2026-09论文中文摘要
ARC-Bench：闭环重规划掩盖了冻结JEPA世界模型中的动作排序缺陷
ARC-Bench: Closed-Loop Replanning Masks Broken Action Ranking in Frozen JEPA World Models
arXiv 2609.05461 · 2026-09-09通过制度整合实现Transformer中的动态窗口化用于金融时间序列
Dynamic Windowing in Transformers via Regime Incorporation for Financial Time Series
arXiv 2609.05460 · 2026-09-09将VGDL编译为因果模型
Compiling VGDL into Causal Models
arXiv 2609.05459 · 2026-09-09评估用于信赖域相场断裂的非线性消去预处理
Assessing Nonlinear Elimination Preconditioning for Trust-Region Phase-Field Fracture
arXiv 2609.05458 · 2026-09-09混凝土中的慢动力学：温度、强度变化和微裂缝损伤的影响
Slow dynamics in concrete: Effects of temperature, strength variation, and microcracking damage
arXiv 2609.05456 · 2026-09-09侵蚀诱发滑坡运动学的运动性研究
Kinematics of erosion-induced landslide mobility
arXiv 2609.05455 · 2026-09-09条件感知学习实现跨多种化学修饰与检测条件下寡核苷酸熔解行为的稳健预测
Condition aware learning enables robust prediction of oligonucleotide melting behavior across diverse chemistries and assay conditions
arXiv 2609.05454 · 2026-09-09慢涨落塑造胚胎干细胞在分岔附近的细胞信号传导
Slow fluctuations shape cell signaling near a bifurcation in embryonic stem cells
arXiv 2609.05453 · 2026-09-09从传感器数据到课堂探究：GenAI支持的学校数字孪生数据探索
From Sensor Data to Classroom Inquiry: GenAI-Supported Exploration of School Digital Twin Data
arXiv 2609.05452 · 2026-09-09ZetaDial：在推理时调节蛋白质结合剂的净电荷以实现治疗可开发性
ZetaDial: dialing net charge of protein binders at inference time for therapeutic developability
arXiv 2609.05451 · 2026-09-09SureRoute：面向无幻觉的逆合成自改进平台
SureRoute: Toward a Hallucination-Free Self-Improving Platform for Retrosynthesis
arXiv 2609.05450 · 2026-09-09三边出行-能源市场设计：作为多周期随机指派博弈
Three-sided mobility-energy market design as a multiperiod stochastic assignment game
arXiv 2609.05449 · 2026-09-09面向视觉与语言Transformer的损伤感知Bandit剪枝
Damage-Aware Bandit Pruning for Vision and Language Transformers
arXiv 2609.05448 · 2026-09-09超扩散二维超导体
Superdiffusive two-dimensional superconductors
arXiv 2609.05447 · 2026-09-09AutoFyn 技术报告：面向长时程智能体的非参数专家迭代
AutoFyn Technical Report: Non-Parametric Expert Iteration for Long-Horizon Agents
arXiv 2609.05446 · 2026-09-09利用固态纳米孔对分子条形码结构进行精确位置读出
Precise Positional Readout of Molecular Barcode Structures using Solid-State Nanopores
arXiv 2609.05445 · 2026-09-09COAST：面向碳感知HPC作业的拥塞感知开始时间推荐
COAST: Congestion-Aware Start-Time Recommendations for Carbon-Aware HPC Jobs
arXiv 2609.05443 · 2026-09-09记忆何时有用？工具使用型LLM智能体中长期记忆的成本感知评估
When Does Memory Help? A Cost-Aware Evaluation of Long-Term Memory in Tool-Using LLM Agents
arXiv 2609.05441 · 2026-09-09扩散模型中的资本注入与绝对连续股息支付优化
Optimization of capital injections and absolutely continuous dividend payments in a diffusion model
arXiv 2609.05440 · 2026-09-09CriticGen：生成感知评估作为可操作反馈
CriticGen: Generation-Aware Evaluation as Actionable Feedback
arXiv 2609.05439 · 2026-09-09超越对与错：评估大型语言模型中的二阶社会推理
Beyond Right and Wrong: Evaluating Second-order Social Reasoning in Large Language Models
arXiv 2609.05437 · 2026-09-09新型混合蛋白质支架缺口填充：加权机器学习集成、束搜索与质量约束重排序
Novel hybrid protein scaffold gap filling using weighted machine learning ensemble, beam search, and mass-constrained reranking
arXiv 2609.05436 · 2026-09-09反射倒向随机微分方程的深度学习：正则化与误差分析
Deep Learning for Reflected BSDEs: Regularization and Error Analysis
arXiv 2609.05434 · 2026-09-09股票永续期货：SpaceX 上市前市场
Perpetual Futures for Stocks: The SpaceX Pre-IPO Market
arXiv 2609.05433 · 2026-09-09伴侣AI与伦理设计：从系统故障与用户需求中学习
Companion AI and Ethical Design: Learning from System Failures and User Desires
arXiv 2609.05432 · 2026-09-09使用可编程货币实现自复制的种子盒服务器
Self-replicating seedbox servers using programmable money
arXiv 2609.05431 · 2026-09-09面向资产元数据信息的去中心化注册表
Towards Decentralized Registries for Assets Metadata Information
arXiv 2609.05430 · 2026-09-09使用RISC-V向量扩展加速OpenFOAM中的稀疏线性求解器
Accelerating Sparse Linear Solvers in OpenFOAM using RISC-V Vector Extensions
arXiv 2609.05429 · 2026-09-09网格检查点：一种用于推测性仿真的创新回滚支持
Grid Checkpointing: an Innovative Rollback Support for Speculative Simulation
arXiv 2609.05428 · 2026-09-09Constella：一种面向低地球轨道太空数据中心的高成本效益分布式AI推理框架
Constella: A Novel Framework for Cost-Efficient Distributed AI Inference in LEO Space Data Centers
arXiv 2609.05427 · 2026-09-09超越引理共享——属性定向可达性的新型并行化策略
Beyond Lemma Sharing -- Novel Parallelization Strategies for Property Directed Reachability
arXiv 2609.05426 · 2026-09-09测量驱动的单GPU LLM服务中主机CPU共置干扰的诊断与缓解（在多GPU服务器上）
Measurement-Driven Diagnosis and Mitigation of Host-CPU Co-location Interference in Single-GPU LLM Serving on a Multi-GPU Server
arXiv 2609.05425 · 2026-09-09具身智能业务算子NPU片内融合部署与优化研究
Research on Intra-Chip Fusion Deployment and Optimization of Embodied Intelligence Business Operator NPU
arXiv 2609.05424 · 2026-09-09看见而不理解：大型语言模型对移动用户界面质量、失败分类与架构解释的评估
Seeing Without Understanding: Large Language Model Evaluation of Mobile User Interface Quality, Failure Taxonomy, and Architectural Explanation
arXiv 2609.05423 · 2026-09-09STORE：面向无服务器函数的自供给存储即服务
STORE: Self-Provisioning Storage-as-a-Service for Serverless Functions
arXiv 2609.05422 · 2026-09-09迈向OGR(29)的形式化验证最优性证书：一份带小规模演示的SAT编码方法笔记
Toward a Formally Verified Optimality Certificate for OGR(29): A SAT-Encoding Methods Note with Small-Case Demos
arXiv 2609.05421 · 2026-09-09在单个 NVIDIA L20 上持续运行 70B 级 AWQ 推理：吞吐量、稳定性、能耗与质量表征
Sustained 70B-Class AWQ Inference on a Single NVIDIA L20: Throughput, Stability, Energy, and Quality Characterization
arXiv 2609.05420 · 2026-09-09使用 NIO Bench 对机器学习工作负载的存储系统进行基准测试
Benchmarking Storage Systems for Machine Learning Workloads Using NIO Bench
arXiv 2609.05418 · 2026-09-09稠密图中的Erdős–Sós猜想
The Erd\H os-Sós conjecture in dense graphs
arXiv 2609.05417 · 2026-09-09WorldSculpt：基于接地视频生成组合式世界
WorldSculpt: Generating Compositional Worlds from Grounded Videos
arXiv 2609.05416 · 2026-09-09Erdős–Sós猜想的极端情形
The extremal cases of the Erd\H os--Sós conjecture
arXiv 2609.05411 · 2026-09-09必要还是充分？基于行为证据评估大语言模型（LLM）的解释
Necessary or Sufficient? Evaluating LLM Explanations With Behavioural Evidence
arXiv 2609.05385 · 2026-09-09第三代探测器中长时引力波信号的快速贝叶斯推断
Fast Bayesian Inference for Long-Duration Gravitational-Wave Signals in 3G detectors
arXiv 2609.05375 · 2026-09-09限制效应可隐藏任意因果结构中的非物理对称性
Restricting the effects hides a nonphysical symmetry from every causal structure
arXiv 2609.05322 · 2026-09-09优化前先询问：交互式优化的动态预建模澄清
Ask Before You Optimize: Dynamic Pre-Formulation Clarification for Interactive Optimization
arXiv 2609.05258 · 2026-09-09LZ实验中的放热暗物质
Exothermic Dark Matter at LZ
arXiv 2609.05204 · 2026-09-09大语言模型能否预测个体对社会政策的行为响应？以中国灵活就业人员的养老金参保预测为例
Can Large Language Models Anticipate Behavioral Responses to Social Policies? A Case of Pension Enrollment Prediction among China's Flexible Workers
arXiv 2609.05189 · 2026-09-09TruthInsightBench：面向开放式科学发现智能体自动化评估的基于证据的基准
TruthInsightBench: An Evidence-Grounded Benchmark for Automated Evaluation of Open-Ended Scientific Discovery Agents
arXiv 2609.05079 · 2026-09-09SwanWeave：单阶段多任务指令引导的3D空间音频编辑
One-Stage Multi-Task Instruction-Guided 3D Spatial Audio Editing
arXiv 2609.04975 · 2026-09-09MINT：一种基于可扩展自我中心流水线监督的世界空间相机与手部运动估计统一模型
MINT: A Unified Model for World-Space Camera and Hand Motion Estimation from Scalable Egocentric Pipeline Supervision
arXiv 2609.04958 · 2026-09-09
