2026-10论文中文摘要
随机次梯度下降在尖锐排斥点处：一个有界噪声反例与高斯噪声规避
Stochastic Subgradient Descent at Sharply Repulsive Points: A Bounded Noise Counterexample and Gaussian Noise Avoidance
arXiv 2610.07777 · 2026-10-07SCSM：一种用于可迁移网站指纹识别的流量原生基础模型
SCSM: A Traffic-Native Foundation Model for Transferable Website Fingerprinting
arXiv 2610.07776 · 2026-10-07本征态量子自举中无有限精确性的收敛性
Convergence Without Finite Exactness in the Eigenstate Quantum Bootstrap
arXiv 2610.07775 · 2026-10-07读取而非操纵：利用路由器对数提升MoE视觉语言模型的多模态安全性
Reading, Not Manipulating: Leveraging Router Logits for Multimodal Safety in MoE Vision-Language Models
arXiv 2610.07774 · 2026-10-07在BESIII上搜索稀有粲衰变
Search for rare charm decays at BESIII
arXiv 2610.07773 · 2026-10-07基于模型的几何感知生成优化用于约束运动规划
Model-Based Geometry-Aware Generative Optimization for Constrained Locomotion Planning
arXiv 2610.07772 · 2026-10-07响应边际遗漏之处：功能后门恢复的自适应查询复杂度
What Response Marginals Miss: Adaptive Query Complexity of Functional Backdoor Recovery
arXiv 2610.07771 · 2026-10-07Clunie--Sheil-Small系数差猜想的反例：通过拟共形映射
A counterexample to the Clunie--Sheil-Small coefficient-difference conjecture via quasiconformal maps
arXiv 2610.07770 · 2026-10-07纤维化的半自由上链模型再探
Semifree Cochain Models for Fibrations Revisited
arXiv 2610.07769 · 2026-10-07有限图上量子游走的脉动
Pulsation of quantum walk on finite graph
arXiv 2610.07768 · 2026-10-07TRACE：面向MoE语言模型FP4强化学习的 rollout 引导量化感知训练
TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language Models
arXiv 2610.07767 · 2026-10-07Furstenberg–Sárközy 定理的一个新界：利用 van der Corput 性质
A new bound for the Furstenberg--Sárközy theorem using the van der Corput property
arXiv 2610.07765 · 2026-10-07无 Transformer 胜过六个协变量：基于童年作文的抑郁症状长期预测
No Transformer Beats Six Covariates: Long-Horizon Prediction of Depressive Symptoms from Childhood Essays
arXiv 2610.07764 · 2026-10-07ST-Bench：面向科学研究任务的多智能体系统生成的空间-时间基准
ST-Bench: A Spatial-Temporal Benchmark for Multi-Agent System Generation on Scientific Research Tasks
arXiv 2610.07763 · 2026-10-07ES-Trace：超越模型卡的代码生成模型伦理来源披露审计
ES-Trace: Auditing Ethical-Sourcing Disclosure of Code Generation Models Beyond Model Cards
arXiv 2610.07762 · 2026-10-07面向可解释推荐的方面表示对比学习
Contrastive Learning for Aspect Representation towards Explainable Recommendation
arXiv 2610.07761 · 2026-10-07金融文档问答中的令牌预算升级：成本可预测，收益是瓶颈
Token-Budgeted Escalation for Financial Document QA: Cost Is Predictable, Benefit Is the Bottleneck
arXiv 2610.07760 · 2026-10-07何时无状态恢复能击败拜占庭法定人数安全性？单步BFT的紧正规形式
When Can Stateless Recovery Defeat Byzantine Quorum Safety? A Tight Normal Form for Single-Step BFT
arXiv 2610.07759 · 2026-10-07延迟更好：分布偏移下视觉Transformer的令牌缩减
Later Is Better: Token Reduction for ViTs Under Distribution Shift
arXiv 2610.07758 · 2026-10-07获取并验证编码智能体的仓库规范
Acquiring and Verifying Repository Norms for Coding Agents
arXiv 2610.07757 · 2026-10-07使用AI评审进行可信方法比较：顺序、批次和聚合效应下的估计与设计
Trustworthy Method Comparison with AI Judges: Estimation and Design under Order, Batch, and Aggregation Effects
arXiv 2610.07755 · 2026-10-07对抗训练的线性Transformer是高斯混合模型的最优鲁棒上下文学习器
Adversarially Trained Linear Transformers Are Optimal Robust In-Context Learners for Gaussian Mixtures
arXiv 2610.07754 · 2026-10-07从证据到行动：工具使用型智能体为何失败
From Evidence to Action: How Tool-Using Agents Fail
arXiv 2610.07753 · 2026-10-07CoRE：通过单一策略学习协作角色专家实现去中心化协作操作
CoRE: Learning Collaboration-Role Experts for Decentralized Collaborative Manipulation with One Policy
arXiv 2610.07752 · 2026-10-07大型语言模型在噪声证据下的推理能力如何？一个主动视觉推理基准
How Well Do LLMs Reason with Noisy Evidence? An Active Visual Reasoning Benchmark
arXiv 2610.07751 · 2026-10-07PLADOS：一种用于Stiefel流形上分散优化的无参数着陆算法
PLADOS: A Parameter-free Landing Algorithm for Decentralized Optimization on the Stiefel Manifold
arXiv 2610.07750 · 2026-10-07半结构化表格上的经济高效数值问答：结构化、消解与规划
Cost-Effective Numerical QA over Semi-Structured Table: Structuring, Resolution, Planning
arXiv 2610.07749 · 2026-10-07当弱报告重要时：面向无人机协同感知的分阶段锚定融合
When Weak Reports Matter: Staged Anchored Fusion for Cooperative UAV Sensing
arXiv 2610.07747 · 2026-10-07因果菱形视界的热力学与力学定律
Thermodynamics and mechanical laws of causal-diamond horizons
arXiv 2610.07746 · 2026-10-07透过位移坐标系：视觉-力觉精密装配的特权噪声蒸馏
Seeing Through the Displaced Frame: Privileged Noise Distillation for Vision-Force Precision Assembly
arXiv 2610.07745 · 2026-10-07一种教学示范模型：可视化从在线交互到个性化推荐的路径
A Pedagogically Demonstrative Model Visualizing the Pathway from Online Interactions to Personalized Recommendation
arXiv 2610.07744 · 2026-10-07利用SKA天文台的甚长基线干涉测量
Very Long Baseline Interferometry with the SKA Observatory
arXiv 2610.07743 · 2026-10-07Cleave：通过解耦代数搜索与算子调度实现张量程序优化的扩展
Cleave: Scaling Tensor Program Optimization via Decoupled Algebraic Search and Operator Scheduling
arXiv 2610.07742 · 2026-10-07随机稳定室友问题可解概率的尖锐渐近性
Sharp Asymptotics for the Solvability Probability of Random Stable Roommates
arXiv 2610.07741 · 2026-10-07基于调和权重的高维在线校准
High-dimensional online calibration from harmonic weights
arXiv 2610.07740 · 2026-10-07引用你所探索的：基于可验证证据的预算感知LLM医学知识图谱推理
Cite What You Explore: Budget-Aware LLM Reasoning over Medical KGs with Verifiable Evidence
arXiv 2610.07739 · 2026-10-07相对论性Vlasov--Poisson系统动量矩的局部传播
Local Propagation of Momentum Moments for the Relativistic Vlasov--Poisson System
arXiv 2610.07738 · 2026-10-07多臂老虎机的纳什社会福利：轨迹级期望与高概率遗憾
Nash Social Welfare for Multi Armed Bandits: Trajectory-wise Expected and High Probability Regret
arXiv 2610.07737 · 2026-10-07两相调和映射热流在给定移动界面下的短时间存在性与唯一性
Short-time existence and uniqueness for two-phase harmonic map heat flow with a prescribed moving interface
arXiv 2610.07736 · 2026-10-07质量作为配对：在对称质量生成中，半代费米子上的显式马约拉纳质量不会跷跷板
Mass as pairing: an explicit Majorana mass on half of a generation does not seesaw in symmetric mass generation
arXiv 2610.07735 · 2026-10-07扭转双层石墨烯中的魔角理论：从非阿贝尔规范场到平带
Theory of Magic Angles in Twisted Bilayer Graphene: from Non-Abelian Gauge Fields to Flat Bands
arXiv 2610.07734 · 2026-10-07简单实心匹配覆盖图的最大边数
The maximum size of simple solid matching covered graphs
arXiv 2610.07733 · 2026-10-07谁在与智能体对话？多用户3D虚拟环境中的LLM
Who Is Talking to the Agent? LLMs in Multi-User 3D Virtual Environments
arXiv 2610.07732 · 2026-10-07在嵌入空间中通过强化学习学习检索
Learning to Retrieve via Reinforcement Learning in Embedding Space
arXiv 2610.07731 · 2026-10-07SanSi：一种用于系统1.5思维的循环类型化决策模型
SanSi: A Looped Typed Decision Model for System 1.5 Thinking
arXiv 2610.07730 · 2026-10-07Foveated Compression：面向令牌高效视觉语言模型的选择性高分辨率保留
Foveated Compression: Selective High-Resolution Preservation for Token-Efficient VLMs
arXiv 2610.07729 · 2026-10-07HiPLEX: 全双工语音语言模型的分层策略分解
HiPLEX: Hierarchical Policy Factorization for Full Duplex Speech Language Models
arXiv 2610.07727 · 2026-10-07结构感知的吞咽荧光透视检查关键点定位
Structure-aware Keypoint Localization for Videofluoroscopic Swallowing Study
arXiv 2610.07726 · 2026-10-07PERSIST：全双工口语对话中跨会话的谁-什么-何时记忆
PERSIST: Who-What-When Memory Across Sessions for Full-Duplex Spoken Dialogue
arXiv 2610.07725 · 2026-10-07整数格上受激随机游走的形状定理
Shape theorem for excited random walks on the integer lattice
arXiv 2610.07724 · 2026-10-07
