2026-09论文中文摘要
组合式安全失效在工具链演化中的识别与运行时监控
Compositional Safety Failures in Harness Evolution: Identification and Runtime Monitoring
arXiv 2609.33123 · 2026-09-30分层安全分布式线性可分计算：任意异构数据分配
Hierarchical Secure Distributed Linearly Separable Computation with Arbitrary Heterogeneous Data Assignment
arXiv 2609.33092 · 2026-09-30拜占庭鲁棒的联邦RAG：通过对齐校准与固定成员共形预测
Byzantine-Robust Federated RAG via Aligned Calibration and Fixed-Membership Conformal Prediction
arXiv 2609.33037 · 2026-09-30通过一阶方法获得光滑非凸-非凹极小极大问题的博弈驻点
Obtaining Game-Stationary Points for Smooth Nonconvex-Nonconcave Minimax Problems via First-Order Methods
arXiv 2609.32970 · 2026-09-30Relic：从多智能体协作到持久组织能力
Relic: From Multi-Agent Collaboration to Persistent Organizational Capability
arXiv 2609.32965 · 2026-09-30承诺-弃权回路：语言模型为何产生幻觉而非弃权
The Commit-Abstain Circuit: Why Language Models Hallucinate Instead of Abstaining
arXiv 2609.32964 · 2026-09-30多模态大语言模型在跨域组织学相似性判断中优于病理学基础模型
Multimodal LLMs Outperform Pathology Foundation Models in Cross-Domain Histological Similarity
arXiv 2609.32876 · 2026-09-30OpenTumorBoard：多学科肿瘤委员会讨论轨迹的真实世界基准
OpenTumorBoard: A Real-World Benchmark of Multidisciplinary Tumor Board Discussion Trajectories
arXiv 2609.32810 · 2026-09-30决策充分的状态表示：度量并减少写入时遗憾
Decision-Sufficient State Representations: Measuring and Reducing Write-Time Regret
arXiv 2609.32805 · 2026-09-30自适应一致性图用于长时程智能体
Adaptive Consistency Graph for Long-Horizon Agents
arXiv 2609.32754 · 2026-09-30GEM：用于模拟相互作用量子系统的ghost-Gutzwiller近似的实现
GEM: An implementation of the ghost-Gutzwiller approximation for simulating interacting quantum systems
arXiv 2609.32633 · 2026-09-30利用耦合流补全进行人-物交互建模
Harnessing Coupled Stream Completion For Human-Object Interaction Modeling
arXiv 2609.32551 · 2026-09-30RepoMAS：面向渐进式指定任务的议题驱动多智能体系统
RepoMAS: Solving Progressively Specified Tasks with Issue-Driven Multi-Agent Systems
arXiv 2609.32490 · 2026-09-30大型强子对撞机上快速且精确的带电粒子轨迹回归学习
Fast and Precise Learned Charged-Particle Trajectory Regression at the Large Hadron Collider
arXiv 2609.32479 · 2026-09-30AdaTutoRank：通过自适应辅导优化学习对文档集合进行重排序，用于RAG与深度研究
AdaTutoRank: Learning to Rerank Document Sets via Adaptive Tutoring Optimization for RAG and Deep Research
arXiv 2609.32472 · 2026-09-30对称线性 Datalog 中的保守 CSP
Conservative CSPs in Logspace are in Symmetric Linear Datalog (under complexity theoretic assumptions)
arXiv 2609.32463 · 2026-09-30DRAM：用于机器人操作策略的Delta规则递归联想记忆
DRAM: Delta-rule Recurrent Associative Memory for Robot Manipulation Policies
arXiv 2609.32453 · 2026-09-30横向双异质结中电可切换的一维四极激子
Electrically switchable one-dimensional quadrupolar excitons in lateral double heterojunctions
arXiv 2609.32392 · 2026-09-30$S_7$ 上的矩阵簇：三维稳定性与连续基数大小的子簇区间
Matrix varieties over $S_7$: dimension-three stability and continuum-sized subvariety intervals
arXiv 2609.32359 · 2026-09-30EyeVQA：从识别到空间定位的眼科视觉语言模型基准测试
EyeVQA: Benchmarking Ophthalmic Vision-Language Models from Recognition to Spatial Grounding
arXiv 2609.32352 · 2026-09-30可供性条件决策：弥合零样本跨楼层视觉与语言导航中的语义-空间鸿沟
Affordance-Conditioned Decision Making: Bridging the Semantic-Spatial Gap in Zero-Shot Cross-Floor Vision-and-Language Navigation
arXiv 2609.32292 · 2026-09-30ScentGen：用于分子气味描述生成的分层多模态嗅觉语义建模
ScentGen: Hierarchical Multimodal Olfactory Semantic Modeling for Molecular Odor Description Generation
arXiv 2609.32291 · 2026-09-30掺杂剂调制的晶格软化驱动β-FeSi2热电材料中热导率的显著降低
Dopant-modulated lattice softening drives drastic thermal conductivity reduction in β-FeSi2 thermoelectrics
arXiv 2609.32261 · 2026-09-30CoMemBench：跨多智能体工作流拓扑的协作记忆边界基准测试
CoMemBench: Benchmarking Collaborative Memory Boundaries across Multi-Agent Workflow Topologies
arXiv 2609.32192 · 2026-09-30残差去噪实现按需的样本高效多智能体协调
Residual Denoising Enables Sample-Efficient Multi-Agent Coordination on Demand
arXiv 2609.32129 · 2026-09-30TriO：面向万物感知的三模态无监督占用世界模型
TriO: Tri-Modal Unsupervised Occupancy World Model for Anything Perception
arXiv 2609.32013 · 2026-09-30油田数字孪生中传感器偏差归因于退化、天气或攻击：概率归因与基于成本的决策的仿真研究
Attributing Sensor Deviations to Degradation, Weather, or Attack in Oilfield Digital Twins: A Simulation Study of Probabilistic Attribution and Cost-Based Decisions
arXiv 2609.31973 · 2026-09-30高效提取有效 E-Graph
Efficient Extraction for Effectful E-Graphs
arXiv 2609.31887 · 2026-09-30稀疏重叠下的LLM裁判验证：从推断到设计
LLM Judge Validation Under Sparse Overlap: From Inference to Design
arXiv 2609.31857 · 2026-09-30平均镜像下降与对偶梯度方法：熵正则化Gromov-Wasserstein问题的收敛算法
Averaged Mirror Descent and Dual Gradient Methods: Convergent Algorithms for Entropic Gromov-Wasserstein Problems
arXiv 2609.31848 · 2026-09-30克尔黑洞周围的非线性电弱斯法勒龙云
Nonlinear electroweak sphaleron clouds around Kerr black holes
arXiv 2609.31845 · 2026-09-30游戏视频到音乐的生成
Video-to-Music Generation for Gameplay Videos
arXiv 2609.31810 · 2026-09-30视频扩散训练过程中的时间注意力头特化
Temporal-Attention Head Specialization During Video Diffusion Training
arXiv 2609.31654 · 2026-09-30OMP-MoE：基于正交匹配追踪的混合专家大语言模型高效专家剪枝
OMP-MoE: Efficient Expert Pruning for Mixture-of-Experts LLMs via Orthogonal Matching Pursuit
arXiv 2609.31631 · 2026-09-30无间隙差分隐私高斯数据主成分分析
Gap-free Differentially Private PCA for Gaussian Data
arXiv 2609.31614 · 2026-09-30双曲多项式的完全单调负幂
Riesz kernels of hyperbolic polynomials: positivity, admissible exponents and Jordan rigidity
arXiv 2609.31592 · 2026-09-30基于学习潜变量贝叶斯跟踪的在线学习
Online Learning via Learned Latent Bayesian Tracking
arXiv 2609.31559 · 2026-09-30锥与多面体交集问题的 ETH 紧致、构造性 FPT 算法
An ETH-Tight, Constructive FPT Algorithm for the Cone and Polytope Intersection Problem
arXiv 2609.31328 · 2026-09-30TaskIR：通过退化适应与任务反馈的任务驱动图像恢复
TaskIR: Task-Driven Image Restoration via Degradation Adaptation and Task Feedback
arXiv 2609.31170 · 2026-09-30DualManip：通过双路径语义推理与几何自适应的智能体动态操作
DualManip: Agentic Dynamic Manipulation via Dual-Path Semantic Reasoning and Geometric Adaptation
arXiv 2609.31112 · 2026-09-30廉价、开放智能体使大语言模型污染更难缓解
Cheap, open agents make LLM pollution harder to mitigate
arXiv 2609.31054 · 2026-09-30PipeDRAM：一种无数据转置的DRAM内架构，支持硬件/软件流水线
PipeDRAM: A Data-Transposition-Free Processing-Using-DRAM Architecture with Hardware/Software Pipelining
arXiv 2609.30998 · 2026-09-30实用确定性线性时间模子集和
Deterministic Linear-Time Modular Subset Sum
arXiv 2609.30992 · 2026-09-30表示感知的传输信息度量：用于非包含离散支撑集
Representation-Aware Transport-Information Measure for Non-inclusive Discrete Supports
arXiv 2609.30944 · 2026-09-30不可补全词与矩阵死亡性的二次界
Quadratic bounds for uncompletable words and matrix mortality
arXiv 2609.30817 · 2026-09-30XPhysICS：面向工业控制系统安全的跨物理域威胁落地
XPhysICS: Cross-Physical-Domain Threat Grounding for Industrial Control Systems Security
arXiv 2609.30805 · 2026-09-30高度倾斜环双星盘中对偏心双星系统的脉冲吸积
Pulsed Accretion onto Eccentric Binaries in Highly Misaligned Circumbinary Disks
arXiv 2609.30782 · 2026-09-30SEA-CLIP-Tiny：面向东南亚语言的高效多语言文本-视觉嵌入模型
SEA-CLIP-Tiny: Efficient Multilingual Text-Vision Embedding for Southeast Asian Languages
arXiv 2609.30739 · 2026-09-30完全兼容3GPP的LEO-ISAC远距离感知：一种窗口网格处理框架
Fully 3GPP-Compatible Long-Range Sensing for LEO-ISAC: A Window-Grid Processing Framework
arXiv 2609.30653 · 2026-09-30音频大语言模型知道自己何时听不到你
Audio LLMs Know When They Can't Hear You
arXiv 2609.30625 · 2026-09-30
