2026-09论文中文摘要
最适者的失败：无性生殖适应变化环境的极限
Failure of the fittest: limits on asexual adaptation to changing environments
arXiv 2609.35933 · 2026-09-30相同字节，不同权威：聊天模板提示注入中的保留令牌表示
Same Bytes, Different Authority: Reserved-Token Representations in Chat-Template Prompt Injection
arXiv 2609.35932 · 2026-09-30关于以AR同调球为边界的自旋$4$-流形的相交形式
On intersection forms of spin $4$-manifolds with AR homology sphere boundary
arXiv 2609.35931 · 2026-09-30Bregman共识
Bregman Consensus
arXiv 2609.35930 · 2026-09-30CADENCE：一种用于快速且准确时间序列分类的置信度自适应双专家网络
CADENCE: A Confidence-Adaptive Dual-Expert Network for Fast and Accurate Time Series Classification
arXiv 2609.35929 · 2026-09-30提示身份降低多智能体LLM系统中的合作性
Prompted Identity Degrades Cooperation in Multi-Agent LLM Systems
arXiv 2609.35928 · 2026-09-30规范损失景观导航：一种基于轨迹的增量学习遗忘缓解方法
Normative Loss Landscape Navigation: A Trajectory-Based Approach to Mitigating Forgetting in Incremental Learning
arXiv 2609.35926 · 2026-09-30实验室生长的硅酸镁太空尘埃类似物的光电子与电子显微镜研究
Photoelectron and electron microscopy investigation of laboratory grown Mg-silicate space dust analogues
arXiv 2609.35925 · 2026-09-30Grab a Coffee: 具有编译目标的离散扩散模型的未来感知引导
Grab a Coffee: Future-Aware Guidance for Discrete Diffusion with Compiled Objectives
arXiv 2609.35924 · 2026-09-30协动乘法过程中的转置奇算子手性：精确尾部级结构与可探测性障碍
Transpose-odd operator chirality in co-moving multiplicative processes: exact tail-level structure and a detectability obstruction
arXiv 2609.35923 · 2026-09-30几乎类人，除非关键之处：VoxParity 与声音应当改变的决策
Almost Human, Except When It Matters: VoxParity and the Decisions a Voice Should Change
arXiv 2609.35922 · 2026-09-30二次二值半径二元胞自动机的有限环障碍
Finite-ring obstructions for quadratic binary radius-two cellular automata
arXiv 2609.35921 · 2026-09-30Fishburn 潜在子集猜想的证明
Proof of Fishburn's latent-subset conjecture
arXiv 2609.35920 · 2026-09-30TEE Anchor：跨TEE组织背书以缓解TEE物理攻击
TEE Anchor: Cross-TEE Organizational Endorsement for Mitigating TEE Physical Attacks
arXiv 2609.35919 · 2026-09-30评估仅名称目录路由用于一次性代码搜索
Evaluating Name-Only Directory Routing for One-Shot Code Search
arXiv 2609.35918 · 2026-09-30道路级事故预测中的复现失败与平凡基线
Replication Failure and Trivial Baselines in Road-Level Crash Prediction
arXiv 2609.35917 · 2026-09-30近似容量受限最小闭系统的Chvátal--Gomory闭包
Approximating the Chvátal--Gomory Closure of Capacity-Bounded Min-Closed Systems
arXiv 2609.35915 · 2026-09-30智能体联邦学习：基于规则的客户端与服务器智能体用于自适应训练
Agentic Federated Learning: Rule-Based Client and Server Agents for Adaptive Training
arXiv 2609.35914 · 2026-09-30UNBIND：面向代码大语言模型的推理时定向引导遗忘
UNBIND: UNlearning By INference-time Directional Steering for Code LLMs
arXiv 2609.35913 · 2026-09-30MMSkillRisk：当多模态技能成为陷阱时，智能体能否保持安全？
MMSkillRisk: Can Agents Stay Safe When Multimodal Skills Become Traps?
arXiv 2609.35912 · 2026-09-30先学后用，信任后置：面向LLM智能体的序贯测试时学习
Learn Now, Use Next, Trust Later: Prequential Test-Time Learning for LLM Agents
arXiv 2609.35911 · 2026-09-30感知计算量的决策价值
The Decision Value of Perception Compute
arXiv 2609.35910 · 2026-09-30相似性并非有效性：防御LLM语义缓存投毒
Similarity Is Not Validity: Defending LLM Semantic Caches Against Poisoning
arXiv 2609.35908 · 2026-09-30网络外部性下的稳健机制设计
Robust Mechanism Design on Networks with Externalities
arXiv 2609.35907 · 2026-09-30双分支矢量量化辅助卫星数字语义通信与索引压缩用于AFDM上的高分辨率遥感图像
Dual-Branch Vector-Quantization-Aided Satellite Digital Semantic Communication with Index Compression for High-Resolution RSI Over AFDM
arXiv 2609.35906 · 2026-09-30惯性尺度下分数耗散的张量扩展
A Tensorial Extension of Fractional Dissipation at Inertial Scales
arXiv 2609.35905 · 2026-09-30复杂网页任务中的结构化交互、视觉定位与鲁棒执行：WebRetriever挑战赛技术报告
Structured Interaction, Visual Localization, and Robust Execution for Complex Web Tasks: A Technical Report on the WebRetriever Challenge
arXiv 2609.35904 · 2026-09-30维度量宇宙学演化的操作意义：用于跨纪元比较的广义宇宙学时间框架
Operational Meaning of the Cosmological Evolution of Dimensional Quantities: A Generalized Cosmological Time Framework for Cross-Epoch Comparisons
arXiv 2609.35903 · 2026-09-30提升中国青少年大语言模型安全性的标准：一个基于文化的细粒度基准
Raising the Bar for Chinese Adolescent LLM Safety: A Culturally-Grounded, Fine-Grained Benchmark
arXiv 2609.35902 · 2026-09-30自由加性卷积下的高阶凸序
Higher-order convex orders under free additive convolution
arXiv 2609.35901 · 2026-09-30重建隐性科学知识：通过端到端复现天文学研究评估LLM智能体
Reconstructing Implicit Scientific Knowledge: Evaluating LLM Agents through End-to-End Reproduction of Astronomy
arXiv 2609.35900 · 2026-09-30Wasserstein因果森林用于分布值结果
Wasserstein Causal Forests for Distribution-Valued Outcomes
arXiv 2609.35898 · 2026-09-30经验时代中的自我发现强化学习：学习历史是资产还是负担？
Self-discovering RL in the Era of Experience: Is Learning History an Asset or a Burden?
arXiv 2609.35897 · 2026-09-30矩形网格图中的最小分辨集：完整刻画与枚举
Minimal Resolving Sets in Rectangular Grid Graphs: A Complete Characterization and Enumeration
arXiv 2609.35896 · 2026-09-30广义Charlier递推系数：Hankel公式、Laguerre--Freud动力学与Painlevé V联系
Generalized Charlier Recurrence Coefficients: Hankel Formulas, Laguerre--Freud Dynamics, and Painlevé V Connections
arXiv 2609.35895 · 2026-09-30动力学方程的波粒分解 II：完整 Boltzmann 方程
Wave-Particle Decomposition for Kinetic Equations II: Full Boltzmann equation
arXiv 2609.35894 · 2026-09-30关于Bringmann、Han、Heim和Kane在eta-商中周期性符号变化的若干猜想的证明
Proofs of some conjectures by Bringmann, Han, Heim, and Kane on periodic sign changes in eta-quotients
arXiv 2609.35893 · 2026-09-30SaplingGuard：面向青少年与大语言模型安全交互的多维画像感知多智能体护栏
SaplingGuard: A Multidimensional-Profile-Aware Multi-Agent Guardrail for Developmentally Safe Adolescent-LLM Interaction
arXiv 2609.35892 · 2026-09-30制品提升控制模型：一个实施案例研究。在目标机器上构建 vs. 一次构建并提升制品
The Artifact Promotion Control Model: An Implementation Case Study. Build on Target Machines vs. Build Once and Promote Artifacts
arXiv 2609.35891 · 2026-09-30表征简单性与电路规模在阈值依赖方式上分离：通过对抗训练进行的受控测试
Representational Simplicity and Circuit Size Dissociate in a Threshold-Dependent Way: A Controlled Test via Adversarial Training
arXiv 2609.35890 · 2026-09-30SINGED：正确输出不能证明LLM智能体的安全执行
SINGED: Correct Outputs Do Not Certify Safe Execution in LLM Agents
arXiv 2609.35889 · 2026-09-30安全领导中的双重契合要求：现代组织中CISO角色履行的扎根理论研究
Dual-Fit Imperative in Security Leadership: A Grounded Theory Investigation of CISO Role Enactment in Modern Organisations
arXiv 2609.35887 · 2026-09-30Agentic Commerce Bench：衡量可花钱智能体的欺诈检测
Agentic Commerce Bench: Measuring Fraud Detection for Agents That Spend Money
arXiv 2609.35886 · 2026-09-30多智能体LLM中的集体状态：推理努力与通信拓扑的影响
Collective Regimes in Multi-Agent LLMs under Reasoning Effort and Communication Topology
arXiv 2609.35885 · 2026-09-30横截子空间中的学习：散度自由算子学习的最小表示
Learning in the Transverse Subspace: A Minimal Representation for Divergence-Free Operator Learning
arXiv 2609.35884 · 2026-09-30从静态策略到离线强化学习中的自适应先验
From Static Policies to Adaptive Priors in Offline Reinforcement Learning
arXiv 2609.35880 · 2026-09-30CruxBench：信息发现基准
CruxBench: A Benchmark of Information Discovery
arXiv 2609.35879 · 2026-09-30PoliVEM：一个用于计算固体力学的Python驱动虚拟元方法框架
PoliVEM: a Python-driven virtual element framework for computational solid mechanics
arXiv 2609.35878 · 2026-09-30BarcodeMAE+：重新思考DNA条形码基础模型的掩码预训练与全局表示
BarcodeMAE+: Rethinking Masked Pretraining and Global Representations for DNA Barcode Foundation Models
arXiv 2609.35877 · 2026-09-30基于EEG的癫痫发作预测的混合集成学习
Hybrid Ensemble Learning for EEG-Based Epileptic Seizure Forecasting
arXiv 2609.35876 · 2026-09-30
