2026-09论文中文摘要
DiDrive：面向自动驾驶安全离线强化学习的风险感知分层扩散框架
DiDrive: A Risk-Aware Hierarchical Diffusion Framework for Safe Offline Reinforcement Learning in Autonomous Driving
arXiv 2609.01609 · 2026-09-03WMLLM：基于“预测后行动”世界建模的自进化优化智能体
WMLLM: Self-Evolving Optimization Agents via Predict-Then-Act World Modeling
arXiv 2609.01608 · 2026-09-03整数上的希尔伯特第十问题的十个未知数
Ten unknowns for Hilbert's tenth problem over the integers
arXiv 2609.01594 · 2026-09-03费米型暗物质吸收与LUX-ZEPLIN实验中的高能事件
Fermionic Dark Matter Absorption and the High-Energy Event in LUX-ZEPLIN
arXiv 2609.01592 · 2026-09-03基于熵的智能体选择性引导：从不完美的VLM教师学习自主策略
Selective Agent Guidance via Entropy: Learning Autonomous Policies from Imperfect VLM Teachers
arXiv 2609.01567 · 2026-09-03TAG-Bench：面向大型音频语言模型的时序音频定位基准测试
TAG-Bench: Benchmarking Temporal Audio Grounding in Large Audio Language Models
arXiv 2609.01542 · 2026-09-03重新思考离线数据驱动优化中的可学习性
Rethinking Learnability in Offline Data-driven Optimization
arXiv 2609.01493 · 2026-09-03关于子集族的可消去对
On cancellative pairs of families of subsets
arXiv 2609.01483 · 2026-09-03波兰测度空间中由成本诱导的Santaló型不等式
On cost-induced Santaló-type inequalities in Polish measure spaces
arXiv 2609.01460 · 2026-09-03MegaStyle++：通过分层风格定义扩展图像风格空间
MegaStyle++: Scaling Image Style Space through Hierarchical Style Definition
arXiv 2609.01423 · 2026-09-03神经符号几何抽象框架（NeuSOGA）：从观测到符号数学表示
Neuro-Symbolic Geometric Abstraction (NeuSOGA): From Observations to Symbolic Mathematical Representations
arXiv 2609.01408 · 2026-09-03利用SKA-Low重建异常空气簇射
Reconstruction of anomalous air showers with SKA-Low
arXiv 2609.01347 · 2026-09-03生产环境中的多臂老虎机：推理时的超参数优化
Bandits in Prod: Hyperparameter Optimization at Inference Time
arXiv 2609.01335 · 2026-09-03关于 pro-$\boldsymbol{\textit{C}}$ 群的可公度性
On commensurations of pro-$\mathcal{C}$ groups
arXiv 2609.01234 · 2026-09-03你的智能体上下文中有什么？针对AI智能体管控工具的上下文权限提升攻击
What's in Your Agent's Context? Context Privilege Escalation Attacks against AI Agent Harness
arXiv 2609.01222 · 2026-09-03基于JWST光谱的红移3.5-7.0低质量星系的物理条件：谱线比率随爆发性的变化
The Physical Conditions of Low-Mass Galaxies at z=3.5-7.0 from JWST Spectroscopy: The Behaviour of Spectral Line Ratios with Burstiness
arXiv 2609.01205 · 2026-09-03FinLifeBench：基于纵向银行对话的详尽生活事件历史与财务状态重建
FinLifeBench: Exhaustive Life-Event History and Financial-State Reconstruction from Longitudinal Banking Dialogue
arXiv 2609.01198 · 2026-09-03通过连续谱中的平带束缚态实现三维腔中的低阈值涡旋激光
Low-threshold vortex lasing in three-dimensional cavities via flatband bound states in the continuum
arXiv 2609.01191 · 2026-09-03一种用于基于里德堡原子量子接收机系统的新型空时编码架构
A Novel Space-Time Coding Architecture for Rydberg Atomic Quantum Receiver-Based Systems
arXiv 2609.01180 · 2026-09-03通过裂缝破解文生图模型：基于多智能体辩论导航异构安全过滤器
Jailbreaking Text-to-Image Models Through Cracks: Navigating Heterogeneous Safety Filters via Multi-Agent Debate
arXiv 2609.01168 · 2026-09-03作为特质方向漂移的阈下学习：SFT蒸馏下的机制与针对性控制
Subliminal Learning as Trait-Direction Drift: A Mechanism and Targeted Control under SFT Distillation
arXiv 2609.01091 · 2026-09-03随机网络上带动态边权重的统计语言竞争模型
Statistical Language Competition Model with Dynamic Edge Weighting on a Random Network
arXiv 2609.01078 · 2026-09-03让置信度改变，而非预测：用于事后校准的预测保留修复
Let Confidence Change, Not the Prediction: Prediction-Preserving Repair for Post-hoc Calibration
arXiv 2609.01072 · 2026-09-03近极端旋转黑洞熵的对数修正
Logarithmic corrections to the entropy of near-extremal rotating black holes
arXiv 2609.01039 · 2026-09-03中国科学技术大学-中国科学院微电子研究所（USTC-IME）低增益雪崩二极管（LGAD）在同步辐射光源聚焦X射线下的时间性能
Time performance of USTC-IME LGAD under synchrotron light source focused X-ray
arXiv 2609.00972 · 2026-09-03实时MIMO-OFDM SDR测试平台中基于轻量级CFR的调制自适应
Lightweight CFR-Based Modulation Adaptation in a Real-Time MIMO-OFDM SDR Testbed
arXiv 2609.00903 · 2026-09-03HELIOS：从午夜到正午，连续户外城市场景重光照
HELIOS: From midnight to noon, continuous outdoor urban scene relighting
arXiv 2609.00901 · 2026-09-03记忆的代价：一种经过校准的计算能量定律
The Price of Remembering: A Calibrated Energy Law for Computation
arXiv 2609.00744 · 2026-09-03超越关键点提取：结构化图像分类中鲁棒几何特征构建的框架
Beyond Landmark Extraction: A Framework for Robust Geometric Feature Construction in Structured Image Classification
arXiv 2609.00634 · 2026-09-03基于双节点蒙特卡洛树搜索的会话式推荐系统有效结构化上下文建模研究
Towards Effective Structured Context Modeling for Conversational Recommender Systems via Dual-node Monte Carlo Tree Search
arXiv 2609.00618 · 2026-09-03宇宙黎明时期超大质量双黑洞的纳赫兹引力波约束
Nanohertz Gravitational-Wave Constraints on Supermassive Binary Black Holes at Cosmic Dawn
arXiv 2609.00613 · 2026-09-03Hausdorff SDL空间的基数界
Cardinality Bounds for Hausdorff SDL Spaces
arXiv 2609.00599 · 2026-09-03基于输出重要性的残差稀疏化用于压缩混合专家大型语言模型
Residual Sparsification via Output Importance for Compressing Mixture-of-Experts LLMs
arXiv 2609.00575 · 2026-09-03VoiceLongMemEval：助手能否记住你的声音？
VoiceLongMemEval: Do Assistants Remember How You Sounded?
arXiv 2609.00570 · 2026-09-03基于见证算子谱范数的虚数见证增强
Imaginarity witnessing enhancement via spectral norms of witnesses
arXiv 2609.00534 · 2026-09-03为什么多层消息传递有效：图神经网络原子间势的完备性理论
Why Multi-Layer Message Passing Works: Completeness Theory for Graph Neural Network Interatomic Potentials
arXiv 2609.00528 · 2026-09-03沿$ab=1/q$的二阶B样条的框架集
The frameset of the second-order B-Spline along $ab=1/q$
arXiv 2609.00481 · 2026-09-03探索语言智能体与非语言智能体之间的协作
Exploring Collaboration between a language and a non-language agent
arXiv 2609.00474 · 2026-09-03QTEA：带稀疏残差显著权重与按列优化的三值大语言模型
QTEA: Ternary LLMs with Sparse Residual Salient Weight and By-Column Optimization
arXiv 2609.00224 · 2026-09-03序数门限、基数投注：将大语言模型（LLM）置信度匹配至金融决策算子
Ordinal Gates, Cardinal Bets: Matching LLM Confidence to the Financial Decision Operator
arXiv 2609.00187 · 2026-09-03否决变量：作为目标独立成本项的人类覆写机制
The Veto Variable: Human Override as a Goal-Independent Cost Term
arXiv 2609.00109 · 2026-09-03RAPIDMap：用于从卫星和街景图像进行可解释灾害制图的快速多智能体流水线
RAPIDMap: Rapid Multi-Agent Pipeline for Interpretable Disaster Mapping from Satellite and Street-view Imagery
arXiv 2609.00046 · 2026-09-03模糊局部降阶模型（fl-ROMs）
Fuzzy local reduced order models (fl-ROMs)
arXiv 2609.00031 · 2026-09-03OpenAgentFlow：为异构AI智能体集群提供系统级安全边界
OpenAgentFlow: Enabling System-Wide Safety Boundaries for Heterogeneous AI Agent Fleets
arXiv 2609.00015 · 2026-09-03基于公开数据的元订单建模与识别
Metaorder modelling and identification from public data
arXiv 2608.30999 · 2026-09-03CogEvol：面向高效且可靠的学习环境生成
CogEvol: Towards Efficient and Reliable Learning Environment Generation
arXiv 2608.30968 · 2026-09-03用DESI DR1目录测量本动速度和层析红偶极子
Measuring peculiar velocity and tomographic redshift dipole with DESI DR1 catalogs
arXiv 2608.30914 · 2026-09-03扭曲三层WSe$_2$和MoTe$_2$中的超莫尔重构与拓扑镶嵌
Supermoiré Reconstruction and Topological Mosaics in Twisted Trilayer WSe$_2$ and MoTe$_2$
arXiv 2608.30813 · 2026-09-03MoS₂单层红外瞬态吸收中的暗激子特征
Dark exciton signatures in the infrared transient absorption of MoS$_2$ monolayer
arXiv 2608.30504 · 2026-09-03两种“无”：金融文献中不显著结果的实际含义
Two Kinds of Nothing: What Insignificant Results in Finance Actually Show
arXiv 2608.30490 · 2026-09-03
