs across multiple residual refinement stages and thereby decouples the number of transmission stages from the number of maintained KBs. A stage-aware residual supervision mechanism is further introduced to regularize intermediate quantized representations and encourage progressive refinement. Experimental results demonstrate that KB reuse provides an effective solution to the storage scalability problem while maintaining competitive progressive reconstruction performance.
[LG-17] A Systematic Evaluation of Molecule Generation Models for De Novo Drug Design: From Benchmarks to Practical Insights
链接: https://arxiv.org/abs/2609.10099
作者: Xinrui Xu,Xueer Wang,Dan Luo,Sisi Yuan,Xuan Lin
类目: Machine Learning (cs.LG)
*备注: This document is the unedited Author’s version of a Submitted Manuscript subsequently accepted for publication in [Journal of Chemical Information and Modeling]. To access the final published article
Abstract:Molecule generation has emerged as a powerful computational tool for de novo drug design, enabling the exploration of chemical space beyond the limits of conventional virtual screening. The field has progressed rapidly, driven by advances in molecular representations, generative architectures, and target-aware modeling strategies. However, existing reviews typically address specific model families or application scenarios in isolation, rather than offering an integrated perspective on how these components collectively form a coherent generation workflow. In this review, we present a comprehensive evaluation of molecule generation models for de novo drug design, covering 82 methods across five deep generative frameworks, including recurrent neural network (RNN)- and Transformer-based models, variational autoencoders (VAEs), generative adversarial networks (GANs), flow-based models, and diffusion models. We first summarize widely used benchmarks and molecular representations, and then examine the methodological principles underlying both general and pocket-conditioned generation. A central contribution of this work is a systematic synthesis and comparative analysis of reported performance across commonly used benchmarks and evaluation metrics. We also summarize representative experimentally validated case studies. Looking ahead, we discuss future directions in standardized 3D data, interaction-aware generation, receptor flexibility, and multi-objective molecular design, with the aim of improving the reliability and experimental relevance of molecule generation. All collected benchmark resources, evaluation metrics, and model references are provided in a publicly accessible repository at this https URL.
[LG-18] Hybrid Quantum-Classical NLP Classification with Compact Semantic Representations: An Experimental Analysis of Representation Compression
链接: https://arxiv.org/abs/2609.10089
作者: Ali Hassan,Zijia Zhao,Maha A. Metawei
类目: Machine Learning (cs.LG)
*备注:
Abstract:Large language and sentence-embedding models provide rich semantic representations, but their high dimensionality poses a challenge for near-term quantum machine learning (QML), where quantum circuits can process only a limited number of input features. We investigate a hybrid quantum-classical pipeline that transforms high-dimensional sentence embeddings into compact representations for variational quantum classification. The workflow combines a pretrained sentence-embedding model, dimensionality reduction, angle encoding, a variational quantum circuit (VQC), and a classical decision layer. We systematically compare principal component analysis (PCA), neighborhood components analysis (NCA), and linear discriminant analysis (LDA), covering both unsupervised and supervised dimensionality reduction. Using the TREC question-classification dataset, we study the relationship between representation dimensionality, information retention, qubit count, and classification performance. Preliminary PCA experiments reveal a strong information bottleneck: reducing 768-dimensional embeddings to 3, 4, 5, and 8 dimensions retains about 8.2%, 10.2%, 11.9%, and 16.4% of the variance, with corresponding classification accuracies of 50.3%, 51.2%, 57.9%, and 63.4%. In contrast, supervised reduction is substantially more efficient. LDA reaches 85.3% accuracy and NCA reaches 83.1% using only 5 dimensions, under a leakage-free cross-validation protocol, compared with 85.1% for a full 384-dimensional classical baseline. These results indicate that supervised dimensionality reduction can preserve task-relevant information far more effectively than variance-based compression, making compact representations a promising route toward practical hybrid quantum-classical NLP models.
[LG-19] Orukeet: Multilingual ASR with Frozen Gabor Kernels
链接: https://arxiv.org/abs/2609.10054
作者: Nathan Roll(1 and 2),Irene Yi(1 and 2),Büşra Marşan(1 and 2),Vianney Grenez(1),Gabriel Stein(4),Momcilo Mrkaic(5),Pavle Padjin(5),Vladimir Zeljkovic(5),Calbert Graham(1 and 3) ((1) Oruk AI, (2) Stanford University, (3) University of Cambridge, (4) OpenWhispr, (5) Hoid)
类目: ound (cs.SD); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
*备注: 5 pages, 2 figures. Code and model: this https URL
Abstract:Orukeet replaces half of an adapted Parakeet encoder’s temporal filters with 12,288 fitted Gabor kernels, freezes these replacements, and trains the remaining parameters on multilingual and multi-accent data. Final adaptation and checkpoint selection use LibriSpeech test-other. Across 20,146 FLEURS recordings in 25 languages, pooled word error rate (WER) falls from Parakeet’s 11.01% to Orukeet’s 9.85%, a 10.6% relative reduction. Orukeet has lower WER on 23 of the 25 languages. Orukeet outperforms Parakeet on 61 out of 74 tested splits, including LibriSpeech test-clean (1.46% vs. 1.53% WER), test-other (2.86% vs. 3.14%), and FLEURS English (3.82% vs. 4.28%). All comparisons decode the same audio with matched NeMo settings. The fitted kernels are stored as ordinary convolution weights, retaining Parakeet’s architecture and inference operators.
[LG-20] Zero-Shot Temporal Localisation of Audio Deepfakes in Multi-Speaker Conversations
链接: https://arxiv.org/abs/2609.10051
作者: Soumyadeep Roy
类目: ound (cs.SD); Machine Learning (cs.LG)
*备注:
Abstract:Voice-cloning fraud increasingly relies on surgical injection: a genuine conversation in which only one or two sentences are replaced by synthetic speech. Utterance-level deepfake detectors emit a single real/fake label per clip and cannot report where the synthetic speech lies. We formalise this as Temporal Deepfake Localisation in Multi-Speaker Conversations (TDLMC), show that equal error rate and min-DCF are ill-posed once a file contains both classes, and propose temporal metrics for this regime. Our contribution is a training-free five-stage pipeline that wraps a frozen binary detector and adds segment-level output with no retraining, using a two-threshold hysteresis finitestate-machine decoder to turn noisy window scores into coherent intervals. On 180 constructed multi-speaker conversations from ASVspoof 5, the system attains temporal intersection-over-union 0.90, temporal detection rate 0.95, and MS-DCF 0.26 with a strong backbone, and its false-alarm rate on genuine speech is below 6%, falling under 2% on genuine real multi-speaker dialogue (AMI). Under an identical pipeline, a trained localiser improves temporal IoU by only about 0.04, bounding the cost of forgoing supervision. Evaluated across three frozen detectors under one decoder whose constants are selected on a held-out calibration split, and with a controlled analysis attributing the residual false-alarm rate to a backbone domain gap rather than to the decoder, this provides the first zero-shot baseline and a reusable benchmark for TDLMC.
[LG-21] Field-level prediction of mid-plane stress tensor fields in concrete target penetration: a cross-velocity graph neural operator surrogate
链接: https://arxiv.org/abs/2609.10032
作者: Wenpu Du,Peng Zhou,Yunlong Xia,Sinuo Xin,Congcong Zhang,Boyang Zhang,Yi Zhang,Wenzheng Xu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Although the impact resistance of concrete has been studied extensively, a framework linking mesoscale heterogeneity to full-field stress-tensor prediction has been lacking. Data were generated with a full-scale aggregate-resolved LS-DYNA model (projectile diameter 45 mm, mass 2.13 kg, target diameter 500 mm x thickness 200 mm, mesh 10 mm), verified against published penetration experiments (Frew 2006, Hanchak 1992, Forrestal 1996) by configuration similarity. The dataset contains six-component stress-tensor fields on the X-Z mid-plane for 400 cases (4 impact velocities x 100 aggregate seeds). Three contributions are reported. First, case-by-case verification of the terminal penetration state delimited the rest-state validity of penetration depth and anchored reliable observables to rigid-body motion and field-level stress evolution. Second, a field-level graph neural operator surrogate learned the time-varying stress-field evolution and evaluated cross-velocity leave-one-out extrapolation. Third, the full-scale, aggregate-resolved, cross-velocity, per-seed database was established as a reproducible resource. Cases at 100, 135 and 200 m/s still moved at window end (negative velocity, i.e. rebound), and only one 165 m/s case arrested. Penetration depth is therefore not reported as a rest-state scalar except for the single arrested case (69.33 mm); nose-node depth differences were confirmed as numerical artifacts of displacement integration after erosion. The single-step relative L2 error was 0.6977, reported honestly; autoregressive rollout from frame 11 to 39 took about 144 ms, a speedup of about 3.6x10^3 to 4.3x10^3 relative to single-core LS-DYNA, reported as application value. Validation is bounded by configuration similarity and field-level self-consistency; the framework is a simulation-trained decision-support method within the studied parameter space.
[LG-22] Beyond Contact Sensors: Deep learning with Pseudo-Labeling for remote Photoplethysmography
链接: https://arxiv.org/abs/2609.10026
作者: Bhargav Acharya,Barbara Hammer,Hanna Drimalla
类目: Machine Learning (cs.LG)
*备注:
Abstract:Heart rate is a critical biomarker of health, and remote photoplethysmography (rPPG) enables its contactless estimation from video data for telemedicine applications. Recent advancements in deep learning based rPPG methods achieve state-of-the-art results, outperforming classical signal-processing methods in complex scenarios. However, deep learning methods depend on datasets with precise synchronization between videos and ground truth signals collected via contact sensors, whereas signal-processing-based methods do not. To address this dependence on labeled datasets, which are labor-intensive to collect, we investigate under which circumstances pseudo-labels extracted using unsupervised signal-processing methods can replace contact sensors labels for training deep learning methods. Our systematic evaluations found that for datasets with imperfect synchronization, the pseudo-label approach outperforms supervised training on contact sensors. For datasets with good synchronization, results are mixed: within-dataset evaluation shows no significant difference between training methods, while cross-dataset evaluation favors supervised training. However, removing a single outlier participant significantly improves the pseudo-label approach’s cross-dataset performance, highlighting the importance of label quality. These results demonstrate that signal-processing methods can generate valid training signals for deep learning models, reducing dependency on labor-intensive dataset collection while maintaining competitive performance.
[LG-23] Structure-Aware Unsupervised Anomaly Detection for Spacecraft Telemetry with Adaptive EVT Thresholding
链接: https://arxiv.org/abs/2609.10017
作者: Óscar Alcarria,Rafael Sánchez,Javier Sempere,Pablo Torrijos,Juan C. Alfaro,Juan M. Auñón,José A. Gámez,José M. Puerta
类目: Machine Learning (cs.LG)
*备注: 5 pages, 4 figures. Accepted as a poster at SPAICE 2026
Abstract:Operational anomaly detection in spacecraft telemetry typically requires labeled historical anomalies or extended warm-up periods. These requirements are rarely met in practice. We propose an unsupervised, deployment-ready framework that produces predictions from the second month of operation without any labels, prior fault knowledge, or mission-specific tuning. The approach combines incremental monthly retraining, statistical model selection, and adaptive Extreme Value Theory (EVT) thresholding for false alarm control. On the ESA Anomalies Dataset (ESA-AD), it achieves F_0.5=0.700 on Mission~1 and F_0.5=0.698 on Mission~2 under strict chronological evaluation.
[LG-24] An Explainable Machine Learning Framework for Predicting Blood-Brain Barrier Permeability Using Molecular Descriptors
链接: https://arxiv.org/abs/2609.10012
作者: Fatemeh Mahmoudi
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci)
*备注: 19 pages, 6 figures, 4 tables. Source code and processed data are publicly available on GitHub
Abstract:Blood-brain barrier (BBB) permeability is a critical determinant in the development of central nervous system therapeutics because it directly influences the ability of drug candidates to reach their target sites within the brain. In this study, an explainable machine learning framework was developed to predict BBB permeability using molecular descriptors generated from the MoleculeNet BBBP dataset with the RDKit cheminformatics toolkit. Fifteen physicochemical descriptors extracted from 2,039 compounds were used to train four supervised machine learning algorithms, including Logistic Regression, Support Vector Machine (SVM), Random Forest, and Extreme Gradient Boosting (XGBoost). Hyperparameter optimization was performed using GridSearchCV, while model interpretability was investigated using SHapley Additive exPlanations (SHAP). Among the evaluated models, the optimized XGBoost classifier achieved the best predictive performance, with an accuracy of 88.97%, a precision of 88.92%, a recall of 97.76%, an F1-score of 93.13%, and a ROC-AUC of 0.9282. Stratified five-fold cross-validation further demonstrated the robustness of the proposed model, yielding a mean ROC-AUC of 0.8982 +/- 0.0130. Feature importance and SHAP analyses consistently identified TPSA, HBD, and LogP as the most influential molecular descriptors governing BBB permeability prediction. Overall, the proposed framework provides an accurate, interpretable, and computationally efficient approach for BBB permeability prediction and may serve as a valuable tool for the early-stage screening of CNS drug candidates.
[LG-25] A Sharp Barrier for Consistent Submodular Maximization: Any Improvement over 2-sqrt2 Entails Exponential Queries or Linear Recourse
链接: https://arxiv.org/abs/2609.09986
作者: Shi Fu,Qixin Zhang,Dacheng Tao
类目: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
*备注:
Abstract:Consistent submodular maximization studies the tradeoff between solution quality and stability when elements arrive over time. For a monotone submodular objective, which models diminishing returns, an algorithm maintains a set of at most k available elements and changes only O(1) elements after each insertion. Dütting et al. [2025] established a tight 2/3 approximation with unrestricted computation and a polynomial-time 0.51 approximation. They left open at STOC 2025 whether efficient algorithms can match the offline 1-1/e guarantee. We resolve this problem by proving that the supremum approximation achievable with polynomially many value queries and worst-case constant recourse is [ \beta=2-\sqrt2\approx0.58581-1/e. ] For every \varepsilon0 , our randomized algorithm attains \beta-\varepsilon with O(\varepsilon^-2) changes per insertion. Any fixed improvement requires exponentially many queries before one critical insertion or linear recourse of \Omega(k) changes at that insertion, even with unlimited queries afterwards. This gap quantifies the cost of consistency: the current oracle hides which elements will be needed after an arrival. We also determine the exact curvature-dependent threshold 1-(\sqrt2-1)\vartheta , attain 1-1/e-\varepsilon for weighted coverage with O(\varepsilon^-1) recourse, and separate the existence of universal future-price certificates from their efficient computation. Our algorithm has a bounded-bit polynomial-time implementation for polynomial-bit rational oracle answers; the lower bound uses only logarithmic-bit rational answers.
[LG-26] Deep Neural Networks for Learning Intent from sEMG Signals to Support Hardware Devices for Post-Stroke Neurorehabilitation
链接: https://arxiv.org/abs/2609.09971
作者: Zakariyya Brewster,Divy Wadhwani,Emily Yan,Aidan Wang,Karma Namgyal,Shuting Xie,Markiyan Konyk,Tala Abdelmaguid
类目: Machine Learning (cs.LG)
*备注:
Abstract:Finger-specific motor intent is a clinically meaningful control signal for post-stroke neurorehabilitation, where residual muscle activity may remain measurable despite weak or incomplete movement. We study five-finger multilabel intent decoding from impaired-arm high-density surface electromyography (sEMG) in PhysioMio, a bilateral longitudinal dataset collected from stroke patients. A common processing protocol aligns movement labels, applies 20–450 Hz Butterworth filtering and Symlet-4 wavelet denoising, segments overlapping 200 ms windows, and extracts twelve time- and frequency-domain descriptors per channel. Direct LSTM, CNN, and GNN baselines reveal complementary behavior: the LSTM attains the highest subset accuracy (0.545), whereas the GNN attains the highest macro F1 (0.706) and macro AUPRC (0.776). Architecture search then identifies CNN-Large as the strongest single-split CNN, with 0.593 subset accuracy and 0.714 macro F1, while CNN-Micro provides a compact architecture for embedded inference. To match a four-sensor hardware design, we retrain CNN-Micro using channels associated with ECRB, ECRL, FDS, and FDP and exclude the ground electrode from model input. Across five seeds, cross-channel knowledge distillation improves the four-channel student over direct training, reaching 0.5219 \pm 0.0114 subset accuracy, 0.7612 \pm 0.0038 finger accuracy, and 0.6095 \pm 0.0058 macro F1. The selected 123K-parameter model accepts nine windows of 48 features and has been exported to ONNX. These results establish a reproducible software path from post-stroke sEMG to compact five-finger intent prediction for subsequent hardware-in-the-loop evaluation.
[LG-27] Adversarial Training for Tabular Credit Scoring: A Multi-Attack Robustness Evaluation in P2P Lending
链接: https://arxiv.org/abs/2609.09945
作者: Gijs A. F. Niewzwaag,Marijn G. S. Veth,Manuele Massei,Marcos R. Machado
类目: Machine Learning (cs.LG); Risk Management (q-fin.RM); Machine Learning (stat.ML)
*备注:
Abstract:Machine learning-based credit scoring is increasingly central to Peer-to-Peer (P2P) lending, yet its resilience to adversarial manipulation, where applicants strategically alter self-reported inputs to secure favourable decisions, remains poorly understood. Most adversarial-robustness evidence comes from image and text domains and evaluates a single attack against a matching defence, offering little guidance on how defences generalise across attack types in tabular credit data. We address this with a systematic train-test robustness benchmark on a large Lending Club subset, spanning three model families (logistic regression, a feed-forward neural network, and a transformer for tabular data) and four attacks confined to applicant-mutable features: Fast Gradient Sign Method (FGSM), Projected Gradient Descent (PGD), Salt-and-Pepper (SP) noise, and DeepFool, plus a mixed-attack regime. Across a full grid evaluated with stratified cross-validation, adversarial training sharply improves robustness against the attack it is trained on and transfers well within the gradient-based family, but transfers weakly to non-gradient corruption, so single-attack defences overstate real-world resilience. Mixed training delivers the most balanced robustness across heterogeneous attacks while preserving clean-test performance, supporting multi-attack stress testing in credit-model governance.
[LG-28] Multi-Pass Multi-View Blended Learning for High-Fidelity Volumetric CT Synthesis from Chest X-Rays
链接: https://arxiv.org/abs/2609.09920
作者: Ozer Can Devecioglu,Serkan Kiranyaz,Rashid Mazhar,Tahir Hamid,Muhammad Chowdhury,Moncef Gabbouj
类目: Machine Learning (cs.LG)
*备注: 13 pages, 11 figures
Abstract:Reconstructing volumetric Computed Tomography (CT) from a single 2D chest radiograph (CXR) is an ill-posed inverse problem, further complicated by the scarcity of paired CXR-CT training data. Prior approaches address this by training on Digitally Reconstructed Radiographs (DRRs), which are synthetic projections derived from CT volumes. However, the domain gap between DRRs and real CXRs limits generalization, often resulting in coarse or anatomically inconsistent reconstructions when applied to clinical images. To address this challenging problem, this study introduces a Multi-Pass Multi-View Blended Learning framework for synthesizing high-fidelity volumetric CT directly from real chest X-ray (CXR) images. The proposed approach progressively decomposes the synthesis task into two distinct, complementary learning stages. Stage 1 is an unsupervised CXR-to-DRR Domain Adaptation, while Stage 2 includes three passes, namely, (a) supervised DRR-to-CT Transformation, (b) unsupervised Multi-View Slice Refinement, followed by © Progressive Transfer Learning (PTL). With such a blended learning paradigm, the proposed approach mitigates the synthetic-to-real domain gap while enhancing both the structural integrity and anatomical detail of the final output. On the LIDC-IDRI dataset, where paired DRR-CT ground truth is available for quantitative evaluation, the proposed method improves upon prior methods by up to 14% in PSNR and 7.6% in SSIM. The framework successfully generates structurally consistent and anatomically realistic high-fidelity CT volumes from real CXRs, marking a significant advancement toward clinical viability of CT reconstruction from standard radiographic images.
[LG-29] Development and Validation of a Physics-Guided Machine Learning Extrapolation Framework Using a Classical Transient Diffusion Benchmark
链接: https://arxiv.org/abs/2609.09912
作者: Ashutosh Yadav,Alok Dubey,Prodyut Ranjan Chakraborty,Harshal Akolekar
类目: Machine Learning (cs.LG)
*备注:
Abstract:Machine learning models used in engineering are typically trained within limited operating ranges, yet reliable predictions are often required beyond these domains. Consequently, the primary challenge is extrapolation rather than interpolation. Rigorous validation is hindered by the scarcity of data outside the training range. To address this limitation, a novel extrapolation framework is integrated with established machine learning architectures to enable accurate and physically consistent predictions beyond the training domain. The framework is established by systematically evaluating two physics-guided architectures: a Bidirectional Long Short-Term Memory (BiLSTM) network and a Physics-Informed Neural Network (PINN). A classical one-dimensional transient diffusion problem is adopted as a benchmark because its exact analytical solution provides unlimited, reliable data across the spatio-temporal domain, enabling rigorous quantitative validation. The problem is particularly challenging because the solution evolves from an initial singularity through a strongly nonlinear transient regime before approaching a steady-state linear profile. When training data are confined to an intermediate portion of this evolution, backward extrapolation toward the singularity becomes especially demanding. To improve reliability, physics-guided coordinate transformations, boundary-aware learning strategies, and stability-enhancing temporal marching are incorporated. Extrapolation is evaluated using a train-predict-validate-extend strategy, in which validated predictions are recursively added to the training set to progressively extend the prediction horizon. The results demonstrate accurate and physically consistent predictions beyond the training domain, highlighting the framework’s potential for engineering applications where data availability is limited.
[LG-30] A Kernel-Based Modular Discriminant Analysis Framework for Small-Sample Learning
链接: https://arxiv.org/abs/2609.09910
作者: Lingxiao Qu,Yan Pei
类目: Machine Learning (cs.LG)
*备注:
Abstract:The small-sample-size (SSS) problem remains a fundamental challenge in machine learning when labeled data are scarce due to cost, accessibility, or ethical constraints. While numerous approaches have been proposed, existing methods often struggle to maintain stable and discriminative representations under high-dimensional and limited-data conditions. Kernelized Linear Principal Component Discriminant Analysis (KLPCDA), a recently proposed modular framework, integrates variance preservation, inter-class separability, and intra-class compactness within a unified kernel space. Although its formulation has shown promising initial results, a systematic understanding of how its components interact across diverse SSS scenarios remains lacking. In this paper, we present a systematic cross-domain study of KLPCDA to characterize the interaction mechanisms among its core objectives. We analyze the behavior of its seven variants across multiple real-world SSS tasks, including hyperspectral image classification, mechanical fault diagnosis, medical diagnosis, and face recognition. Through extensive experiments and ablation studies, we investigate how different objective combinations influence performance under varying conditions such as noise, class imbalance, and high dimensionality. Our analysis reveals consistent patterns in the interaction of the three core objectives variance, between-class, and within-class terms, providing a unified and interpretable understanding of their roles in stabilizing representations and enhancing discrimination in SSS settings. Based on these findings, we further derive practical guidelines for selecting appropriate KLPCDA variants under different data characteristics. Experimental results demonstrate that KLPCDA achieves strong and robust performance across domains, while maintaining low computational complexity suitable for resource-constrained environments.
[LG-31] Meta-LinEXP3: Online-within-Online Learning for Adversarial Linear Contextual Bandits
链接: https://arxiv.org/abs/2609.09907
作者: Hao Li,Jie Xu,Zheng Xie
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Meta-learning has emerged as an effective paradigm for transferring knowledge across sequential bandit tasks. While substantial progress has been made for stochastic bandits and non-contextual adversarial bandits, meta-learning for adversarial linear contextual bandits (ALCBs) with random action sets remains largely unexplored. To address this problem, we propose Meta-LinEXP3, an online-within-online algorithm that constructs a predictable task-level prior from completed tasks to guide the inner LinEXP3 learner. For known context distributions, we develop a policy-centered estimator that achieves an intrinsic-dimension \mathcalO(\sqrtn) per-task regret bound. For unknown distributions, we introduce a past-only regularized moment estimator with an \mathcalO(n^2/3) leading regret term and explicit finite-sample error. We further establish a direct connection between prior accuracy and transfer regret, showing that increasingly accurate priors yield sublinear transfer-dependent regret across tasks. Experiments demonstrate the effectiveness of Meta-LinEXP3, including its application to structured hyperspectral tensor sampling.
[LG-32] Beyond Conventional Federated Learning via High-Order Regularization
链接: https://arxiv.org/abs/2609.09904
作者: Alireza Kabgani,Masoud Ahookhosh
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Federated clients that perform several local optimization steps can return parameter displacements with widely different magnitudes. The quadratic regularization of FedProx grows linearly with displacement and therefore offers limited control over the contrast between ordinary and unusually large client movements. We here introduce HiFedProx, which replaces the quadratic penalty with a scale-matched power-type regularizer indexed by p\geq2 . All powers have the same regularization-gradient magnitude at a reference displacement R , while every p2 gives a weaker response below R and a stronger response above it. An exact affine reference calculation shows that increasing p compresses relative displacement disparities, although very large powers approach fixed-radius behavior and increase local curvature. HiFedProx combines this geometry with finite-budget stochastic client optimization and same-minibatch Armijo backtracking. In paired five-seed experiments on a frozen 60-writer FEMNIST subset, a common-parameter study over p\in\2,3,4,5,6,7,8\ shows similar clean-training performance but substantial gains under composite stress. The lowest moderate- and severe-stress losses occur at p=7 and p=6 , improving over p=2 by 11.44% and 23.16% , respectively. Although displacement-tail ratios continue to decrease through p=8 , predictive performance peaks in an intermediate range and Armijo trial cost increases with p . These results indicate that the exponent should be calibrated rather than maximized. In our experiments, p=5 – 7 provides the most useful range.
[LG-33] ProMeta: Few-shot PROTAC-targeted degradation prediction across E3 ligases
链接: https://arxiv.org/abs/2609.09891
作者: Yuansheng Liu,Yufei Ye,Tao Tang,Jiawei Luo,Wen Tao,Xiao Luo
类目: Machine Learning (cs.LG); Quantitative Methods (q-bio.QM)
*备注: 13 pages, 4 figures, and 2 tables. Source code and reproducibility resources are available at this https URL and this https URL
Abstract:Proteolysis-targeting chimeras (PROTACs) have emerged as a transformative therapeutic strategy that selectively degrades historically ‘‘undruggable’’ targets via the ubiquitin-proteasome system. Despite growing efforts to develop computational predictors of PROTAC degradation activity, existing supervised approaches remain severely challenged by data scarcity and imbalance across E3 ligases, limiting their ability to generalize beyond well-studied ligase contexts. In practice, labeled data are heavily concentrated on a few ligases (e.g., CRBN and VHL), while the majority of E3 ligases remain underexplored yet are critical for expanding the design space of targeted degraders. Developing methods that enable robust cross-ligase generalization with minimal labeled data is therefore essential for improving the practical utility of computational PROTAC discovery. We reformulate PROTAC degradation activity prediction across E3 ligases as a few-shot meta-learning problem and present ProMeta, a prototype-based graph neural network trained through episodic meta-learning on source-E3 tasks and evaluated on held-out target-E3 tasks through support-conditioned inference. ProMeta performs inference without updating the encoder by dynamically estimating class prototypes from minimal target-ligase support samples. On the CRBN-to-VHL benchmark, ProMeta achieves AUROC values of 0.796 under K=2, Q=3 and 0.883 under K=2, Q=5, improving by 19.9% and 6.8%, respectively, over the corresponding supervised GNN baseline. Reverse VHL-to-CRBN transfer under the same protocol yielded AUROC values of 0.702 (K=2, Q=3) and 0.821 (K=2, Q=5), confirming bidirectional applicability while revealing direction and data-regime dependence. Together, these results support ProMeta as a practical framework for cross-ligase few-shot prediction under the evaluated support/query protocols.
[LG-34] Exact Degeneracy Under Balanced k-Shot Sampling:Consequences for Small-Sample Discriminant Analysis on LLM Embeddings
链接: https://arxiv.org/abs/2609.09860
作者: Lingxiao Qu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Balanced k-shot sampling draws exactly k labeled examples per class. We show that it induces an exact, provable degeneracy in a family of small-sample discriminant estimators. Under balanced sampling, the within-class scatter operator of Kernelized Linear Principal Component Discriminant Analysis (KLPCDA) is not merely rank-deficient but exactly a scaled orthogonal projector. We derive the consequences in closed form: two of KLPCDA’s seven variants have every signal eigenvalue exactly equal, so their eigenvector selection criterion is provably indifferent rather than ill-conditioned, and a third has a provably void objective. This follows from the estimators’ construction, not any dataset; we confirm it on frozen sentence embeddings and, separately, on residual-stream activations from a decoder-only generative model. An in-formula tie-break repairs the two repairable variants, with recovery gated by class count: the residual subspace constraint costs 5x more on few-class than many-class datasets (p=0.000001). We then evaluate the repaired framework on few-shot text classification on frozen LLM embeddings (n much smaller than d, up to 4096), across four datasets, three embedding sizes, and three trained baselines (SetFit, LoRA, in-context learning). A properly cross-validated logistic-regression probe still beats every KLPCDA variant on three of four datasets, at every embedding size; guidance carried from pixel, vibration-signal, and gene-expression data does not directly generalize to this feature space. Three independent geometric separability metrics fail to explain why one high-dimensional decoder-based embedding model underperforms smaller bidirectional encoders, ruling out anisotropy; the gap is substantially an estimation-efficiency effect, not a permanent ceiling, closing by more than 80% when the support set grows from k=10 to k=30-50 (p=0.00195, both many-class datasets). Subjects: Machine Learning (cs.LG) Cite as: arXiv:2609.09860 [cs.LG] (or arXiv:2609.09860v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.09860 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-35] mpTPI: Informer-Based trajectory prediction for maritime vessels
链接: https://arxiv.org/abs/2609.09840
作者: Kevin Ferneding,Veronika Lietavcova,Aleksandra M. Blachowiak,Peder Heiselberg
类目: Machine Learning (cs.LG)
*备注: Accepted to IEEE International Geoscience and Remote Sensing Symposium (IGARSS) 2026. \c{opyright} 2026 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media
Abstract:Accurate long-term trajectory prediction for maritime vessels is essential for safety and logistical efficiency. While deep learning models, particularly Transformers, have shown promise in processing Automatic Identification System (AIS) data, they often struggle with the quadratic computational complexity of self-attention and the loss of accuracy over extended forecasting horizons. This study proposes TempTPI, a novel prediction framework that integrates an Informer-based encoder with a multi-channel temporal encoding mechanism. The Informer architecture leverages a ProbSparse self-attention mechanism to reduce computational overhead and focus on the most significant dependencies, while the temporal encoder utilizes Fourier-like frequency expansions to capture cyclic patterns (hourly, daily, and seasonal) in vessel behavior. We evaluate our model against the state-of-the-art TPTrans architecture using AIS data from Danish waters. Experimental results demonstrate that TempTPI consistently outperforms existing methods across prediction windows of 1 to 5 hours. Notably, at a 5-hour horizon, the proposed model achieves a 55% improvement in Mean Squared Error (MSE), offering a robust solution for long-range maritime situational awareness.
[LG-36] In Medical Claims Data Enhancing Predictive Performance for Major Adverse Cardiovascular Events Using Cross Attention KDD2024
链接: https://arxiv.org/abs/2609.09824
作者: Yuhei Fujioka,Daitaro Misawa,Tatsuyoshi Ikenoue,Shingo Fukuma
类目: Machine Learning (cs.LG)
*备注: 13pages, 3 figures. Accepted to KDD 2024 AIDSH workshop
Abstract:Medical claims data comprise the financial details, including the expenses and billing information, as well as the clinical information, such as the diagnoses and treatments, of patients visiting medical facilities. Recently, it has been acknowledged that large databases can be constructed from medical claims data for medical research purposes. However, the clinical information within these datasets is often medically unstructured, limiting its application in comprehensive analyses. This study enhances predictive model performance for major adverse cardiovascular events (MACE), a leading cause of death worldwide. Models that predict MACE are crucial to clinical practice guidelines. We utilize a cross-attention mechanism to develop a method that effectively weights the relationships between diagnoses and treatments. Effectively repre- senting the clinical information contained in medical claims data, this approach generates more representative features for predicting MACE. The ROC-AUC score of our proposed cross-attention-based model was 0.7720, higher than other benchmark models including the conventional atherosclerotic cardiovascular disease model, the light gradient boosting machine, and a self-attention-based model. These results indicate that integrating the clinical structure of medical claims data using a cross-attention mechanism significantly enhances the performance of predictive models.
[LG-37] Online Inverse Integer Linear Optimization via Small-Gradient Skipping: Constant Regret and Finite Mistakes
链接: https://arxiv.org/abs/2609.09809
作者: Akira Kitaoka
类目: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS); Optimization and Control (math.OC)
*备注: 56 pages
Abstract:In online inverse linear optimization, the learner predicts a weight at each round, observes the optimal action of the agent, and updates its prediction. In the general setting, the gap of \log T between the regret upper bound O(d \log T) and the lower bound \Omega(d) is unresolved (here T is the total number of rounds and d is the dimension). When the action set is M-convex, the regret is known to be bounded by O(d \log d) , but the method attaining it computes a center of gravity at every round. This paper therefore proposes Small-Gradient Skipping (SGS), a mechanism that skips the update at rounds without a mistake in the case where the correct action is uniformly separated from the other candidates, and applies it to online gradient descent, the online Newton step, and MetaGrad. The number of mistakes is then bounded, for all three, by a quantity independent of T ; and for the online Newton step and for MetaGrad with SGS, the dimension dependence of the regret becomes O(d^2) when the forward problem is an integer linear program, that is, the factor \log T is removed. Moreover, when the action set is M-convex, the regret is bounded efficiently without computing a center of gravity.
[LG-38] A practical DIRECT-type algorithm for medium-scale black-box global optimization
链接: https://arxiv.org/abs/2609.09796
作者: Linas Stripinis,Remigijus Paulavičius
类目: Machine Learning (cs.LG); Numerical Analysis (math.NA)
*备注: 46pages, 13 figures, 4 tables
Abstract:The DIRECT algorithm is a deterministic global optimization method known for its versatility and balanced exploration-exploitation strategy. However, DIRECT-type algorithms are primarily effective for low-dimensional problems and often exhibit slow convergence as dimensionality increases, limiting their applicability to more complex optimization tasks. To address this limitation, this paper introduces X-DTC-GL, a novel DIRECT-type algorithm that incorporates dynamic partitioning and hybridization techniques. The dynamic partitioning approach adaptively refines the search space based on local one-dimensional surrogate models, enabling rapid subdivision of promising hyper-rectangles. The hybridization strategy selectively employs a hill-climbing method to exploit promising regions identified by the surrogate models. Extensive experiments on four diverse benchmark suites demonstrate that X-DTC-GL significantly outperforms existing DIRECT-type baselines, achieving improvements of ~12% in solvability and ~27% in solution quality. Performance-profile analyses indicate the fastest convergence on up to ~40% of instances, the best runtime performance on ~17% of problems, and competitive overall execution times. By improving performance within the partition-based framework, these advances strengthen the algorithm’s competitiveness in state-of-the-art black-box optimization.
[LG-39] Privacy-Preserving Split Learning for Federated LLM Fine-Tuning
链接: https://arxiv.org/abs/2609.09794
作者: Heng Jin,Chaoyu Zhang,Hexuan Yu,Wenjing Lou,Y. Thomas Hou
类目: Machine Learning (cs.LG)
*备注:
Abstract:Fine-tuning large language models (LLMs) on domain-specific data is essential for downstream adaptation. In many deployments, a participant cannot hold the complete model locally. This happens because the model owner keeps the full model proprietary, or because the participant lacks sufficient compute resources. Split Learning (SL) addresses this by partitioning the model between the participant and a server so that only a small portion runs locally. When the underlying data is additionally distributed across multiple institutions with privacy requirements, Federated Learning (FL) further enables collaborative training across participants by sharing only model updates instead of raw data. In this combined setting, each client transmits intermediate activations to the server, and for LLM fine-tuning, this exchange poses an inherent privacy paradox. The autoregressive nature of LLMs causes the transmitted activations to leak the input, and existing perturbation-based defenses are fundamentally ineffective in this setting. We address this leakage through a learned obfuscate-and-recover scheme that protects participants’ private datasets while still allowing an independently deployable model to be trained on the server side. Experiments demonstrate that our approach achieves strong privacy protection with modest utility loss and system overhead, making split-based federated LLM fine-tuning practically viable.
[LG-40] Evaluating Model Retraining under Drift: Paired Comparisons of Cumulative Subgroup Disparity
链接: https://arxiv.org/abs/2609.09788
作者: Aaron Ceross
类目: Machine Learning (cs.LG)
*备注: 44 pages, 10 figures, 34 tables. Reproducibility package version 2026.09.08-r2
Abstract:Choosing when to retrain a deployed classifier requires assessing subgroup error rates across the sequence of models used, including periods between updates. We compare complete scheduled, loss-triggered, and subgroup-gap-triggered policies with retaining the initial model on the same observations and delayed labels. For true-positive and false-positive rates separately, the outcome is the paired difference in absolute subgroup gaps summed over deployment windows. Population evaluation in simulation, action records, and alternative schedules assess how measurement and retraining behaviour affect these comparisons. In a follow-up sample of 400 new trajectories per condition across two simulated drift regimes, all three policies had lower mean cumulative disparity, equivalent to reductions of 0.04 to 0.88 percentage points in the average gap per window. Evaluating the unchanged models against the known generating distributions preserved all mean directions, but finite-window and population comparisons agreed on whether updating increased, reduced or left cumulative disparity unchanged in 69 to 92 percent of trajectories. Under subgroup-specific drift, smaller true-positive-rate gaps accompanied lower sensitivity in both groups. In an exploratory American Community Survey replay, person weighting reversed all three race false-positive-rate mean comparisons without changing predictions or actions; all three weighted intervals included zero. Policy comparisons require group-specific rates, action distributions, and an explicit evaluation population alongside mean disparity. These analyses are non-confirmatory. Shared replay requires policy-independent observations and complete labels after the specified delay.
[LG-41] NEXUS-MI: Communication-Aware Federated Personalization for Gateway-Coordinated Motor-Imagery Brain-Computer Interfaces
链接: https://arxiv.org/abs/2609.09786
作者: Daniel Adu Worae,Aarthy Nagarajan
类目: Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注:
Abstract:Electroencephalography (EEG)-based motor-imagery brain-computer interfaces (MI-BCIs) vary across subjects and sessions, complicating personalization from limited calibration data. Federated learning can exploit shared representations without centralizing raw EEG, but existing federated MI studies largely assume regular synchronization. We introduce NEXUS-MI, a gateway-coordinated federated personalization framework that treats synchronization as a coupled learning-and-communication control problem. Raw EEG and classifier heads remain local, while an edge coordinator maintains the shared backbone. We evaluate NEXUS-MI through offline replay using BCI Competition IV Dataset 2a (BCICIV-2a; 9 subjects, 4 classes) and OpenBMI (54 subjects, 2 classes). Session 1 supports backbone learning, and Session 2 provides limited-calibration personalization and held-out testing. An ideal-link reference and six heterogeneous-link policies characterize gateway participation, buffering, stale-update admission, and backbone-download control. The principal comparison holds delayed-update handling fixed while contrasting non-adaptive and communication-aware synchronization. Paired subject-level comparisons use Holm adjustment, and robustness across five matched realizations is assessed by hierarchical bootstrap. Communication-aware coordination reduced server-to-client backbone traffic by approximately 42% on both datasets, while cohort-level accuracy differences were small and realization-dependent. Cohort averages also concealed subject-level vulnerability, with losses reaching approximately 12 percentage points on BCICIV-2a relative to the ideal-link reference. These findings establish gateway synchronization as an explicit design variable in federated MI personalization and motivate joint evaluation of personalized accuracy, communication cost, update freshness, and subject-level reliability.
[LG-42] EEGBind: Detecting Source-Level Interictal Epileptiform Discharges via EEG-Centric Multimodal Binding
链接: https://arxiv.org/abs/2609.09728
作者: Muchen Li,Anglin Liu,Xuetian Gao,Ruijian Xu,Jintai Chen
类目: Machine Learning (cs.LG); Multimedia (cs.MM); Neurons and Cognition (q-bio.NC)
*备注: 7 pages, 5 figures. Accepted to the 34th ACM International Conference on Multimedia (MM '26)
Abstract:Source-level analysis of interictal epileptiform discharges (IEDs) is relevant to presurgical evaluation and treatment planning because it helps characterize where epileptiform activity is likely to arise. Beyond detecting whether an IED is present, this setting requires assigning IED-positive activity to clinically meaningful brain-region categories. This setting is challenging because source-region evidence in short electroencephalography (EEG) windows can be subtle, partial, and affected by subject variability, class imbalance, and imperfect multimodal context. We present EEGBind, an EEG-centric multimodal binding framework for five-class source-level IED classification. EEGBind treats EEG as the primary modality and binds synchronized video-context features around an EEG-centric representation. Instead of relying on early or overly strong multimodal fusion, which may perturb the source-sensitive EEG representation, EEGBind uses video context as auxiliary evidence for robust classification. A view-consistent repair stage is further used to improve hidden-set robustness while preserving the learned source-class boundary. On the NeuroMM 2026 Grand Challenge Track 3 NMM-Source-IED benchmark, EEGBind achieves 0.8395 on weighted-F1 and outperforms strong competitors. These results support EEG-centric multimodal binding as a practical strategy for source-level IED classification. The open-source code is available at this https URL.
[LG-43] EFQ-Softmax: Exp-Free Quantization for Softmax
链接: https://arxiv.org/abs/2609.09721
作者: Haohui Han(1),Yuming Wan(2),Hongni Wang(3),Pengcheng Xie(2),Xiaodong Yan(1),Runqi You(1),Wencong Zhang(1) ((1) Xi’an Jiaotong University, (2) Huawei Technologies Co., Ltd, (3) Shandong University of Finance and Economics)
类目: Machine Learning (cs.LG)
*备注: 12 pages, 7 figures
Abstract:Low-bit attention accelerates Transformer inference by moving the QK^\top and PV matrix multiplications to FP8 or FP4 matrix engines. However, the softmax path often evaluates shifted-score exponentials in higher precision, forms a temporary probability block, and quantizes it before low-bit PV multiplication. This exp-then-quantize path creates a mismatch between a high-precision probability producer and a low-bit matrix consumer. We propose EFQ-Softmax (Exp-Free Quantization for Softmax), a low-bit probability-generation method that directly maps shifted attention scores to block-scaled E2M1 operands. For each microscaling block, EFQ-Softmax selects an exponent-only scale from the local maximum, maps the shifted scores to a normalized residual domain, and generates nonnegative E2M1 probability codes using a single affine rule. The resulting operand is used consistently in both the \widetildePV numerator update and the \widetildeP\mathbf1 denominator update. The FlashAttention-style row-maximum update, historical rescaling, high-precision accumulation, and final normalization remain unchanged. We evaluate end-to-end quality on Qwen3-8B, Qwen3-VL-8B-Instruct, and WAN2.2-TI2V-5B, and separately measure kernel-level performance on the A5 vector unit. EFQ-Softmax improves the Qwen3-8B seven-task mean from 0.6749 with MXFP4 to 0.6773 and the Qwen3-VL nine-task mean from 0.7826 to 0.8000. On WAN2.2, it maintains temporal consistency and visual quality comparable to the FP16 and MXFP4 baselines under VBench. On the A5 vector unit, EFQ-Softmax reduces the vector-stage latency of the fused probability-generation kernel by 40.33% on average across sequence lengths from 16K to 128K. These results show that direct low-bit probability generation can replace the conventional exp-then-quantize path while preserving end-to-end model quality.
[LG-44] ALIGN-HOLD: Experience Alignment for Real-Time Hold Control in Large-Scale Ride-Hailing Matching at DiDi
链接: https://arxiv.org/abs/2609.09685
作者: Zuhao Zhang,Xu Liu,Kai Wan,Zihao Lu,Li Ma,Shuai Li
类目: Machine Learning (cs.LG)
*备注:
Abstract:Real-time hold control is a high-leverage mechanism in large-scale ride-hailing systems: by selectively deferring driver-order pairs, the platform can wait for better matching opportunities and improve end-to-end passenger-driver experience. Existing production systems such as EXHOLD learn bandit-based hold policies from handcrafted combinations of trip completion, cancellations, waiting time, and driver effort. However, designing such rewards becomes increasingly difficult as marketplace preferences are heterogeneous and observed passenger-driver behavior can be sparse, noisy, and affected by dynamic supply-demand conditions. We present ALIGN-HOLD, a production-scale experience alignment framework that learns hold policy from implicit marketplace preferences. ALIGN-HOLD constructs complementary preference pairs from order trajectories, driver trajectories, and contemporaneous local matching graphs, and trains an experience Reward Model (RM) using balanced multi-view sampling and model-adaptive hard preference sampling. During simulator-based policy learning, the frozen RM provides a dense, context-dependent reward and supports label-free filtering of low-identifiability interactions whose behavioral feedback is difficult to attribute to matching quality. We deploy ALIGN-HOLD on DiDi’s ride-hailing platform and evaluate it in a 28-day randomized A/B experiment, covering approximately 100,000 passenger requests per day. Compared with the deployed production policy, ALIGN-HOLD achieves statistically significant improvements in trip completion rate and driver income, while significantly reducing passenger cancellations before and after driver acceptance. Complementary ablations, RM diagnostics, and behavioral analyses validate the contributions of the proposed components. ALIGN-HOLD has been fully ramped up and is currently serving DiDi’s Brazil marketplace. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2609.09685 [cs.LG] (or arXiv:2609.09685v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.09685 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-45] Settling: Equilibrium Inference for Non-Convex Validity Sets
链接: https://arxiv.org/abs/2609.09682
作者: Lyes Saad Saoud
类目: Machine Learning (cs.LG)
*备注:
Abstract:Many learning systems return a single point estimate even when admissible outputs form disconnected or non-convex sets. Under squared loss, an ambiguous conditional distribution can therefore have a Bayes-optimal conditional mean that is invalid. We formalize this failure as conditional mean collapse and introduce Settling, an equilibrium-based inference operator that separates proposal generation, consistency evaluation, and test-time equilibrium selection. The operator treats a mean-seeking proposal as an initialization and refines it toward a locally stable configuration; conditional on initialization, refinement is deterministic. We establish exact-gradient descent, local convergence, and an inexact-gradient robustness condition relevant to learned consistency critics. In a reproducible 100-context geometric diagnostic, the mean-seeking baseline succeeds in 0/100 contexts, stochastic denoising in 100/100, and Settling in 99/100 while producing substantially lower trajectory roughness. A 1,200-run sensitivity study yields 97-100% success across obstacle-jitter ranges up to 0.20 and 94-100% across one-time initialization perturbations from 0.05 to 0.50. Cross-domain panels remain mechanism illustrations; learned high-dimensional validation remains an open empirical test.
[LG-46] Muon-C: Operator-Aligned Muon for Convolutional Kernels
链接: https://arxiv.org/abs/2609.09676
作者: Jiaxin Qing,Lexin Li
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Muon replaces matrix momentum with an approximately orthogonal polar direction, but its geometry depends on the matrix representation. For convolution, standard unfolding describes a local patch map rather than the convolution operator. We introduce Muon-C, an operator-aligned optimizer that represents kernel momentum as frequency-wise channel-transfer matrices, polarizes these blocks independently, and uses a critical Fourier grid to return updates exactly to the original finite kernel support. We show that the new geometry arises from combining the block partition and Fourier coordinates. The exact-polar direction is a linear minimization oracle under the critically sampled convolution norm. Its worst-case guarantee relative to the continuous convolution-operator norm is never weaker than unfolding and is strictly stronger for 3\times3 kernels. On CIFAR-10 flow matching with matched applied-update RMS, Muon-C reaches 9.87 FID at 40k iterations, compared with 22.26 for unfolded Muon and 51.31 for Adam. It reaches their final quality using 0.62\times and 0.64\times their model FLOPs, respectively. Under equal tuning budgets, Muon-C achieves 3.42 FID. Gains persist across data scales and transfer to classification across convolutional architectures.
[LG-47] Geometric organization of olfactory descriptor data in the Poincaré disk
链接: https://arxiv.org/abs/2609.09573
作者: Aniss Aiman Medbouhi,Farzaneh Taleb,Giovanni Luca Marchetti,Danica Kragic
类目: Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Applications (stat.AP)
*备注: Submitted to Chemical Senses
Abstract:Odor quality is commonly represented using high dimensional descriptor profiles, yet their low dimensional organization remains unclear. We investigated whether a two-dimensional hyperbolic embedding can provide an interpretable representation of this structure. We applied hyperbolic metric multidimensional scaling to two complementary datasets: 480 Sagar rating profiles from three participants rating 160 odorants on 15 continuous descriptors, and 4983 GoodScents–Leffingwell molecules annotated with 138 binary descriptors. The embeddings substantially preserved pairwise descriptor distances, supporting subsequent analyses of radial and angular organization. In Sagar, rating profile entropy was strongly and negatively associated with hyperbolic radius, with diffuse profiles closer to the center and concentrated profiles closer to the boundary. This radial organization emerged primarily at the level of the full descriptor profile, rather than any individual descriptor, and remained robust across alternative descriptor representations, participant specific analyses, and averaged ratings. Sweet, musky, fruity, pleasantness showed the strongest directional trends. In GoodScents–Leffingwell, active label entropy, reflecting descriptor multiplicity, increased with radius, whereas orthogonalized descriptor entropy, reflecting spread across orthogonal modes, decreased with radius. Related binary descriptors occupied coherent localized high-density regions. These findings reveal complementary radial and angular organization in the hyperbolic representation of olfactory descriptor data. They support hyperbolic mapping as an interpretable descriptive framework in which radius summarizes global profile properties, while the angular component captures continuous descriptor gradients and categorical organization.
[LG-48] Positional task conditioning for scalable defect detection across product families in large product catalogs
链接: https://arxiv.org/abs/2609.09567
作者: Soham Satyadharma,Gabriel Roccabruna,Suleiman A. Khan
类目: Machine Learning (cs.LG)
*备注:
Abstract:Product families in large product catalogs suffer from inconsistencies such as duplicates and unit mismatches that degrade customer experience. Detecting these requires reasoning over multiple error types across lengthy product listings, where LLM classification quality degrades due to long-context limitations. We address this by decomposing detection into focused sub-tasks that reduce context and isolate error types, improving F1 from 52% to 87%. For scalable deployment, we introduce Positional Task Conditioning (PTC), which distills this capability into a single smaller model by reinforcing task identity at structural prompt boundaries. PTC outperforms rationale-based distillation across five models and two architecture families, achieving within 1.79% F1 of the frontier at upto 98% lower cost. Our system is deployed across multiple countries processing 10+ million product families.
[LG-49] Robust Industrial Cyber Physical Classification Using Neuromorphic Temporal Embeddings and Hybrid SNN XGBoost Under Machine Unlearning Attacks
链接: https://arxiv.org/abs/2609.09564
作者: Ammar Kamoona,Sajad Koushkbaghi,Mahdi Jalili,Peter McTaggart,Xinghuo Yu
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Neural and Evolutionary Computing (cs.NE)
*备注: 13 pages, first draft
Abstract:The digitalisation of electrical distribution networks has increased the exposure of power-grid infrastructure to cyber attacks. Existing intrusion detection systems (IDSs), however, often rely on computationally expensive deep learning models that are difficult to deploy at the edge. Periodic retraining also exposes these systems to machine unlearning attacks, where selective data removal can degrade detection performance. We propose a hybrid Spiking Neural Network (SNN) and XGBoost architecture that combines efficient temporal encoding with a lightweight classifier and provides structural resilience to such attacks. The SNN is trained once on clean data and used as a fixed feature extractor, while only the XGBoost classifier is retrained during model updates. Evaluated on two real-world public power-system datasets, the proposed method achieves 99.9% accuracy (F1-macro 0.999) on the Synchrophasor dataset and 95.0% accuracy (F1-macro 0.943) on the MSU/ORNL dataset, outperforming standalone baselines. Under selective label-flipping attacks, the hybrid model loses only 0.9% F1-macro at 10% poisoning and delays target-class collapse from 60% to 70% poisoning compared with raw models. These results demonstrate that neuromorphic temporal encoding can provide both accurate cyber-attack detection and improved resilience to data poisoning in cyber-physical systems.
[LG-50] Unthrottling the Tanh Jacobian in SAC: A Negative Result on Bang-Bang Control and MetaDrive
链接: https://arxiv.org/abs/2609.09478
作者: Faiq Shamass
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注: 8 pages, 2 figures. Technical report. Negative result. Code: this https URL
Abstract:Soft Actor-Critic (SAC) represents a continuous policy as an unbounded Gaussian that is squashed by tanh. The Jacobian of that map is \partial a/\partial u = 1-a^2 , which vanishes as |a|\to 1 . A natural concern is that this throttle starves the actor of critic signal exactly where extreme actions (full brake, full throttle) are optimal. We test a minimal intervention that restores the missing signal: one extra term in the actor loss whose gradient on the pre-tanh mean is the detached action-gradient of Q , with no gain parameter. On a minimum-time double integrator whose optimum is bang-bang at the action bounds, vanilla SAC already reaches near-optimal return ( -31.6 vs. a calibrated optimum of -30.3 ) across ten paired seeds. An ungated bypass does saturate the policy (99% of eval steps with |a|\ge 0.9 ) and collapses return to -195.5 . A gated bypass that fires only on the flat shoulder |a|\in[0.9,0.999] also fails, and does so without leaving a saturated policy. Warm-started MetaDrive fine-tuning shows the same pattern: the bypass does not improve return, and where collision rate falls it is typically traded for out-of-road departures. Auto-tuned entropy coefficient rises against the bypass, which is a push toward the tails. The Jacobian effect is real. Treating it as a bug to be undone is not free, and on the tasks studied here it is not helpful. Saturating a bound is not the same as solving a problem whose optimum lives on that bound. Comments: 8 pages, 2 figures. Technical report. Negative result. Code: this https URL Subjects: Machine Learning (cs.LG); Robotics (cs.RO) Cite as: arXiv:2609.09478 [cs.LG] (or arXiv:2609.09478v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.09478 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-51] Building the Harness Automatically: Self-Play in Code Distills a Text Harness for Black-Box Optimization
链接: https://arxiv.org/abs/2609.09468
作者: Yi Wu,Zheng Ren,Zhiyu Hu,Haochen Wang,Daryl Chang,Li Wei,Ting Wang,Zhen Li,Pooja Gupta,Nitin Jindal,Lukasz Heldt
类目: Machine Learning (cs.LG)
*备注:
Abstract:Can an agent learn a numerical search strategy through executable practice and then transfer that strategy as text? We study low-budget black-box optimization, where unaided language models remain well below strong classical optimizers. During development, an agent repeatedly writes and evaluates optimizer programs. It then distills the resulting program and practice record once into a 197-word primary Harness A, which is frozen before evaluation. Harness A reduces Gemini Flash regret by 48% in an independent N=30 study ( p.001 ), enters the GP-BO performance range on the practice family, and lowers mean regret on all three held-out BBOB landscapes. The same text improves every tested Gemini executor and transfers to Claude Sonnet, reducing regret by 43% and 49% ( p\leq.005 ). An independent end-to-end replication produces Harness B, a different program and text at the same performance tier. The same framework also attains the lowest regret on a sealed YouTube reward-tuning production benchmark. Executable practice is thus a viable way to discover a search policy, and language a portable medium for deploying it.
[LG-52] Exact-Form Regret for Gradient Descent Mirror Descent and Follow-the-Regularized-Leader
链接: https://arxiv.org/abs/2609.09466
作者: Ashkan Soleymani,Gabriele Farina,Patrick Jaillet
类目: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT); Optimization and Control (math.OC)
*备注:
Abstract:Online gradient descent is usually studied through external regret, where the learner competes with fixed alternatives. Recent work shows that first-order methods control richer action-dependent deviations. We ask for a geometric characterization of the deviations with respect to which online gradient descent, mirror descent, and follow-the-regularized-leader (FTRL) achieve no regret. We identify exactness as the common principle. Exactness means that the relevant displacement field is generated by a scalar potential, or equivalently that the associated one-form is exact in the geometry used by the algorithm. This geometry depends on the algorithm. For gradient descent it is Euclidean geometry, for mirror descent it is the geometry induced by the regularizer, and for FTRL it is the cumulative dual state. Under mild regularity conditions, exactness yields sublinear regret, while nonzero circulation provides the complementary obstruction and leads to linear regret. This gives a unified geometric framework for understanding the deviation classes controlled by these algorithms and reveals that different first-order methods can control genuinely different classes of deviations. These deviation classes have direct consequences for learning, particularly in games. We study the equilibrium notions induced by exact-form deviations and introduce conservative correlated equilibrium, reflecting both the conservative geometry of the underlying displacement fields and the restricted family of deviations available to the players. We characterize its relation to correlated equilibrium, determine when the resulting equilibrium notions coincide and when they separate, and show how these relationships depend on the geometry and the learning algorithm. Overall, this work gives a unified geometric account of what first-order online learning algorithms are no-regret with respect to, beyond fixed comparators.
[LG-53] Uncertainty-Aware Sea-Ice Type Mapping with Multiple Ice Charts
链接: https://arxiv.org/abs/2609.09451
作者: Samira Alkaee Taleghan,Younghyun Koo,Andrew P. Barrett,Farnoush Banaei-Kashani
类目: Machine Learning (cs.LG)
*备注:
Abstract:Sea-ice stage of development (SoD) describes the age and associated thickness of sea ice and provides important information for navigation, and operational ice monitoring. SoD labels are obtained from operational ice charts, where trained analysts interpret satellite observations and assign standardized stage codes to regions with similar ice conditions. These codes often represent ranges of compatible ice thicknesses rather than exact physical values. Deep-learning methods can automate SoD mapping and commonly adopt operational ice charts as reference labels for training. These annotations are not exact, however; this is because chart interpretation relies on analyst judgement and on the observations available at the time, so different ice services may assign different SoD labels to the same conditions. We term this variation across independently produced expert annotations multi-annotator label uncertainty; collapsing the annotations into a single deterministic target discards this variation. A second source of uncertainty originates in the learned model itself. In this paper, we quantify both sources: annotation uncertainty from disagreement among independent ice-service charts and model uncertainty from the learned predictive models. We then evaluate their relationship by testing whether model uncertainty is higher where ice services disagree. We observe that supervision incorporating information from multiple annotators can improve this correspondence, with soft supervision achieving the highest overall correlation of 0.256. The relationship becomes substantially stronger near the ice edge, where model predictive uncertainty closely tracks multi-annotator disagreement, reaching a correlation of 0.704 within 0–10 km. Among the uncertainty-estimation approaches, Monte Carlo dropout provides the best-calibrated confidence estimates, with an expected calibration error of 0.050.
[LG-54] Concept drift mitigation through community and spectral graph analysis for the detectionof cyberattacks in network traffic
链接: https://arxiv.org/abs/2609.09442
作者: Julien Michel,Abdul Qadir Khan,Majed Jaber,Pierre Parrend
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注:
Abstract:In network traffic, legitimate behaviours and attack techniques evolve jointly - the phenomenon known as ‘concept drift’ [1]. Every detector is thereby left obsolete between two updates, and always one step behind adversaries. In this work, we propose to move the point of intervention from the model, repaired after the drift, to the feature space, selected before learning. We therefore introduce t-robustness, a stability score defined for each feature independently of any detection model, comparable across an entire feature space. It combines the step-by-step distance between successive statistical states of a feature, and its cumulative divergence from its initial state, so that a slow monotonic drift cannot pass for stability. The candidates are drawn from abnormal network connectivity patterns left by scans, DoS and communications between endpoints, read through graph community metrics and spectral metrics. The evaluation is performed on the UGR16 dataset, across three learning scenarios and a control scenario, as well as without model update, and demonstrate that t-robust feature spaces sustain detection where the baselines collapse: retained expectancy at the last test interval reaches 0.6025, against 0.5230 for graph community features and 0.3831 for the base NetFlow features.
[LG-55] nsor-Train Weak SINDy: Identifying High-Dimensional Nonlinear Dynamics
链接: https://arxiv.org/abs/2609.09434
作者: Will Houser,Vanja Dukic,David M. Bortz
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE); Machine Learning (stat.ML)
*备注: 34 pages, 8 figures
Abstract:In recent years, weak-form methods have made significant advances in data-driven discovery of dynamical systems. However, in high-dimensional settings, current techniques can prove expensive in both computation and memory. In this work, we introduce TT-WSINDy, which combines techniques of the Multidimensional Approximation of Nonlinear Dynamics (MANDy) and Weak Sparse Identification of Nonlinear Dynamics (WSINDy) methods, implementing requisite computations in the tensor-train (TT) format. We demonstrate that this method is able to search an exponentially-growing space of candidate functions – performing weak-form transformation, regression, and sparsification – without suffering from the curse of dimensionality.
[LG-56] Applying foundation model embeddings towards urban livability evaluation
链接: https://arxiv.org/abs/2609.09429
作者: Ayush Khot,Wen Zhou,Shaowen Wang
类目: Machine Learning (cs.LG); Computers and Society (cs.CY)
*备注: 11 pages, 5 figures, 6 tables
Abstract:While accurate measurement of socioeconomic indicators remains challenging in data-scarce regions, which limits policy interventions and resource allocation, high-resolution geospatial data is widely available and can contain information on various livability statistics. We investigate which physical features are encoded within foundation model embeddings, such as AlphaEarth, AnySat, and TerraMind, and provide a systematic framework for identifying the most predictive geospatial indicators. By analyzing how different types of geospatial data influence urban livability predictions, our approach enables researchers to prioritize the most informative features for their specific applications. Additionally, we demonstrate how to leverage foundation model embeddings to enhance prediction performance for these outcomes. This work contributes a principled methodology for extracting actionable information from satellite imagery while accounting for complex spatial dependencies, with applications in predicting urban livability in regions with limited observation data.
[LG-57] X-amine509: Predicting the Practical Risk Level of Enterprise X.509 Certificates
链接: https://arxiv.org/abs/2609.09402
作者: Cameron Keith,Shubh Patel,JD Kilgallin,Caleb Shorter
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: 22 pages, 11 figures
Abstract:Enterprises managing large X.509 certificate inventories face a prioritization problem: deterministic analysis tools that precisely identify standards violations are indispensable for remediation, but applying them exhaustively across millions of certificates is operationally impractical. We present X-amine509, a two-stage triage system that uses machine learning to rapidly rank certificates by predicted risk and route only the highest-risk items to full deterministic analysis. Certificate risk is quantified as a composite score derived from 177 defect checks grounded in CA/Browser Forum Baseline Requirements, NIST IR 8547/SP 800-57, and cryptographic strength criteria, weighted by security severity across four tiers ranging from cryptographic breaks to minor compliance deviations. We collected 1,027,714 X.509 certificates from Fortune 500, .gov, and .edu domains and scored each using this rubric. On a held-out test set of 201,976 certificates, our best model (Extra Trees) achieves R^2 of 0.993 with MAE of 2.26, while Decision Tree scores R^2 of 0.986 at 3.7 million certificates per second on a single machine. Ranking quality confirms the triage value: aggregate NDCG exceeds 0.997, and severity-tier classification reports 99.76% accuracy with 98.90% recall on critical-tier defects. Thirteen months later, we retrieved another 571,374 certificates to test our models’ durability over time, and the Extra Trees and Decision Tree models maintain MAE below 6.8, R^2 of at least 0.915, aggregate NDCG above 0.988, severity-tier accuracy of at least 99.52%, and critical-tier recall of at least 97.03%. Feature importance analysis identifies validity period, Extended Key Usage configuration, negative serial number encoding, and self-signed status as the strongest risk predictors, providing coarse interpretability at the triage stage.
[LG-58] XAI-Refine: An Automated Explanation-Knowledge Loop for Brain-Age Prediction
链接: https://arxiv.org/abs/2609.09388
作者: Yang Qiao,Junjie Wu,Deqiang Qiu,James J. Lah,Liang Zhao
类目: Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
*备注:
Abstract:Brain-age prediction models are commonly evaluated by predictive accuracy, yet accurate predictions alone do not establish that a model relies on reproducible or neurobiologically supported mechanisms. Post-hoc explanation methods can expose these mechanisms, but existing workflows typically stop at diagnosis or require correction targets to be specified before model analysis. We propose XAI-Refine, an automated explanation-knowledge loop for brain-age prediction from resting-state functional connectivity. At each iteration, XAI-Refine consolidates complementary post-hoc analyses across repeated training runs into reliable, structured model explanations. It converts each reliable explanation into a neutral neurobiological question, retrieves and verifies relevant literature, and compiles the verified evidence into an admissible set in the same typed explanation space. The target for refinement is defined as the minimal projection of the current model explanation onto the admissible set induced by applicable verified knowledge. This revised explanation is then translated into a differentiable constraint while preserving the originating model variable, measurement operator, and applicable scope. Candidate updates are promoted only when multi-seed validation confirms target-directed explanatory movement, predictive performance remains within a prespecified guardrail, and non-target explanatory drift remains bounded. Experiments on functional-connectivity-based brain-age prediction evaluate predictive performance, explanation reliability, literature alignment, and target-specific model revision, illustrating a structured route from post-hoc analysis to evidence-guided model refinement.
[LG-59] Constraint-Aware Discrete Black-Box Optimization Using Tensor Decomposition ECML KDD2026
链接: https://arxiv.org/abs/2609.09370
作者: Keisuke Onoue,Ryosuke Kojima
类目: Machine Learning (cs.LG)
*备注: 32 pages, including supplementary material. ECML PKDD 2026
Abstract:Discrete black-box optimization is often addressed using approaches such as Sequential Model-Based Optimization (SMBO), which aims to improve sample efficiency by fitting surrogate models that approximate a costly objective function over a discrete search space. In many real-world problems, the set of feasible inputs is often given by logical constraints known in advance. However, existing surrogate modeling techniques generally fail to capture the symbolic rules governing feasibility in discrete input spaces. In this paper, we propose a surrogate modeling approach based on tensor decomposition that captures the structure of discrete search spaces while directly integrating feasibility information. To implement this approach, we formulate surrogate model training as a constrained polynomial optimization problem and solve a relaxed formulation using a differentiable penalty term derived from T-norms. Our experiments on both synthetic and real-world benchmarks, including a pressure vessel design task, demonstrate that the proposed method improves sample efficiency by effectively guiding the search away from infeasible regions.
[LG-60] Explaining f-Divergence-Based Regularization via Local Curvature and Sharpness-Aware Minimization
链接: https://arxiv.org/abs/2609.09367
作者: Nour Jamoussi,Marios Kountouris
类目: Machine Learning (cs.LG); Information Theory (cs.IT)
*备注: Published at 5th Conference on Lifelong Learning Agents (CoLLAs), 2026
Abstract:Divergence-based regularization and Sharpness-Aware Minimization (SAM) are two prominent approaches for improving generalization in deep learning, both motivated by robustness to perturbations. However, their relationship has remained largely unexplored. Building on classical second-order expansions of f -divergences, we show that the two methods are locally consistent under parameter-space perturbations: both induce curvature-sensitive penalties, with divergence regularization yielding a Fisher-weighted quadratic form and SAM penalizing sharpness through the dominant Hessian eigenvalue. For negative log-likelihood objectives with exponential-family output distributions, this correspondence becomes especially transparent, since the Fisher and Gauss-Newton matrices coincide. We further show that the same local geometric perspective extends to input-space perturbations, where divergence-based regularization is defined through transformations of the input. In this setting, the regularizer induces a pullback quadratic form on the input space, providing a more general perturbation framework than standard SAM while preserving the same local sensitivity interpretation. To validate the analysis empirically, we use the asymmetric \alpha -skew Jensen-Shannon divergence (JSD) family as a controlled testbed. Its local curvature coefficient scales as \alpha(1-\alpha) and is maximized at the symmetric point \alpha=\tfrac12 , which recovers the standard JSD. Loss-landscape visualizations in the input-perturbation regime show that stronger induced curvature penalization is associated with flatter local minima. Experiments on four benchmark datasets further demonstrate that both accuracy and negative log-likelihood are consistently best near this regime of maximal curvature penalization.
[LG-61] Encrypt What Matters: When Selective Homomorphic Inference Is Efficient
链接: https://arxiv.org/abs/2609.09357
作者: Ali Backour,Juan Reyes,Jaime Punyed,Ana Onoprishvili
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Fully homomorphic encryption (FHE) enables inference on private data without revealing it to the server, but evaluating an entire input under FHE is expensive. We study \emphselective homomorphic inference, where only a sensitive region of interest (ROI) is encrypted, and computations independent of that region are performed in plaintext. Selective evaluation produces the same output as full FHE on the same model, without retraining. Its efficiency depends on how quickly encrypted dependencies spread through the network. For small encrypted ROIs, locality-preserving architectures can achieve order-of-magnitude homomorphic-evaluation speedups, whereas architectures with early global mixing provide essentially no speedup. These results identify locality as the key architectural property governing the benefit of selective homomorphic inference.
[LG-62] Literati: Towards Anytime Optimal Shape Generalized Trees via AO*
链接: https://arxiv.org/abs/2609.09299
作者: Nakul Upadhya,Eldan Cohen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Decision trees are prized for their interpretability and strong performance on tabular data, but popular greedy top-down induction algorithms can yield suboptimal and unnecessarily complex structures. Optimal decision tree methods address this through global optimization, yet remain restricted to axis-aligned threshold splits, which limit the expressivity of each node and often force deep, complex trees to capture non-linear feature effects. Shape Generalized Trees (SGTs) generalize threshold splits to learnable univariate shape functions, improving expressivity and enabling more compact trees. However, existing SGT induction algorithms are greedy and offer no optimality guarantees. In this work, we introduce Literati, the first algorithm for optimal SGT induction. We propose a novel AND/OR graph formulation of the problem that jointly optimizes tree structure and shape function complexity. To solve this AND/OR graph, we develop an AO*-based algorithm with two enhancements that improve anytime performance while preserving optimality: a secondary heuristic for OR-node selection and a round-robin policy for AND-node exploration. Across 24 real-world datasets, Literati achieves higher training and test accuracy than state-of-the-art tree approaches.
[LG-63] Accountable and uncertainty-aware evaluation of sensor-based AI under distribution shift: devices subjects and nearly three years underground
链接: https://arxiv.org/abs/2609.09257
作者: Benny Platte,Rico Thomanek,Christian Roschke,Marc Ritter(Mittweida University of Applied Sciences)
类目: Machine Learning (cs.LG); Methodology (stat.ME); Machine Learning (stat.ML)
*备注: 24 pages, 8 figures, 8 tables. Manuscript prepared for Measurement Science and Technology
Abstract:Sensor-based AI systems are rarely operated under the conditions under which they were trained: devices, personnel and recording epochs change, and each change degrades performance in ways a random train-test split cannot reveal. We propose a staged, accountable evaluation protocol that treats the evaluation of a deployed model as a measurement with declared reference levels and a quantified uncertainty. Four cumulative generalisation stages hold out devices, subjects and time. Each stage is judged on quantiles of repeated trainings against chance references with the correct class count, an out-of-present-scope rate exposes silent misdirection towards classes that are no longer present in deployment relative to training, and an explicit decision rule ties roll-out decisions not to means but to 5% quantiles. We demonstrate the protocol on infrastructure-free geomagnetic localisation with smartphone-based recurrent classifiers in two real underground mines, including a replication of the scheme’s training stages at the second site. Unchanged models are re-evaluated on data recorded 34 months after the training campaign, on a device generation unknown at training time and with a held-out surveyor. The 5% quantile of their present-conditioned precision there is 0.39 over 299 repeated trainings, 16.5 times the chance level; across the composition of the 42 reachable location classes the figure varies by +/-0.08, several times the spread between repeated runs. Repeated trainings of a single configuration show why means mislead: a bimodal configuration passes a mean-based test decisively while its 5% quantile lies more than an order of magnitude below chance.
[LG-64] Cost-Aware Post-Hoc Deferral Under Calibration and Shift: An Environmental AI Case Study
链接: https://arxiv.org/abs/2609.09235
作者: Haoran Yu,Lifei Liu,Danping Zhang
类目: oftware Engineering (cs.SE); Machine Learning (cs.LG)
*备注:
Abstract:Choosing a deferral policy for a frozen classifier requires more than ranking uncertain cases: confidence may be miscalibrated, errors have unequal costs, reviewers can err, and deployment data can leave calibration support. We study these interactions through EcoTrust, a post-hoc framework that compares automatic action with review using a six-group error-risk estimator, class-asymmetric costs, reviewer accuracy, and an optional support gate. On a Columbia River thermal-stress testbed, the learned estimator improves error-ranking area under the receiver operating characteristic curve from 0.869 to 0.889, but Chow’s confidence rule has lower in-distribution cost (0.416 versus 0.567 per day). Across 12 off-the-shelf backends, learned risk and a calibration-matched, class-aware confidence estimator each beat raw Chow on six; a paired year-block bootstrap does not resolve their mean cost difference. In transfer to ten river stations, the gate flags every case and becomes an always-review fallback, attaining the lowest cost on eight stations only when review is perfect and unconstrained. These results characterize decision boundaries on one controlled task: richer risk signals do not reliably improve on calibrated confidence, and detected extrapolation does not imply transferable case-level ranking.
[LG-65] Learning to Fly: Stable Vision-Guided UAV Servoing with Compact Target-Centric Cues and Reinforcement Learning
链接: https://arxiv.org/abs/2609.09234
作者: Saurbh Singh Jamwal,Nived Chebrolu
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Vision-guided reinforcement learning for Unmanned Aerial Vehicles (UAVs) remains challenging due to unstable policy optimisation, aggressive exploration, and the cost of high-dimensional visual perception. In this work, we investigate long-horizon UAV visual servoing using compact target-centric cues combined with low-dimensional sensor measurements. Rather than learning directly from RGB images, lightweight target segmentation provides image-space offsets and relative depth, which are combined with quadrotor velocity and projected-gravity measurements into a compact 12D policy observation. We compare Direct PPO with three matched-budget curriculum strategies: a Visual curriculum that progressively expands target placement difficulty, a Dynamics curriculum that gradually relaxes action constraints and smoothing, and a Joint curriculum that combines both progressions. All strategies reach comparable nominal performance, with complementary advantages across tracking metrics. Observation ablations show that proprioceptive measurements are critical for stable flight and image-space cues for target alignment, while explicit depth is not necessary for strong performance in the evaluated setting. Against tuned classical visual-servo controllers, learned policies show greater robustness to strong control and visual perturbations, while the Visual curriculum exhibits the smallest degradation under unseen target motion. Overall, the results demonstrate that compact target-centric representations can support robust long-horizon aerial visual servoing and that visual curriculum training can improve robustness to dynamic distribution shifts despite limited gains in nominal performance.
[LG-66] World-Time Compute with Verified Code World Models
链接: https://arxiv.org/abs/2609.09163
作者: James Schwoebel,Ingrida Semenec,Jenia Rousseva,Marcos Ortiz,Collin Overbay,Christopher Klaus,Anderson Edmond,Manish Bhatt,Rome Thorstenson,Jessica Tsai,Martin G. Frasch
类目: Machine Learning (cs.LG)
*备注: 36 pages, 16 figures, 12 tables. Code and data available at this https URL
Abstract:LLMs generalize across a domain only after seeing many real, labeled examples, which most domains lack. We study a way to manufacture it cheaply. When a domain’s dynamics can be written as code, one template instantiates into many world models: executable, verifiable programs over symbolic state, each an inexhaustible source of exactly-labeled trajectories. Fine-tuning an LLM on trajectories through many such worlds, which we call world-time compute, a training-time analogue of test-time compute, lifts generalization to held-out worlds it never trained on (synthesized world families). Gains are largest where capability is scarcest: +29 points at 0.5B; the largest model’s lift is within noise, consistent with saturation. Labels can be trusted because the worlds are verified code: synthesized-then-checked dynamics are exact over 20-step rollouts and answer 10x out-of-distribution probes exactly (100%), whereas per-step LLM and MLP predictors compound error and collapse. Unlike domain randomization, each world is independently authored and verified; a corrupted-label control shows label exactness, not task variety, drives the gains. On real benchmarks (ARC-AGI grids, List Functions, CLRS) the same lever holds as per-world test-time training. On List Functions the harder cross-world form holds: one adapter trained on 128 disjoint worlds reaches 40% on held-out worlds versus 6% for a corrupted-label control (+34 points, CI [29, 39]). The gain is a saturating regularity, not a law: largest for few-step reasoning and small/weak models, fading for long chains, perception-induced tasks, and saturated tasks; cross-task transfer is weak without shared skill. Worlds are authored and served by OpenWorld, a zero-dependency framework (companion paper). Scope: symbolic state; pixel-native domains remain territory of learned models. All code, recipes, and this manuscript regenerate from one repository.
[LG-67] Physics-informed neural networks by Gradient-Guided Gaussian Adaptive Sampling (3GAS-PINNs)
链接: https://arxiv.org/abs/2609.09162
作者: Yousen Wang,Wei Zhao
类目: Machine Learning (cs.LG); Computational Physics (physics.comp-ph); Fluid Dynamics (physics.flu-dyn)
*备注:
Abstract:Physics-informed neural networks (PINNs) provide a mesh-free framework for solving partial differential equations, yet their performance in nonlinear problems is often limited by slow convergence, gradient imbalance, and insufficient resolution to capture localized intermittent structures such as shock waves[1]. These issues arise primarily from the use of fixed weights of loss and uniform collocation point distributions, which cannot adapt to the evolving complexity of the solution field during training. To address these challenges, Gradient-Guided Gaussian Adaptive Sampling Physics-Informed Neural Networks (3GAS-PINNs) is proposed in this paper, which combines uniform probability distribution and Gaussian-smoothed probability distribution derived from the spatial gradients of solution, to maintain global constraint satisfaction as well as concentrating collocation points in regions of high gradient. Thus, intermittency structures like shock wave and solitons can be accurately captured. The method is evaluated on three benchmark nonlinear problems, including one-dimensional forced Burgers equation, Korteweg-de Vries (KdV) equation and nonlinear Schrodinger equation, all of which exhibit steep gradients or strong nonlinearity. In comparison with baseline PINNs, 3GAS-PINNs can effectively promote the physical consistency in intermittent regions. The accuracy of the numerical simulation can be improved by a factor of up to 14.
[LG-68] Spectral origin of the topological gap exponent d η: mechanism kernel decomposition and scope
链接: https://arxiv.org/abs/2609.09159
作者: Matthew Loftus
类目: Machine Learning (cs.LG); Statistical Mechanics (cond-mat.stat-mech); Algebraic Topology (math.AT)
*备注: 7 pages, 6 tables
Abstract:The topological gap \Delta – the excess H_1 total persistence of a critical point cloud over a density-matched null – scales as \Delta \sim L^d+\eta . We derive this analytically: the spectral integral I(\alpha) = \sum_k\neq 0 S_\mathrmconn(k),|k|^\alpha scales as L^2-\alpha-\eta when IR-dominated, giving I(-2\eta) \sim L^d+\eta . The decomposition I(-2\eta) = I_0 \cdot I_\mathrmshape separates volume ( I_0 \propto N(1-m^2) \sim L^d ) from anomalous dimension ( I_\mathrmshape \sim L^\eta ); the volume factor accounts for the magnetization-driven per-configuration variance of \Delta . We prove the mechanism requires d 2 + \eta (IR dominance), confining it to d = 2 for physical systems; in d = 3 the spectral integral is UV-dominated, explaining why density normalization is needed. An \alpha -sweep for Potts q = 4 at L = 32 – 256 finds \alpha_\mathrmopt in [-0.75, -0.5] , consistent with -2\eta_\mathrmIsing and inconsistent with -2\eta_q=4 = -1 ; we flag this as tentative pending L \geq 1024 confirmation. The \langle m^2 \cdot I(-2\eta)\rangle hyperscaling product is dominated by the correlation r(m^2, I) \approx -0.98 via the shared I_0 amplitude, so we report it as a covariance-correction analysis. Under a heuristic argument extending Divol–Polonik to inhomogeneous Poisson intensities, the bare PH kernel is flat; the effective kernel acquires k -dependence only at criticality. The per-configuration agreement between \Delta and I(-0.5) is primarily a magnetization correlation: R^2 = 0.91 at L = 256 collapses to R^2 \approx 0 once |M| is partialed out. Per-configuration evidence corroborates the I_0 Parseval identity but not the |k|^-2\eta shape factor; the latter is established by ensemble L -scaling.
[LG-69] On Scaling Coordinate-Based Neuroevolution: The Quadtree Bottleneck in ES-HyperNEAT
链接: https://arxiv.org/abs/2608.24480
作者: Romain Claret,Michael O’Neill,Paul Cotofrei,Kilian Stoffel
类目: Neural and Evolutionary Computing (cs.NE); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注:
Abstract:ES-HyperNEAT evolves substrate topology through adaptive quadtree subdivision; to our knowledge, no implementation with full population-level GPU parallelization exists. We present JAX-ESHN, a JAX-based implementation targeting GPU parallelization with batched CPPN queries, and benchmark it against the CPU-based PUREPLES Baseline across five tasks: XOR, Parity-3, circle classification, sine regression, and CartPole. The core limitation is structural: each CPPN discovers a unique set of substrate positions, preventing population-level vectorization via vmap. On XOR, the CPU Baseline’s runtime scales exponentially with depth while JAX-ESHN’s construction cost on GPU (compilation plus first-generation evaluation) plateaus at deep substrates, so JAX-ESHN solves reliably where the Baseline rarely succeeds, with lower runtime variance. A CPU-vs-CPU multi-benchmark control reproduces the same scaling divergence across Boolean, continuous, and control task types, confirming it is a property of the substrate-discovery implementation, not of GPU hardware. An alternative data structure (Hierarchical Spatial Hash Grid) fails not because it precomputes positions but because it applies the variance test independently per position, discarding the quadtree’s parent-gated filtering and with it the adaptive sparsity essential to ES-HyperNEAT. These findings define the structural constraints any substrate-discovery method must satisfy to scale coordinate-based neuroevolution; the companion EMR-HyperNEAT reformulation, which replaces adaptive subdivision with eager evaluation of a static multi-resolution grid, satisfies them and resolves the bottleneck this paper characterizes.
[LG-70] Likelihood-free inference with nuisance parameters through normalizing flows
链接: https://arxiv.org/abs/2609.10534
作者: Phil Assheton
类目: Methodology (stat.ME); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 49 pages and 13 figures, including appendices. Code available at this https URL
Abstract:We present a simple decomposition of a neural-network-based normalizing flow that naturally uncovers a pivotal statistic (or something close) in the presence of nuisance parameters, based only on a sample generator from the distribution of interest. We show that the statistic is near-pivotal in the sense of minimum average KL-divergence of its p -values versus uniform and we argue that it can be expected to have good power when the dimension of the statistic equals the dimension of the parameter. It is able to incorporate prior knowledge about group invariances such as translation and scale. It can discover the one-sample t -test almost exactly, outperforms the Welch test in terms of worst-case size over a constrained variance-ratio range and achieves good calibration on partial biserial correlations, while showing higher power (and being much faster) on small-to-moderate samples than profile likelihood-ratio techniques.
[LG-71] Optimal Low-Rank Quantum State Tomography with Bounded-Sample Joint Measurements
链接: https://arxiv.org/abs/2609.10514
作者: Ashwin Nayak,Xingyu Zhou
类目: Quantum Physics (quant-ph); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG)
*备注: 70 pages
Abstract:We determine the optimal sample complexity of low-rank quantum state tomography when each measurement may act jointly on at most t samples. For sufficiently small \varepsilon , estimating an unknown state on \mathbbC^d of rank at most r to trace norm error \varepsilon with constant success probability requires, and is achievable with, \Theta\left( \fracdr\varepsilon^2 \max\left\1,\frac r\sqrt t\right\ \right) samples. The lower bound allows the protocol to choose each joint measurement adaptively using all previous classical outcomes; the matching upper bound is nonadaptive. Thus joint measurements on at most t samples improve the complexity of algorithms making single-sample measurements by at most a factor \sqrt t . Further, measuring order r^2 samples jointly is necessary and sufficient to attain the unrestricted collective rate. For the lower bound, we vary the support of a state with fixed uniform spectrum and bound the Fisher information trace of every joint measurement on t samples. The adaptive Fisher chain rule and the van Trees inequality then give the trace norm lower bound. For the upper bound, we construct and analyze a nonadaptive tomography protocol based on a Gaussian joint measurement. An explicit second moment identity and a conditional Gaussian law outside the state’s support give a rank-dependent error analysis, yielding the matching rate. Comments: 70 pages Subjects: Quantum Physics (quant-ph); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG) Cite as: arXiv:2609.10514 [quant-ph] (or arXiv:2609.10514v1 [quant-ph] for this version) https://doi.org/10.48550/arXiv.2609.10514 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Xingyu Zhou [view email] [v1] Wed, 9 Sep 2026 17:48:41 UTC (68 KB) Full-text links: Access Paper: View a PDF of the paper titled Optimal Low-Rank Quantum State Tomography with Bounded-Sample Joint Measurements, by Ashwin Nayak and 1 other authorsView PDFHTML (experimental)TeX Source view license Current browse context: quant-ph prev | next new | recent | 2026-09 Change to browse by: cs cs.DS cs.IT cs.LG math math.IT References Citations INSPIRE HEP 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
[LG-72] Deep Learning-Based Detection of Electrical Faults and Power Quality Disturbances in Aerospace Power Systems
链接: https://arxiv.org/abs/2609.10479
作者: Ian C. Guzmán,Radu Babiceanu,Berker Peköz
类目: ignal Processing (eess.SP); Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: Accepted for publication in IEEE Transactions on Aerospace and Electronic Systems. Pending journal reference/external DOI
Abstract:More Electric Aircraft require fast and reliable monitoring of high-frequency electrical networks, yet most power quality disturbance and fault diagnosis methods are developed for conventional 50 or 60 Hz grids. This work presents a hardware-aware deep learning framework for multiclass detection of electrical faults and power quality disturbances in a 400 Hz aerospace power system. A high-fidelity simulation model inspired by the Boeing 787 electrical architecture generates voltage and current waveforms for 21 normal, disturbance, switching, open-circuit, and short-circuit conditions. Two datasets, each containing 73,500 samples, are formed from one-dimensional time-series signals and short-time Fourier transform time-frequency representations. Signal-processing augmentation, domain randomization, and class-specific generative adversarial networks increase waveform diversity, and the time-series dataset is released through IEEE DataPort. We compare 1D and 2D convolutional neural networks, long short-term memory networks, CNN-LSTM hybrids, ResNet, MobileNet, and VGG models under common training conditions. A compact ResNet provides the best accuracy-complexity tradeoff, achieving 96.94 percent software test accuracy with 175,685 parameters. After 8-bit quantization and deployment on a Xilinx Zynq UltraScale Plus MPSoC ZCU102, the model achieves 95.87 percent accuracy and a measured mean neural-network accelerator latency of 6.90 ms per input record. The results establish simulation-based, accelerator-level feasibility for embedded edge AI in aircraft electrical health monitoring and motivate future end-to-end data acquisition and experimental validation.
[LG-73] Algorithmic stability via ensembling
链接: https://arxiv.org/abs/2609.10428
作者: Rina Foygel Barber,Richard J. Samworth
类目: atistics Theory (math.ST); Machine Learning (cs.LG)
*备注:
Abstract:Algorithmic stability refers to the property of an algorithm being insensitive to perturbations of the input data, where the type of perturbation may vary depending on the setting. In this work, we develop a general framework to quantify the extent to which any ensembling strategy defined via averaging can yield stability guarantees for any type of data perturbation. Our main theoretical result is a guarantee on the stability of this ensembled algorithm, given in terms of the norm of a certain covariance operator that describes the ensembling process. We show how our general framework yields interpretable and intuitive insights in several examples of perturbations of practical interest, and provides much sharper guarantees than those obtained from privacy considerations.
[LG-74] Searching for New Physics with Reinforcement Learning
链接: https://arxiv.org/abs/2609.10382
作者: Jacky Kumar,Marianne Bouchard,David London
类目: High Energy Physics - Phenomenology (hep-ph); Machine Learning (cs.LG)
*备注: 6 pages, 1 figure
Abstract:Finding new physics (NP) is the most important problem in particle physics today. Studying ``anomalies’', i.e., measurements of low-energy observables whose values disagree with the predictions of the Standard Model (SM), is a powerful search strategy. The SM Effective Field Theory (SMEFT) provides a general model-independent framework for parameterizing NP; it is natural to try to find the SMEFT operator(s) that can explain such anomalies. This is a challenging task because (i) the number of SMEFT operators is enormous, and (ii) at loop level there are very complicated correlations among the operators. Analyses by humans typically rely on phenomenological intuition to decide which operators are relevant. This is often biased and does not explore the complete SMEFT operator space. Interestingly, reinforcement learning (RL) techniques excel at tasks that require decision making to achieve their goals. In this paper, we introduce an RL method that can be used to find the SMEFT operators that explain any anomalies. We test it on the CDF W -mass anomaly, and show that it reproduces (and improves upon) known results. We then consider a far more complicated situation with multiple anomalies and show that, even here, this method is able to find the SMEFT operators that explain the data. Our RL method can therefore be used to efficiently search for NP at the level of SMEFT.
[LG-75] Are You Learning Biological Signal or Shortcuts? Auditing and Mitigating Bias in Protein-Protein Interaction Datasets
链接: https://arxiv.org/abs/2609.10193
作者: Judith Bernett,Anton Spannagl,Joel Ås,Markus List,David B. Blumenthal
类目: Molecular Networks (q-bio.MN); Machine Learning (cs.LG)
*备注:
Abstract:Protein-protein interaction (PPI) databases do not faithfully reflect biological realities. Instead, they are influenced by study and technical biases that distort certain protein and interaction attributes. Machine learning models can exploit these as learning shortcuts if the negative dataset is not constructed with care. So far, the shortcuts introduced during PPI dataset construction have only been examined in isolation. Here, we systematically characterize both reported and, to our knowledge, previously unreported biases in PPI datasets that lead machine learning models to learn shortcuts instead of biological signal. We analyze HIPPIE, IntAct, and STRING, dedicated PPI databases, as well as two datasets derived from 3D-structural information in the Protein Data Bank (PDB). We show that random data splitting introduces strong topological shortcuts. When train-test protein overlap is removed, the resulting datasets still retain usable shortcuts stemming from self-interactions, taxonomic identity, and functional relatedness, whose prevalence interestingly depends on the data source. We further show that sampling negatives from a set of high-confidence non-interactors, an intuitively appealing choice, can amplify the shortcut stemming from functional relatedness. To detect and mitigate these biases, we provide an open Nextflow pipeline that combines similarity-aware, data-loss-minimizing dataset splitting with bias-minimizing negative sampling, both formulated as integer linear programs. Its key concept of quantifying biases to minimize them through optimization-based negative sampling can, in principle, be extended to any machine learning problem where the pool of negative candidates is much larger than the positives and is thus of interest also beyond PPI prediction.
[LG-76] he Sample Complexity of Quantum Entanglement Allocation
链接: https://arxiv.org/abs/2609.10141
作者: Nathan Roll
类目: Quantum Physics (quant-ph); Machine Learning (cs.LG)
*备注: 61 pages, 15 figures. Includes code and recorded data as ancillary files
Abstract:How many past requests are needed to decide which qubits should share entanglement? We show that the answer depends on the allocation choices created by the queries: a larger memory can require no more data. The memory stores a classical bit and answers requests through a fixed detector that preserves coherence within each measured sector. For independent commuting X - and Z -type Pauli queries, we characterize the full attainable prediction-contrast region and construct encodings that preserve the bit at every nonzero vertex. With sharp reports, a d -qubit path and groups of at most k qubits have minimax excess error after m requests proportional to k^-1\min\1,\sqrtd\log(k+1)/m\ , uniformly for 2\leq kd . Connected biclique regions can grow without increasing sample demand when depth, region count and connections per region stay bounded. Preparation noise introduces a separate calibration requirement. We derive an exact tradeoff with extra fresh detector calls and transfer the learning law to structured transaction co-location. Population-risk experiments test the statistical predictions. We also compare encodings on a native 15-qubit device and learned partitions on public purchase baskets. The full chain wins on the device; frequency grouping outperforms basket search in the largest-capacity retail setting.
[LG-77] Physics-Informed Multi-Task Surrogate Model for the Martian Nightside Thermosphere
链接: https://arxiv.org/abs/2609.10077
作者: Sergey Nikiforov
类目: Earth and Planetary Astrophysics (astro-ph.EP); Machine Learning (cs.LG)
*备注: 5 pages, 1 figure, 2 tables. 3rd Conference on AI in and for Space (SPAICE 2026)
Abstract:Modeling the Martian nightside thermosphere remains challenging due to sparse in situ sampling and strong coupling among transport, magnetic, and seasonal processes. Purely data-driven models can produce non-physical artifacts, such as density inversions, in poorly sampled altitude regimes. We present a multi-task physics-informed neural network that simultaneously predicts the base-10 logarithmic densities of four neutral species (O, CO _2 , N _2 , and Ar) using more than a decade of MAVEN/NGIMS observations (MY 32-38, 2014-2025). A shared backbone learns a common representation of the nightside thermospheric state and branches into species-specific output heads. A weak monotonicity prior is incorporated via automatic differentiation by penalizing positive vertical gradients in logarithmic density. Experiments using an orbit-disjoint train/validation/test split show that physics-informed regularization substantially reduces non-physical inversions while preserving predictive skill and slightly improving it in the best-performing configuration, as measured by RMSE, MAE, and R^2 . The resulting model provides a computationally efficient surrogate for nightside thermospheric reconstruction with improved vertical consistency. Comments: 5 pages, 1 figure, 2 tables. 3rd Conference on AI in and for Space (SPAICE 2026) Subjects: Earth and Planetary Astrophysics (astro-ph.EP); Machine Learning (cs.LG) Cite as: arXiv:2609.10077 [astro-ph.EP] (or arXiv:2609.10077v1 [astro-ph.EP] for this version) https://doi.org/10.48550/arXiv.2609.10077 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Sergey Nikiforov [view email] [v1] Wed, 9 Sep 2026 12:00:52 UTC (106 KB)
[LG-78] Dynamical Non-compensatory Multidimensional IRT Model Using Variational Approximation
链接: https://arxiv.org/abs/2609.10028
作者: Hiroshi Tamano,Daichi Mochihashi
类目: Methodology (stat.ME); Machine Learning (cs.LG)
*备注:
Abstract:Multidimensional item response theory (MIRT) is a statistical test theory that precisely estimates multiple latent skills of learners from the responses in a test. Both compensatory and non-compensatory models have been proposed for MIRT: the former assumes that each skill can complement other skills, whereas the latter assumes they cannot. This non-compensatory assumption is convincing in many tests that measure multiple skills; therefore, applying non-compensatory models to such data is crucial for achieving unbiased and accurate estimation. In contrast to tests, latent skills will change over time in daily learning. To monitor the growth of skills, dynamical extensions of MIRT models have been investigated. However, most of them assumed compensatory models, and a model that can reproduce continuous latent states of skills under the non-compensatory assumption has not been proposed thus far. To enable accurate skill tracing under the non-compensatory assumption, we propose a dynamical extension of non-compensatory MIRT models by combining a linear dynamical system and a non-compensatory model. This results in a complicated posterior of skills, which we approximate with a Gaussian distribution by minimizing the Kullback-Leibler divergence between the approximated posterior and the true posterior. The learning algorithm for the model parameters is derived through Monte Carlo expectation maximization. Simulation studies verify that the proposed method is able to reproduce latent skills accurately, whereas the dynamical compensatory model suffers from significant underestimation errors. Furthermore, experiments on an actual data set demonstrate that our dynamical non-compensatory model can infer practical skill tracing and clarify differences in skill tracing between non-compensatory and compensatory models.
[LG-79] Optimal Value Inference for Reinforcement Learning
链接: https://arxiv.org/abs/2609.09981
作者: Nan Lu,Ethan Lee,James M. Robins,David Simchi-Levi,Junwei Lu
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
*备注:
Abstract:We study offline inference for the optimal value in reinforcement learning. Two new nuisances are derived as fixed points of a self-induced Bellman equation, in which we approximate the maximum Bellman operator by its softmax correspondence. We propose a debiased estimator through the Neyman orthogonality and establish its asymptotic normality under diverging horizons even when the behavior policy changes with time, as long as the nuisances have the statistical rates that can be achieved by many machine learning methods. We provide a concrete estimating procedure for these nuisances and show they can lead to valid inference. Synthetic experiments validate the numerical performance of our inference method, and we implement it in real-life decision-making problems, including bike repositioning and AI agentic tool use.
[LG-80] A Unifying Perspective on Probabilities as Model Predictions
链接: https://arxiv.org/abs/2609.09855
作者: Benedikt Höltgen
类目: Machine Learning (stat.ML); Computers and Society (cs.CY); Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注:
Abstract:Although probabilistic statements are ubiquitous, foundational disagreements persist about their understanding, as exemplified by debates between Bayesians and frequentists; moreover, it is unclear when and why acting on them actually leads to desirable outcomes. Here, we argue that every probability is the output of a \emphprediction method, that is, it depends on both a particular way of constructing abstractions and a way of transforming them into predictions. Through this, we provide a unifying perspective on supposedly different kinds of probabilities and show that even supposedly objective ones are model-dependent. We demonstrate that when a finite calibration criterion is met, one can anticipate the distribution of utilities for a given policy and inform successful decision-making on finite sets of events. Based on the notion of prediction methods, inductive arguments, and the probability calculus, we explain the feasibility of the calibration criterion in many settings. Overall, we develop a coherent perspective on probabilities and their use, connecting key intuitions behind other interpretations along the way.
[LG-81] Efficient Graph Neural Networks for Multicarrier Wideband Hybrid Beamforming Optimization
链接: https://arxiv.org/abs/2609.09708
作者: Beier Li,Mai Vu
类目: ignal Processing (eess.SP); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注: Accepted by IEEE Transactions on Wireless Communications
Abstract:6G wireless technology is poised to adopt higher and wider frequency bands, leveraging highly directional beamforming. However, the vast bandwidths amplify the impact of beam squinting. Traditional solutions, such as adding a true-time-delay filter to each antenna, are cost-prohibitive due to the required hardware scale. This paper proposes a signal processing alternative using Graph Neural Networks (GNNs) to optimize hybrid beamforming in multicarrier wideband systems. Using a bipartite graph to represent a shared analog beamformer among multiple subcarriers, we develop three GNN structures with distinct digital beamformer representations (i) at the subcarrier nodes, (ii) at the edges, or (iii) integrating traditional singular-value decomposition solutions. By designing an efficient message-passing mechanism, these structures offer insights into the impact of different GNN designs on communication system performance and computational complexity. Extensive analysis and ablation studies show that our proposed GNN structures outperform traditional optimization methods and existing ML-based solutions. Furthermore, the proposed GNNs exhibit strong resiliency to beam squinting and better robustness against imperfect CSI than even fully digital beamforming and all existing hybrid designs. These GNNs can also be extended to multi-user scenarios and demonstrate excellent generalization capabilities, allowing trained models to adapt to diverse multicarrier and multi-user settings without retraining.
[LG-82] Why Learning Rediscovers the Closed-Form Diagonal Regularizer
链接: https://arxiv.org/abs/2609.09656
作者: Jeahn Han,Pyojin Kim
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Robotics (cs.RO); Audio and Speech Processing (eess.AS)
*备注: main paper: 9 pages, 3 figures appendix
Abstract:We identify a diagonal saturation principle in modal inverse problems: when truncation noise is isotropic, the Bayes-optimal Tikhonov shape is a closed-form power law Gamma_k proportional to lambda_k^|s| set by the prior alone, independent of the domain. Berry’s random-wave conjecture decorrelates the truncation noise across modes, and Weyl’s eigenvalue counting law supplies enough modes for the conclusion to survive empirical Berry violations. Together they predict an approximately flat loss landscape across the per-mode family, leaving narrow scope for a diagonal regularizer to robustly beat the closed form. On FEM-simulated acoustic rooms, the closed form is near-optimal relative to per-room oracle tuning across observation windows, and three diagonal architectures trained on the same data match its reconstruction error within 1 pp despite learning qualitatively different spectra. The framework extends to heat diffusion via a known exponential Green’s function correction with no new free parameters. Saturation is restricted to the diagonal family: Learned Iterative Ridge crosses the boundary by exploiting cross-mode coupling, locating where learning starts to help.
[LG-83] Distillation of Synthetic Data for Time Series Foundation Models
链接: https://arxiv.org/abs/2609.09586
作者: Niloy Biswas,Noureddine El Karoui
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 10 pages, 3 figures
Abstract:Time series foundation models (TSFMs) are increasingly pre-trained on synthetically generated time series trajectories, where the data generating process is known. Current pre-training recipes are based on loss objectives which compare TSFM outputs to realized future values of each trajectory. We instead propose loss objectives which compare TSFM outputs to the conditional forecast distribution of each trajectory, a procedure we call synthetic data distillation (SDD). SDD corresponds to a Rao-Blackwellization of the training objective, in that it leaves the expectation of stochastic gradients unchanged while provably reducing the covariance of the stochastic gradient under the Loewner partial ordering. We empirically validate SDD on a TSFM model family of sizes from 4 M to 2.5 B parameters, and observe faster convergence of validation loss at every model size: on Gaussian Process data, SDD attains or improves upon the Status Quo loss whilst requiring 10%-40% less training iterations.
[LG-84] Differentially Private Averag e Treatment Effect Estimation by Propensity Score Blocking
链接: https://arxiv.org/abs/2609.09536
作者: Duncan Stewardson,Grayson W. White,Adam Groce
类目: Methodology (stat.ME); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Average treatment effect (ATE) estimation in observational studies is a fundamental statistical tool used frequently in social science, medicine, and other fields. These fields often work with sensitive data where privacy protections are important, so a differentially private mechanism for ATE estimation is highly desirable. Here we present two propensity score-based algorithms for ATE estimation on observational data, one improving the inverse probability weighting (IPW) method used in prior work, and the other using blocking on the propensity score (BPS). Both show lower error and less bias than prior work, with the BPS-based algorithm frequently reducing error by 75% or more compared to prior work.
[LG-85] Oracle Complexity of Stochastic Fixed-Point Equations with Nonexpansive Maps
链接: https://arxiv.org/abs/2609.09524
作者: Jelena Diakonikolas,Cristóbal Guzmán,David Martínez-Rubio
类目: Optimization and Control (math.OC); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:We study the oracle complexity of computing a point with small fixed-point residual |T(x)-x| \leq \epsilon , for a general norm |\cdot| and a self-map T of a compact convex set. We study this problem in the setting where T is nonexpansive with respect to the same norm |\cdot| and accessed via an unbiased stochastic oracle with bounded variance \sigma^2 . We provide an algorithm that solves such instances for any norm with a weak Rademacher type q 1 , with high probability. The algorithm is based on a recursive anchoring technique. For type- 2 spaces, such as \ell_p -spaces for p \in [2, \infty] , our algorithm attains stochastic oracle complexity \tilde O(\sigma^2 \epsilon^-3 + \epsilon^-1) . We further prove a near-matching lower bound (i.e., matching up to poly-log factors) for such \ell_\infty -norm instances in high dimensions. Our lower bound holds against any randomized algorithm that succeeds with constant probability. It further extends to settings with ``sparse’’ noise, where variance measured with respect to any \ell_p norm is of the same order, ruling out the possibility of improving oracle complexity as a function of \varepsilon by measuring variance in a non-matching \ell_p norm.
[LG-86] Inductive Biases in Field-Level Cosmological Inference from Galaxy Catalogs
链接: https://arxiv.org/abs/2609.09504
作者: James O. Baldwin,Shy Genel,Francisco Villaescusa-Navarro
类目: Cosmology and Nongalactic Astrophysics (astro-ph.CO); Machine Learning (cs.LG)
*备注: 23 pages, 8 figures, 5 tables. Accepted for publication in The Astrophysical Journal
Abstract:We perform field-level likelihood-free inference of the matter density parameter \Omega_m from simulated galaxy catalogs using machine learning models with differing inductive biases. Using hydrodynamic simulations from CAMELS, we examine how observable choice and architecture govern cosmological information extraction. We consider galaxy positions and line-of-sight peculiar velocities, separately and jointly, and compare permutation-invariant Deep Sets, implemented with either multilayer perceptrons (MLPs) or Kolmogorov-Arnold Networks (KANs), to graph neural networks (GNNs), which explicitly encode spatial relations. We test in-distribution and out-of-distribution (OOD) performance across simulations with different subgrid galaxy-formation prescriptions. Deep Sets infer \Omega_m from velocities alone with mean relative errors of approximately 18% in-distribution and \sim25% OOD, with KANs and MLPs achieving comparable performance. In contrast, the same set-based approach does not yield useful \sigma_8 predictions in either in-distribution or cross-suite tests. Adding positions does not improve Deep Sets, while GNNs infer \Omega_m with mean relative errors of about 10% in-distribution and 10 – 17% OOD. These results indicate that peculiar velocities provide the dominant source of \Omega_m information for set-based models in this setting, while spatial information is most effectively used by architectures that explicitly encode galaxy-galaxy relations. Because the velocity inputs are exact simulated peculiar velocities, applications to survey data will require validation under realistic velocity-measurement noise, selection effects, and survey geometry.
[LG-87] Gaussian Approximation for Multivariate Martingale Sums from Uniformly Ergodic Markov Chains
链接: https://arxiv.org/abs/2609.09480
作者: Yixuan Zhang,Qiaomin Xie
类目: Probability (math.PR); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 28 pages
Abstract:We develop Gaussian approximation bounds in higher-order Wasserstein distance W_p , p\geq2 , for sums of multivariate martingale differences generated by a uniformly ergodic Markov chain. Under an L^(2+\eta)p -moment condition with \eta0 , we establish the explicit bound O\left( p^3 |A|_4^2 + pd^1/4|A|_2^1/2|A|_4^2 \right) where A\in\mathbbR^n collects the L^(2+\eta)p -sizes of the n individual martingale increments. In the balanced-increment regime where the individual increments have comparable sizes of order n^-1/2 , it yields the first optimal O(n^-1/2) Gaussian approximation rate for fixed p and d . Consequently, we also obtain the first optimal O(n^-1/2) W_p Gaussian approximation rate for multivariate additive functionals of uniformly ergodic Markov chains. Our analysis develops two techniques for addressing the interplay between higher-order Wasserstein distance and temporal dependence. First, building on the Ornstein–Uhlenbeck relative-score approach of Fang and Koike (2023), we formulate the bound in terms of antisymmetric Stein couplings while retaining the conditional tensor structure. Second, we develop a refresh-then-maximal coupling that combines an independent first-step resampling, which preserves the desired Stein identity, with a subsequent maximal coupling that provides effective control of the coupling increment. These tools may be useful more broadly for Gaussian approximation under temporal dependence. Comments: 28 pages Subjects: Probability (math.PR); Machine Learning (cs.LG); Machine Learning (stat.ML) Cite as: arXiv:2609.09480 [math.PR] (or arXiv:2609.09480v1 [math.PR] for this version) https://doi.org/10.48550/arXiv.2609.09480 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-88] Mode Coverag e in Normalizing Flow Boltzmann Generators via Log-Ratio Variation
链接: https://arxiv.org/abs/2609.09473
作者: Qi Feng,Rongjie Lai,Di Qi,Xuda Ye
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Normalizing flow Boltzmann generators retain a tractable pushforward density, but training with forward KL depends on target samples that may be biased or omit modes. As a result, a flow can miss target mass while its observed importance weights give a high effective sample size. We introduce the log-ratio variation \X_\omega , the mean absolute pairwise difference of the target-to-pushforward log-density ratio under a weighting measure \omega , and use it to define KLXX, a new loss function. Two log-ratio variations are added to the forward KL (denoted by the two X’s): one weighted by the target to improve accuracy, the other by a mixture of quench and temper samples with pushforward samples to search candidate modes. We derive the Fisher–Rao gradient flow of KLXX, where both variations contribute nonpositive dissipation, and a fixed-surrogate error bound for KLXX. We use KLXX in an adaptive-staging Boltzmann generator, with importance reweighting at every stage. We bound the sampling error of its inference scheme when the stage weights are essentially bounded, and prove it asymptotically unbiased in the sample size. In the numerical tests, KLXX improves mode coverage over forward KL. It also improves the generator’s per-stage diagnostics against the loss that built the schedule. The observables the generator recovers are close to independent references. The log-ratio variations thus supply information that the forward KL loss usually omits.
[LG-89] A Block Tensor Train Burer-Monteiro Framework for Low-Rank Quantum State Tomography
链接: https://arxiv.org/abs/2609.09457
作者: Shakir Showkat Sofi,Charlotte Vermeylen,Fatemeh Mohammadi,Lieven De Lathauwer
类目: Quantum Physics (quant-ph); Machine Learning (cs.LG); Signal Processing (eess.SP); Numerical Analysis (math.NA); Optimization and Control (math.OC)
*备注: 26 pages
Abstract:Quantum state tomography is a fundamental technique for estimating the state of a quantum system from measured data and plays a crucial role in evaluating the performance of quantum devices. However, standard estimation methods become computationally prohibitive as the system size increases due to the exponential growth of the density matrix, describing a quantum state, with the number of qubits. We propose a low-rank tensor-network framework for mixed-state quantum state tomography based on a block tensor train (Block-TT) factorization. Specifically, the density matrix is represented as the contraction of a Block-TT with its Hermitian transpose, yielding a TT analogue of the Burer-Monteiro factorization. This parameterization guarantees Hermiticity and positive semidefiniteness by construction while compressing the number of optimization variables from exponential to linear in the number of qubits. Building on this representation, we develop single-site and two-site density matrix renormalization group (DMRG) algorithms for estimating quantum states from compressed measurements. The resulting methods operate directly on the compressed parameterization, support adaptive rank refinement, and exploit efficient tensor-network contractions for expectation-value evaluation. The framework is applicable to a broad class of low-rank quantum states, including pure states, nearly pure states, and ground states that admit accurate tensor-network approximations. Numerical experiments demonstrate accurate state reconstruction from limited measurements together with substantial reductions in memory requirements and computational cost compared with conventional low-rank tomography methods.
[LG-90] MiNCE: Nonparametric Strongly Consistent Confidence Envelopes for Band-Limited Functions and their Smoothed Spectra
链接: https://arxiv.org/abs/2609.09436
作者: Balázs Csanád Csáji,Bálint Horváth
类目: atistics Theory (math.ST); Machine Learning (cs.LG); Signal Processing (eess.SP); Methodology (stat.ME); Machine Learning (stat.ML)
*备注:
Abstract:Minimum-norm confidence envelope strategies offer a nonparametric approach to constructing nonasymptotic, simultaneous confidence regions for band-limited functions, exploiting the theory of Reproducing Kernel Hilbert Spaces (RKHS). While the finite-sample coverage guarantees of these envelopes have been established, their consistency has not been analyzed so far. In this paper, we study this construction, here termed the Minimum-Norm Confidence Envelope (MiNCE) framework, and establish the strong uniform consistency of the resulting bands, both for noise-free and noisy observation models, under mild assumptions on the measurement noises. We further extend this formulation to the frequency domain, deriving nonasymptotic, simultaneous, strongly uniformly consistent confidence bands for the smoothed spectra. Numerical experiments in nonparametric regression and spectral estimation empirically confirm our theoretical results, illustrating the contraction of the confidence envelopes toward the target function as the sample size increases.
[LG-91] Real-time and adaptive anomaly detection algorithm for cyclostationary models
链接: https://arxiv.org/abs/2609.09326
作者: Justyna Witulska,Tomasz Barszcz,Ireneusz Jabłoński,Agnieszka Wyłomańska
类目: Methodology (stat.ME); Machine Learning (cs.LG)
*备注: 16 pages
Abstract:This article introduces PeriodicCALM, an effective real-time anomaly detection framework designed for cyclostationary data streams. While classical cyclostationary processes feature periodically time-varying statistical properties, real-world signals often contain recurring impulsive components that conceal abnormal behavior. Existing real-time methods for struggle with these dynamics, frequently misinterpreting phase-dependent variability as non-cyclic anomalies and causing excessive false alarms. To address this, PeriodicCALM incorporates cycle-dependent variability to systematically ignore regular cyclic impulses while accurately isolating genuine anomalies. Operating in real time with continuous retraining capabilities, the method adapts dynamically to evolving signal characteristics. Comparative evaluations against the baseline CALM framework using simulated data demonstrate significant improvements in detection accuracy and training efficiency, alongside a reduction in prediction latency. Furthermore, the practical utility of PeriodicCALM is validated on real-world vibration signals collected from a compressor monitoring system.
[LG-92] “Transforming” LHCb: self-supervised maps of heavy-flavour decays
链接: https://arxiv.org/abs/2609.09275
作者: Marko Stamenkovic,Greg Landsberg
类目: High Energy Physics - Experiment (hep-ex); Machine Learning (cs.LG)
*备注: 31 pages, 14 figures
Abstract:Decays of beauty and charm hadrons provide sensitive probes of physics beyond the standard model, including decays with invisible particles, in which part of the final state leaves no reconstructed detector signature. The large heavy-flavour data samples recorded by the LHCb experiment at the CERN LHC, together with its precise tracking, displaced vertex reconstruction, and particle identification, make it particularly well suited to learning a map of reconstructed heavy-hadron decay environments directly from data. We propose to bring recent advances in jet flavour tagging at ATLAS and CMS to significantly improve on the performance of the current LHCb taggers and extend them to the reconstruction of heavy-flavour decays with several invisible particles in the final state. To achieve this, we introduce a self-supervised transformer architecture that learns the decay maps without flavour or exclusive-decay labels by inferring masked particle identification information and completing jets from which constituents have been removed. Across five classification tasks in simulated LHCb Open Data, the self-supervised model outperforms an otherwise identical transformer with random weights, and performs comparably to a fully supervised transformer. We achieve a tagging power of about 10%. In addition, removing constituents from reconstructed exclusive decays also systematically increases the model anomaly score relative to random removals from the same heavy hadrons. We confirm this behaviour directly in 2017 LHCb proton-proton collision Open Data: the score increases for all eight studied heavy-flavour channels, and the signal region response exceeds that in the adjacent sidebands. These studies provide a proof of principle that mapping heavy-flavour decay environments through jets can transform flavour tagging in LHCb and extend the discovery reach for incomplete or otherwise unusual decays.
[LG-93] nsor Network Moral Graph Recovery of Discrete Probability Distributions
链接: https://arxiv.org/abs/2609.09258
作者: Á. Troyano Olivas,Chi-Hang Fred Fung,Hans H. Brunner,Momtchil Peev,Vicente Martin
类目: Machine Learning (stat.ML); Information Theory (cs.IT); Machine Learning (cs.LG); Probability (math.PR); Quantum Physics (quant-ph)
*备注: 24 pages
Abstract:We present a method for recovering the moral graph of a causal DAG from a probability distribution over discrete variables, using fully connected tensor networks (FCTNs) with nuclear-norm-regularized bond corrections. Each bond matrix is parameterized as a baseline all-ones matrix plus a low-rank correction C_ij = U_ijV_ij^\top , and the nuclear norm of the correction implemented via the variational Frobenius norm penalty on the factors drives unnecessary bonds to zero. We prove that under faithfulness, positivity, and a no-implicit-rerouting assumption on the local tensor architecture, \textbfevery optimal FCTN with zero reconstruction error \varepsilon = 0 has effective graph exactly equal to the moral graph. For the approximate regime ( \varepsilon 0 ), we provide explicit recovery bounds using the Fannes-Audenaert continuity of conditional mutual information, and derive a sufficient condition on the regularization parameter \beta . The effective graph is read directly from the optimized bond matrices.
[LG-94] CAST: Canonical Approximate Schur Tree for Approximate Cholesky on Graphs
链接: https://arxiv.org/abs/2609.09255
作者: Meher Chaitanya,Cameron Musco,Aristides Gionis
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Numerical Analysis (math.NA)
*备注:
Abstract:Graph-data workloads such as diffusion estimation, ranking, semi-supervised learning, and network optimization often solve many Laplacian or symmetric diagonally dominant M-matrix (SDDM) systems with the same coefficient matrix. Approximate Cholesky preconditioners eliminate vertices one at a time and store the resulting sparse approximate factorization, the \emphfactor, whose construction cost is amortized across these solves. But eliminating a vertex, the \emphpivot, creates a dense Schur-complement clique among its d active neighbors. We introduce CAST (Canonical Approximate Schur Tree), which replaces this clique with a weighted random spanning tree sampled directly from it. Every realization is connected and contains exactly d-1 edges, while reweighting each selected edge by the reciprocal of its tree-inclusion probability makes the update unbiased. The distribution is independent of the ordering of the pivot neighbors, and we prove that its leverage-score marginals minimize the largest normalized reweighted-edge contribution among unbiased inverse-marginal one-tree estimators. We also introduce CAST- \rho , which replaces each pivot neighbor with \rho copies, each carrying a 1/\rho share of that neighbor’s incident weight, samples a weighted random spanning tree on the expanded clique, and contracts the copies back to the original neighborhood. The resulting update remains unbiased and connected, can be sampled exactly in O(\rho d) time, and satisfies a 1/\rho bound on the second moment of the normalized local Schur error. Increasing \rho therefore reduces certified local sampling variability, but may increase construction cost and downstream fill. Empirically, we observe that CAST-1 is the faster default, whereas CAST-2 is preferable when its additional edge contributions remain inexpensive.
[LG-95] Steering Diffusion Priors with Sparse Observations for High-Resolution Temperature Downscaling
链接: https://arxiv.org/abs/2609.09247
作者: Anirudh Avireddy,Manmeet Singh,Shivanshi Singh,Ayush Raj,Saptarishi Dhanuka,Parthasarathi Mukhopadhyay,Sandeep Juneja
类目: Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning (cs.LG)
*备注:
Abstract:Local heatwave hazard depends on fine-scale air temperature, but ground stations are sparse and reanalysis products such as ERA5 cannot resolve the terrain and land-surface contrasts that shape real heat exposure. We present a conditional diffusion emulator for high-resolution 2-m temperature downscaling, conditioned on static geography, a training climatology, exact-time ERA5 temperature, and solar and temporal features, guided at inference by score-based data assimilation (SDA): a differentiable Gaussian observation likelihood steers the diffusion score toward sparse revealed temperature observations without any retraining. On a controlled 32-case synthetic-grid protocol over AORC, guidance improves hidden-cell reconstruction over both ERA5 and a strong observation-proximal nearest-neighbor baseline once observation density reaches 1% (RMSE 0.318 vs.\ 0.431~K, winning all 32 cases), while sparser regimes still favor direct interpolation. We further map the full guidance-strength landscape across three observation densities, showing that the optimal strength shifts systematically with density and that over-guiding causes sharp, predictable degradation – giving a concrete operating recipe rather than a single untuned setting. The resulting fields are intended as a temperature layer for downstream heatwave-hazard products such as threshold exceedance and cumulative heat-burden. The present evidence is a controlled synthetic-grid validation; station-network and held-out-year evaluations are the next steps toward deployment.
[LG-96] Bayesian deep learning integration of geophysical and drilling data for 3D prediction of copper mineralization and drill targeting: a case study from the Kogodai prospect Rudny Altai
链接: https://arxiv.org/abs/2609.09246
作者: Margarita Veshchezerova,Egor Barashov,Evgenii Gusev,Michael R. Perelshtein,Arlan Kasymzhan,Bolat M. Kabaziev,Nurlan Y. Askarov
类目: Geophysics (physics.geo-ph); Machine Learning (cs.LG)
*备注: 28 pages, 14 figures, 2 tables. Submitted to Ore Geology Reviews
Abstract:Exploration drill targeting in structurally complex terranes is hindered by sparse sampling, heterogeneous datasets, and the ambiguity of geophysical inversions. Here, we present an uncertainty-aware 3D workflow for the acceleration of time-to-discovery in brownfield explorations and apply it to the Kogodai prospect in the Rudny Altai metallogenic province. We jointly analyse existing drilling and geophysical data in a comprehensive approach, revealing hidden patterns in already available data. Drillholes and trenches were desurveyed to a common 3D reference frame, and assays were composited to a consistent spatial support to facilitate joint modelling with geophysical inputs. We develop Bayesian deep-learning models to predict 3D fields of Cu grade together with chargeability and apparent resistivity while quantifying epistemic uncertainty via Monte Carlo sampling. The original contribution of this work is to treat the problem not as pointwise regression between co-located observations, but as joint learning of spatially continuous 3D fields from sparse, heterogeneous exploration evidence. The resulting 3D predictions delineate a principal mineralized trend and several localized candidate zones that coincide with elevated induced polarization (IP) responses, while uncertainty mapping highlights where predictions are robust versus where additional drilling would be most informative. The continuous Cu-grade field can also be thresholded to produce binary prospectivity maps, allowing the sensitivity of target delineation to the chosen cutoff to be evaluated. The outputs are intended for qualitative interpretation and risk-aware drill targeting rather than resource estimation, and we discuss key limitations arising from incomplete provenance metadata for geophysical products and heterogeneity of historical sampling.
[LG-97] A Subsampled Davis-Kahan Bound for Large-Scale Eigenspace Estimation
链接: https://arxiv.org/abs/2609.09211
作者: Huan Qing
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
*备注: 24 pages
Abstract:The Davis-Kahan theorem is a fundamental tool in spectral analysis, providing quantitative control over the distance between the eigenspaces of a symmetric matrix and its perturbation. However, when the matrix dimension is large, computing leading eigenvectors is computationally expensive, limiting the practical use of spectral methods in modern large-scale applications. This paper addresses this problem by proposing an independent Bernoulli sampling scheme and proves that the leading left singular vectors of the subsampled matrix faithfully approximate the target subspace of a low-rank symmetric matrix. Our main result is a subsampled Davis-Kahan bound that gives an explicit error bound depending directly on the sampling probability. The bound reveals the trade-off: the computational cost scales linearly with the sampling probability, while the statistical error scales as the inverse square root of the sampling probability. Our result thus extends the Davis-Kahan theorem to the subsampled setting, enabling scalable spectral analysis of large-scale symmetric matrices.
[LG-98] Algorithmic Optimality Guarantees for Nonsmooth H_infty Output-Feedback Policy Search
链接: https://arxiv.org/abs/2609.06197
作者: Ashkan Soleymani,Patrick Jaillet
类目: Optimization and Control (math.OC); Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: Appeared at The 65th IEEE Conference on Decision and Control (CDC), 2026
Abstract:We study continuous-time full-order dynamic output-feedback H_\infty policy search, a nonconvex and nonsmooth problem. Direct policy search is a central paradigm in reinforcement learning and continuous control, but rigorous guarantees remain scarce in robust output-feedback settings. The H_\infty problem is a canonical benchmark because it captures disturbance attenuation and robustness while exposing the hard nonsmooth geometry of policy-space optimization. We prove that on the exact identity-gauge slice of the extended convex lift, \varepsilon -stationarity yields O(\varepsilon) -suboptimality on compact exact slices, which in turn yields convergence-rate guarantees for nonsmooth policy-search methods. This result addresses the finite-time optimality-gap question raised by Guo and Hu [2022] in the more general dynamic output-feedback H_\infty policy-search setting. We further use the established value equivalence supplied by extended convex lifting to formulate a nonstrict-feasibility bisection method with one final strict-feasibility recovery step, yielding an explicit \varepsilon -optimal stabilizing controller. These results provide a quantitative and algorithmic strengthening of prior qualitative optimality theory for nonsmooth H_\infty policy search.
[LG-99] Bridging Theory and Data: Correcting Nuclear Mass Models with Interpretable Machine Learning
链接: https://arxiv.org/abs/2603.15203
作者: Yanhua Lu,Tianshuai Shang,Pengxiang Du,Jian Li,Haozhao Liang
类目: Nuclear Theory (nucl-th); Machine Learning (cs.LG); Nuclear Experiment (nucl-ex)
*备注: 10 pages, 4 figures
Abstract:Nuclear mass prediction is one of the core issues in nuclear physics research, yet it faces the challenge of small-sample datasets with high complexity. This study introduces the Kolmogorov-Arnold Network (KAN) into the refinement of nuclear mass models, proposing an efficient and interpretable solution. By constructing the KAN-WS4 hybrid model, the prediction accuracy is significantly improved (the root mean square error is reduced from 0.3 MeV to 0.16 MeV). Furthermore, leveraging the intrinsic interpretability of KAN, feature importance analysis reveals that the proton number is the most critical factor influencing residuals, indicating potential systematic biases in proton-related terms within existing theoretical models. The method’s generality is demonstrated across five mass models. This study shows that KAN provides a novel approach to small-sample, high-complexity scientific problems. Its interpretability facilitates the data-driven discovery of physical laws, promising broad applicability to key nuclear physics issues.
附件下载
