2026-10论文中文摘要
$C^1([0,1])$ 正单位球面上的范数可加映射
Norm-Additive Maps on the Positive Unit Sphere of $C^1([0,1])$
arXiv 2610.02862 · 2026-10-05遏制自主操作者：在Kubernetes上保护AI智能体的纵深防御框架与参考架构
Containing the Autonomous Operator: A Defense-in-Depth Framework and Reference Architecture for Securing AI Agents on Kubernetes
arXiv 2610.02861 · 2026-10-05反事实动作评估、观测瓶颈与联合嵌入预测世界模型中的表示几何
Counterfactual Action Evaluation, Observation Bottlenecks, and Representation Geometry in Joint-Embedding Predictive World Models
arXiv 2610.02860 · 2026-10-05从热到同调：全温度下精确量子吉布斯采样的谱隙传递
From Heat to Homology: Spectral Gap Transfer for Exact Quantum Gibbs Sampling at All Temperatures
arXiv 2610.02859 · 2026-10-05面向小型语言模型智能体的感知工具环境蒸馏
Harness-Aware Distillation for Small Language Model Agents
arXiv 2610.02858 · 2026-10-05三维单位球中临界 Neumann 问题的变号解
Sign-changing solutions for the critical Neumann problem in the three-dimensional unit ball
arXiv 2610.02857 · 2026-10-05自适应互蒸馏用于大语言模型平衡多任务后训练
Adaptive Mutual Distillation for Balanced Multi-Task Post-Training of Large Language Models
arXiv 2610.02856 · 2026-10-05J-PARC μ子到电子转换实验中低动量负μ子束的首次测量
First measurement of the low-momentum negative-muon beam for the muon-to-electron conversion experiment at J-PARC
arXiv 2610.02855 · 2026-10-05具有记忆效应的重复相互作用模型中的热化
Thermalization in a repeated-interaction model with memory effects
arXiv 2610.02854 · 2026-10-05有界可达性与基于收缩约束状态空间模型的越狱检测
Bounded Reachability & Jailbreak Detection via Contraction-Constrained State Space Models
arXiv 2610.02853 · 2026-10-05二元Zinbiel代数的幂零性与幂零子簇
Nilpotency and nil subvarieties of binary Zinbiel algebras
arXiv 2610.02852 · 2026-10-05ByteSplat：通过GPU内与GPU间通信减少实现高效分布式3D高斯溅射训练
ByteSplat: Efficient Distributed 3D Gaussian Splatting Training via Intra- and Inter-GPU communication reduction
arXiv 2610.02851 · 2026-10-05线性可检测条件下最优滤波器定量渐近稳定性的PDE方法
A PDE approach to quantitative asymptotic stability of the optimal filter in the linear detectable setting
arXiv 2610.02850 · 2026-10-05非局域非线性静电学：同号电荷吸引与非对称溶剂化
Nonlocal, nonlinear electrostatics: like-charge attraction and asymmetric solvation
arXiv 2610.02849 · 2026-10-05排列鲁棒性并不足够：多智能体Transformer策略中的动作坍缩
Permutation Robustness Is Not Enough: Action Collapse in Multi-Agent Transformer Policies
arXiv 2610.02848 · 2026-10-05开放团队多智能体强化学习中的换员正交信用分配
Turnover-Orthogonal Credit Assignment for Open-Team Multi-Agent Reinforcement Learning
arXiv 2610.02847 · 2026-10-05理解强化学习中的富集（Enrichment）
Understanding Enrichment in Reinforcement Learning
arXiv 2610.02846 · 2026-10-05稳定自主生成中的时间之箭
An Arrow of Time in Stable Autonomous Generation
arXiv 2610.02845 · 2026-10-05DNAlign：面向LLMs的动态零空间安全对齐
DNAlign: Dynamic Null-Space Safe Alignment for LLMs
arXiv 2610.02844 · 2026-10-05关于OFDM在非周期测距中的最优性
On the Optimality of OFDM for Aperiodic Ranging
arXiv 2610.02843 · 2026-10-05无限流评估何时能支持动态规划？兼容性、递归压缩与广义求和
When Can an Infinite-Stream Evaluation Support Dynamic Programming? Compatibility, Recursive Compression, and Generalized Summation
arXiv 2610.02842 · 2026-10-05多模态声明验证对LLM重写的鲁棒性如何？
How Robust Is Multimodal Claim Verification to LLM Rewriting?
arXiv 2610.02841 · 2026-10-05PointWAM：用于灵巧机器人操作的3D世界动作建模
PointWAM: 3D World Action Modeling for Dexterous Robotic Manipulation
arXiv 2610.02840 · 2026-10-05探索数据分布之外的新世界：系统行为、因果税与非因果基础模型
To Explore The Strange New World Beyond Data Distribution: System Behavior, Causality Tax, and Non-causal Base Model
arXiv 2610.02839 · 2026-10-05Axient：链上金融协议的清单绑定证据——七层推导、相关性、防篡改与可复现声明提升
Axient: Manifest-Bound Evidence for On-Chain Financial Protocols: Seven-Layer Derivation, Correlation, Tamper Rejection, and Reproducible Claim Promotion
arXiv 2610.02838 · 2026-10-05Transformer 应该回溯多远？音乐序列模型中的重复与复制
How Far Back Should a Transformer Look? Repetition and Copying in Music Sequence Models
arXiv 2610.02837 · 2026-10-05基于相关性的蒸馏产生更易合并的语音-音乐编码器
Correlation-Based Distillation Yields More Mergeable Speech-Music Encoders
arXiv 2610.02836 · 2026-10-05All Work And No Play Makes Jack a Dull Boy: 理解并防止RLVR中的灾难性策略崩溃
All Work And No Play Makes Jack a Dull Boy: Understanding and Preventing Catastrophic Strategy Collapse in RLVR
arXiv 2610.02835 · 2026-10-05Axient：杠杆事件市场的规范协议图组合：单一状态权威、原子组合、持久Saga与恰好一次恢复
Axient: Canonical Protocol-Graph Composition for Leveraged Event Markets: Single State Authority, Atomic Composition, Durable Sagas, and Exactly-Once Recovery
arXiv 2610.02834 · 2026-10-05解决 Erdős-Ulam 单色并封闭族猜想
Resolving Erdős-Ulam Monochromatic Union-Closed Family Conjectures
arXiv 2610.02833 · 2026-10-05FastOPD：面向轻量级VLA部署的在线策略蒸馏
FastOPD: On-Policy Distillation for Lightweight VLA Deployment
arXiv 2610.02832 · 2026-10-05AMBER：列表式视觉-语言重排序的多视图自适应预算分配
AMBER: Multi-View Adaptive Budget Allocation for Listwise Vision-Language Reranking
arXiv 2610.02831 · 2026-10-05关于电池储能系统在连续日内电力市场中的交易策略
On BESS-Backed Trading on the Continuous Intraday Electricity Market
arXiv 2610.02830 · 2026-10-05FSPO：预算受限的LLM强化学习后训练中的策略一致风险与帕累托可行控制
FSPO: Policy-Consistent Risk and Pareto-Feasible Control for Budgeted LLM RL Post-Training
arXiv 2610.02828 · 2026-10-05MLCommons 越狱基准 v1.0
MLCommons Jailbreak Benchmark v1.0
arXiv 2610.02827 · 2026-10-05复杂任务的缩放轨迹：通过递归自我重写
Scaling Trajectories for Complex Tasks through Recursive Self-Rewrite
arXiv 2610.02826 · 2026-10-05TerrainForge：面向反事实自动驾驶的物理基础道路几何编辑
TerrainForge: Physics-Grounded road geometry Editing for Counterfactual Autonomous Driving
arXiv 2610.02825 · 2026-10-05MetaRubric：学习为基于评分标准的强化学习进行奖励
MetaRubric: Learning to Reward for Rubric-Based Reinforcement Learning
arXiv 2610.02824 · 2026-10-05究竟是什么让基于相关性的SSL蒸馏具有噪声鲁棒性？一种机制性修正
What Actually Makes Correlation-Based SSL Distillation Noise-Robust? A Mechanistic Correction
arXiv 2610.02823 · 2026-10-05自适应谱-Koopman动力学建模用于时间域泛化
Adaptive Spectral-Koopman Dynamics Modeling for Temporal Domain Generalization
arXiv 2610.02822 · 2026-10-05使用预计算数值格林函数的快速纳米光子逆向设计
Fast Nanophotonic Inverse Design using Precomputed Numerical Green Functions
arXiv 2610.02821 · 2026-10-05通过内轫致辐射实现SIMP暗物质与辐射之间的动力学平衡
Kinetic Equilibrium between SIMP Dark Matter and Radiation via Internal Bremsstrahlung
arXiv 2610.02820 · 2026-10-05面向全模态推理的以文本为中心的后训练
Text-Centric Post-Training for Omni-Modal Reasoning
arXiv 2610.02819 · 2026-10-05RMCW：一种基于Reed--Muller码的语言模型抗删除水印
RMCW: A Deletion-Robust Watermark Based on Reed--Muller Codes for Language Models
arXiv 2610.02817 · 2026-10-05门控槽注意力-2：线性注意力中的双侧联想记忆修正
Gated Slot Attention-2: Two-Sided Associative Memory Correction in Linear Attention
arXiv 2610.02816 · 2026-10-05iS-KV：基于块增量SVD的在线低秩KV缓存压缩
iS-KV: Online Low-Rank KV Cache Compression via Block-Incremental SVD
arXiv 2610.02815 · 2026-10-05VeriPy：Python组件的源保留验证与兼容性检查
VeriPy Source-Preserving Verification and Compatibility Checking for Python Components
arXiv 2610.02814 · 2026-10-05寄存器路由延迟融合：在视觉运动模仿中重连易捷径的观测融合
Register-Routed Delayed Fusion: Rewiring Shortcut-Prone Observation Fusion in Visuomotor Imitation
arXiv 2610.02813 · 2026-10-05与有色人种年轻成年人（YOC）共同设计心理健康支持人工智能：需求、期望及对AI素养的启示
Co-Designing AI For Mental Health Support With Young Adults of Color (YOC): Needs, Expectations, and Implications for AI Literacy
arXiv 2610.02812 · 2026-10-05学习接触丰富操作中的反射行为
Learning Reflexive Behavior for Contact-Rich Manipulation
arXiv 2610.02811 · 2026-10-05
