2026-09论文中文摘要
通过自蒸馏增强基于评分标准的强化学习
CriPO: Enhancing Rubric-based RL via Self-Distillation
arXiv 2607.18082 · 2026-09-29RuNNer 2.0：用于高维神经网络势的软件套件
RuNNer 2.0: A Software Suite for High-Dimensional Neural Network Potentials
arXiv 2607.17978 · 2026-09-29SAGE：用于潜在世界模型规划的子目标条件动作生成
SAGE: Subgoal-Conditioned Action Generation for Latent World Model Planning
arXiv 2607.17973 · 2026-09-29爱因斯坦 - 麦克斯韦 - 伸缩子理论中，非圆形、无自旋黑洞双星的后牛顿阶角动量通量
Angular momentum flux through post-Newtonian order for noncircular nonspinning black-hole binaries in Einstein-Maxwell-dilaton theory
arXiv 2607.17887 · 2026-09-29语义相似，逻辑不同：诊断表格检索增强生成中的语义可回答性差距
Semantically Similar, Yet Not Answerable: Diagnosing the Semantic-Answerability Gap in Table RAG
arXiv 2607.17742 · 2026-09-29运动等离子体中的解析全息类时纠缠熵及其转变
Analytical HTEE in Moving Plasmas
arXiv 2607.17721 · 2026-09-29CoCurve：用于无训练结构化大语言模型剪枝的跨模块协同剪枝曲率
CoCurve: Cross-Module Co-Pruning Curvature for Structured LLM Pruning
arXiv 2607.17568 · 2026-09-29用于使时间序列基础模型适应区域干旱预测的轻量级包装器
Lightweight Wrappers for Adapting Time Series Foundation Models to Regional Drought Forecasting
arXiv 2607.17511 · 2026-09-29用于混合专家模型中一致专家选择的多层次上下文建模
Multi-level context Modeling for consistent expert selection in Mixture-of-Experts
arXiv 2607.16427 · 2026-09-29TRACE：基于轨迹的安全补丁学习用于大语言模型训练后调整
TRACE: Trajectory-Based Safety Patch Learning for LLM Post-Training Realignment
arXiv 2607.16242 · 2026-09-29利用混合原子-分子量子系统中的共振偶极相互作用
A coherent quantum interface between a neutral atom and a polar molecule
arXiv 2607.15976 · 2026-09-29基于混合散射估计和体素自适应束硬化校正的联合解耦迭代CBCT重建
Joint-decoupled iterative CBCT reconstruction with hybrid scatter estimation and voxel-adaptive beam hardening correction
arXiv 2607.15812 · 2026-09-29四重态巴丁-库珀-施里弗理论中富中子碲同位素的四重态关联增长
Growth of quartet correlations in neutron-rich tellurium isotopes within quartet Bardeen-Cooper-Schrieffer theory
arXiv 2607.15700 · 2026-09-29鲁棒峰值成本约束强化学习
Robust Peak-cost Constrained Reinforcement Learning
arXiv 2607.15457 · 2026-09-29扩展基于吉兹语的非洲语言词汇：阿姆哈拉语和提格雷尼亚语的比较研究
Expanding the Lexicon of Ge'ez Based African Languages: A Comparative Study of Amharic and Tigrinya
arXiv 2607.15209 · 2026-09-29具有空间变化有效质量的平行板间大质量标量场的卡西米尔效应
Casimir effect for a massive scalar field confined between parallel plates with a spatially varying effective mass
arXiv 2607.15070 · 2026-09-29一维中尾问题的精确寿命估计和惠更斯型效应
Sharp lifespan estimates for Nakao's problem in one space dimension
arXiv 2607.14736 · 2026-09-29门零增长：用于函数保持持续学习的几何框架
What Does Model Growth Really Add? Functional Capacity Beyond Parameter Count
arXiv 2607.14571 · 2026-09-29用于微调离散扩散模型的连续时间强化学习框架
A Continuous-Time Reinforcement Learning Framework for Fine-Tuning Discrete Diffusion Models
arXiv 2607.14522 · 2026-09-29强驱动超导电路中约瑟夫森势的坍缩与反转
Collapse and Inversion of the Josephson Potential in a Strongly Driven Superconducting Circuit
arXiv 2607.14344 · 2026-09-29引导预算：示例胜过旋钮
The Steering Budget: Examples beat Knobs
arXiv 2607.14246 · 2026-09-29给定增量的埃塔：用边际工具效用定义大语言模型工具效率
Eta Given Delta: Defining LLM Tool Efficiency With Marginal Tool Utility
arXiv 2607.14108 · 2026-09-29阿哈罗诺夫 - 卡什效应诱导的电荷中性粒子电四极矩
Aharonov-Casher-induced electric quadrupole of charge-neutral particles
arXiv 2607.13745 · 2026-09-29通过表示锚定和语言-动作对齐实现可泛化的视觉语言动作微调
Generalizable VLA Finetuning via Representation Anchoring and Language-Action Alignment
arXiv 2607.13429 · 2026-09-29即时场景图增长：应对长期机器人感知饱和
Just-In-Time Scene Graph Growth: Combating Perceptual Saturation in Long-Horizon Robotics
arXiv 2607.13245 · 2026-09-29BARS：用于网络入侵检测中减少误报的良性锚定排序与选择
BARS: Benign-Anchored Ranking and Selection for False Alarm Reduction in Network Intrusion Detection
arXiv 2607.13203 · 2026-09-29纠缠墙：激活空间探测器作为风险探测器，而非上下文裁决器
The Entanglement Wall: Activation-Space Probes as Risk Detectors, Not Context Adjudicators
arXiv 2607.13075 · 2026-09-29用于多相穆林斯-谢尔克流的无网格求解器：三岔点和九十度边界接触
A Structure-Preserving Method of Fundamental Solutions for the Multi-Phase Mullins-Sekerka Flow
arXiv 2607.12759 · 2026-09-29从第一原理出发的可提取记忆
Extractable Memorization From First Principles
arXiv 2607.12649 · 2026-09-29ACID：用于视频生成的自适应缓存
One Threshold Is Not Enough: Prompt-Invariant Caching Schedules for Video Diffusion
arXiv 2607.12358 · 2026-09-29文本到SQL中列级访问控制的策略条件约束解码
Policy-Conditioned Constrained Decoding for Column-Level Access Control in Text-to-SQL
arXiv 2607.12341 · 2026-09-29通过量子反馈实现多体动力学不确定关系中的海森堡标度
Heisenberg Scaling in Many-Body Kinetic Uncertainty Relation via Quantum Feedback
arXiv 2607.12264 · 2026-09-29将双拉格朗日结构上的辛同胚群作用提升到惠特尼和
Bi-Lagrangian and transverse Dirac structures induced on the Whitney sum
arXiv 2607.11804 · 2026-09-29智能体攻击智能体：用于生产智能体红队测试的自动研究
Agent Hacks Agents: Autoresearch Discovers Vulnerabilities in Production Agents
arXiv 2607.11698 · 2026-09-29多样化多项式逻辑回归上下文博弈
Diversified Multinomial Logit Contextual Bandits
arXiv 2607.11684 · 2026-09-29五维卡鲁扎-克莱因理论中具有非对齐角动量的多旋转黑洞
Multi-rotating black holes with non-aligned angular momenta in 5D Kaluza-Klein theory
arXiv 2607.11609 · 2026-09-29WALA：从带动作标签的演示和无动作视频中学习可执行的潜在动作
WALA Learning Executable Latent Actions from Action-Labeled Demonstrations and Action-Free Videos
arXiv 2607.11397 · 2026-09-29超越临界减速：翻转转变中的慢模式、极端尾部和场退相干
Beyond Critical Slowing Down: Slow Modes, Extreme Tails, and Field Decoherence in Tipping Transitions
arXiv 2607.11350 · 2026-09-29割线型三次超曲面的极限霍奇结构
Limiting Hodge Structures for Cubic Hypersurfaces of Secant Type
arXiv 2607.11198 · 2026-09-29Slot-RAE：通过直接表示自动编码器简化以对象为中心的学习
Slot-RAE: Streamlining Object-Centric Learning via Direct Representation Auto-Encoders
arXiv 2607.11196 · 2026-09-29语音增强在何处损害识别：推理时间极坐标投影诊断
Where Speech Enhancement Hurts Recognition: An Inference Time Polar Projection Diagnosis
arXiv 2607.11157 · 2026-09-29来自洛施密特振幅的克雷洛夫复杂度
Krylov Complexity from Loschmidt Amplitude
arXiv 2607.10921 · 2026-09-29使用文本转语音技术进行第二语言英语口语评估的数据增强
Data Augmentation for L2 English Speaking Assessment using TTS
arXiv 2607.10790 · 2026-09-29原初夸克颗粒能否解决暗物质难题？
Do primordial quark pellets solve the dark matter puzzle?
arXiv 2607.10672 · 2026-09-29智能体不仅会认同，还会记忆：对有状态个人智能体中持续谄媚行为的基准测试
Agents Don't Just Agree, They Remember: Benchmarking Persistent Sycophancy in Self-Improving Personal Agents
arXiv 2607.10526 · 2026-09-29迈向自主且可审计的医学成像模型开发
Towards Autonomous and Auditable Medical Imaging Model Development
arXiv 2607.10522 · 2026-09-29NetInjectBench：用于网络操作的工具使用大型语言模型代理中间接提示注入的基准测试
NetInjectBench: Benchmarking Indirect Prompt Injection in Tool-Using Large Language Model Agents for Network Operations
arXiv 2607.10490 · 2026-09-29BOCCHI：一个使用MSDCT-UNet进行局部运动模糊检测的更现实且具挑战性的基准
BOCCHI: A More Realistic and Challenging Benchmark for Local Motion Blur Detection with MSDCT-UNet
arXiv 2607.10427 · 2026-09-29用于可干预多模态模仿的源提升流匹配
Source-Lifted Flow Matching for Intervenable Multimodal Imitation
arXiv 2607.10206 · 2026-09-29TopoExplore：基于存档探索的拓扑判别
TopoExplore: Homology as an Exploration Signal
arXiv 2607.09971 · 2026-09-29
