2026-09论文中文摘要
Log-Euclidean Rényi 条件互信息
Log-Euclidean Rényi Conditional Mutual Information
arXiv 2609.34699 · 2026-09-29CMS分析流程在集成挑战中的研究
A study of CMS analysis pipelines through the Integration Challenge
arXiv 2609.34698 · 2026-09-29软-共线手征微扰理论用于大反冲下 $B \to \pi$ 形状因子
Soft-Collinear Chiral Perturbation Theory for $B \to π$ form factors at large recoil
arXiv 2609.34696 · 2026-09-29零阶奖励塑形对镇定控制中策略梯度的充分性
Sufficiency of Zeroth-Order Reward Shaping for Policy Gradient in Stabilization Control
arXiv 2609.34695 · 2026-09-29向量值子Bergman空间：值域包含与非刚性
Vector-Valued Sub-Bergman Spaces: Range Inclusions and Nonrigidity
arXiv 2609.34694 · 2026-09-29基于Juno/JADE与Juno/JEDI联合测量的木星极光区沉降电子能量分布
Energy distributions of precipitating electrons in Jupiter's auroral regions from combined Juno/JADE and Juno/JEDI measurements
arXiv 2609.34693 · 2026-09-29使用LLM检测LLM生成文本：跨代分析
Using LLMs to Detect LLM-Generated Texts: A Cross-Generation Analysis
arXiv 2609.34691 · 2026-09-29对称感知流匹配用于端到端分子晶体生成
Symmetry-Aware Flow Matching for End-to-end Molecular Crystal Generation
arXiv 2609.34690 · 2026-09-29Vilenkin系统上的向量值Littlewood--Paley--Rubio de Francia不等式
Vector-valued Littlewood--Paley--Rubio de Francia Inequalities on Vilenkin Systems
arXiv 2609.34689 · 2026-09-29统一轨迹匹配策略优化：多样化T2I生成与VLA泛化
Unified Trajectory Matching Policy Optimization: Diverse T2I Generation and VLA Generalization
arXiv 2609.34688 · 2026-09-29VCN-Bench：面向先验视觉经验空间推理的视频情境化导航基准
VCN-Bench: A Video-Contextualized Navigation Benchmark for Spatial Reasoning over Prior Visual Experience
arXiv 2609.34687 · 2026-09-29越狱上下文挥之不去：工具代理中的分歧安全路由及其跨任务可预测性
Jailbreak Context Lingers: Divergent Safety Routing and Its Cross-Task Predictability in Tool Agents
arXiv 2609.34686 · 2026-09-29P2D - 一种二维多航天器太阳风持续性模型
P2D - A Two-Dimensional Multi-Spacecraft Solar Wind Persistence Model
arXiv 2609.34685 · 2026-09-29自然状态预测精度可能掩盖VLA读出中较弱的受控响应性
Natural State-Prediction Accuracy can Hide Weak Controlled Responsiveness in VLA Readouts
arXiv 2609.34684 · 2026-09-29AgentPerfBench：智能体LLM推理性能的基准测试与评估套件
AgentPerfBench: A Benchmarking and Evaluation Suite for Inference Performance of Agentic LLMs
arXiv 2609.34683 · 2026-09-29V-Gym：通过技能-数据协同进化增强智能体视觉推理
V-Gym: Enhancing Agentic Visual Reasoning via Skill-Data Co-Evolution
arXiv 2609.34682 · 2026-09-29SOLAR：一种面向大语言模型预训练的状态驱动在线学习率调度器
SOLAR: A State-Driven Online Learning Rate Scheduler for LLM Pretraining
arXiv 2609.34681 · 2026-09-29QuantForge：发现MXFP4训练后量化的残差分解
QuantForge: Discovering Residual Decompositions for MXFP4 Post-Training Quantization
arXiv 2609.34680 · 2026-09-29从偏好到互惠：基于经验校准的LLM智能体建模的分散匹配
From Preference to Reciprocity: Decentralized Matching with Empirically Grounded LLM-agent Based Modeling
arXiv 2609.34679 · 2026-09-29公平事实核查：用RoSh缩小LLM事实判断中的跨语言差距
Fair Fact-Checking: Closing the Cross-Lingual Gap in LLM Factual Judgement with RoSh
arXiv 2609.34678 · 2026-09-29学习该回忆什么：用于世界模型的自适应多线索情景记忆
Learning What to Recall: Adaptive Multi-Cue Episodic Memory for World Models
arXiv 2609.34677 · 2026-09-29两色划分正则性毕达哥拉斯三元组的一个新证明
A new proof of two-color partition regularity of Pythagorean triples
arXiv 2609.34676 · 2026-09-29HOI-Retarget：面向人-物交互的接触中心重定向方法
HOI-Retarget: Contact-Centric Retargeting for Human-Object Interaction
arXiv 2609.34674 · 2026-09-29FestDPO：基于直接偏好优化的少步生成器对齐
FestDPO: Few-step Generator Alignment with Direct Preference Optimization
arXiv 2609.34673 · 2026-09-29证据对齐的多模态在线自蒸馏用于细粒度视觉理解
Evidence-Aligned Multimodal On-Policy Self-Distillation for Fine-Grained Visual Understanding
arXiv 2609.34672 · 2026-09-29DISCCO：基于距离的复杂对象贝叶斯空间聚类与节点脆弱性中心性
DISCCO: Distance-Based Bayesian Spatial Clustering of Complex Objects with Node-Frailty Centrality
arXiv 2609.34671 · 2026-09-29有向图上的层次聚类与信号去噪
Hierarchical Clustering and Signal Denoising on Digraphs
arXiv 2609.34670 · 2026-09-29欧拉方程的全离散活动通量方法：探索不同重构与新的多维演化算子
Fully Discrete Active Flux Methods for the Euler Equations: Exploring Different Reconstructions and New Multi-Dimensional Evolution Operators
arXiv 2609.34669 · 2026-09-29巨型流体天线系统（E-FAS）中的隐蔽通信：信道硬化对检测的影响
Covert Communications over Enormous Fluid Antenna Systems (E-FAS): Impact of Channel Hardening on Detection
arXiv 2609.34668 · 2026-09-29双时间尺度微调可证明学习两层ReLU网络的新特征
Two-Timescale Fine-tuning Provably Learns New Features for Two-Layer ReLU Networks
arXiv 2609.34667 · 2026-09-29可微物理优化在机器人材料操控中的数值可靠性研究
On the Numerical Reliability of Differentiable Physics-Based Optimization for Robotic Material Manipulation
arXiv 2609.34666 · 2026-09-29二元行动多智能体合约的复杂性景观定界
Settling the Complexity Landscape of Multi-Agent Contracts with Binary Actions
arXiv 2609.34665 · 2026-09-29有界函子上同调与比较定理
Bounded functor cohomology and comparison theorem
arXiv 2609.34664 · 2026-09-29工具等待与重新到达在编译期静态LLM服务中的成本机制与配置选择
Tool Waiting and Re-arrival in Compile-Time-Static LLM Serving: Cost Mechanisms and Configuration Selection
arXiv 2609.34663 · 2026-09-29基于漂移的无监督语音增强
Unsupervised Speech Enhancement via Drifting
arXiv 2609.34662 · 2026-09-29Codoku：面向前沿编码智能体的可再生程序推理挑战
Codoku: Renewable Program-Reasoning Challenges for Frontier Coding Agents
arXiv 2609.34661 · 2026-09-29深度最优的量子编译
Depth-Optimal Quantum Compilation
arXiv 2609.34659 · 2026-09-29Torch-PIM：面向PyTorch的自动化配置文件引导的PIM卸载
Torch-PIM: Automated Profile-Guided PIM Offloading for PyTorch
arXiv 2609.34657 · 2026-09-29具有观测动作的矩阵博弈中的极小极大最后迭代收敛
Minimax Last-Iterate Convergence in Matrix Games with Observed Actions
arXiv 2609.34656 · 2026-09-29AI工具在双钻工作流中的采用：设计师实践中的阶段、模式与障碍
AI Tools Adoption across the Double Diamond Workflow: Phase, Mode, and Barriers in Designer Practice
arXiv 2609.34655 · 2026-09-29蛋白质基础模型适应性预测的通用框架
A General Harness for Protein Foundation Model Fitness Prediction
arXiv 2609.34654 · 2026-09-29以视觉信息检索重振医学时间序列：一种视觉-语言视角
Revitalizing Medical Time Series with Vision-Informed Retrieval: A Vision-Language Perspective
arXiv 2609.34652 · 2026-09-29注意力头何时可以被静态定义？
When Can Attention Heads Be Statically Defined?
arXiv 2609.34650 · 2026-09-29超越技能进化：面向长时程智能体框架的自进化上下文管理策略
Beyond Skill Evolution: Self-Evolving Context Management Policies for Long-Horizon Agent Harnesses
arXiv 2609.34649 · 2026-09-29SEmoEdit：探测与利用预训练语音流的情感可编辑性
SEmoEdit: Probing and Harnessing the Editability of Pre-trained Speech Flows
arXiv 2609.34648 · 2026-09-29重子SU(2)多重态中的质量劈裂测量
Measuring Mass Splittings in Baryon SU(2) Multiplets
arXiv 2609.34647 · 2026-09-29循环片段：C11程序重试操作的有限事件结构与操作语义
Episodic Loops: Finitary Event Structures and Operational Semantics for C11 Programs with Retries
arXiv 2609.34646 · 2026-09-29Nereus：面向LLM后训练的自适应并行策略
Nereus: Adaptive Parallelism for LLM Post-Training
arXiv 2609.34645 · 2026-09-29光前全息QCD中轻赝标量和矢量介子的径向激发谱
Radial Excitation Spectra of Light Pseudoscalar and Vector Mesons in Light-Front Holographic QCD
arXiv 2609.34644 · 2026-09-29通用动态投资组合
Universal Dynamic Portfolios
arXiv 2609.34643 · 2026-09-29
