2026-10论文中文摘要
银河系近邻双胞胎（MWNeTs）。I. 一种识别银河系演化对应体的分层框架
Milky Way Near Twins (MWNeTs). I. A Hierarchical Framework for Identifying Candidate Evolutionary Counterparts of the Milky Way
arXiv 2607.08336 · 2026-10-06卡罗尔超引力
Carroll supergravities
arXiv 2607.08329 · 2026-10-06可解群与生成可解子群的概率
Supersoluble groups and the probability of generating a supersoluble subgroup
arXiv 2607.08165 · 2026-10-06NFTR：从可证明的模式平均到离线目标条件强化学习中的测地线子目标选择
NFTR: From Provable Mode-Averaging to Geodesic Subgoal Selection in Offline Goal-Conditioned RL
arXiv 2607.07855 · 2026-10-06Co-LMLM：连续查询有限记忆语言模型
Co-LMLM: Continuous-Query Limited Memory Language Models
arXiv 2607.07707 · 2026-10-06DNA 手柄偏置力依赖的环化时间
DNA handles bias force-dependent looping times
arXiv 2607.07561 · 2026-10-06重新审视确定性等价：欠驱动非线性系统中估计与控制的结构耦合
Revisiting Certainty Equivalence: The Structural Coupling Between Estimation and Control in Underactuated Nonlinear Systems
arXiv 2607.07276 · 2026-10-06多元均匀分布保持变换的随机反演
Stochastic Inversion of Multivariate Uniform-Distribution-Preserving Transformations
arXiv 2607.07174 · 2026-10-06TACoS：从涂鸦注释到精确分割的二维材料弱监督学习
TACoS: Weakly Supervised Learning of Two-Dimensional Materials from Scribble Annotations to Precise Segmentation
arXiv 2607.07169 · 2026-10-06超越人类水平的智能测量
Measuring Intelligence Beyond Human Scale
arXiv 2607.07040 · 2026-10-06驾驭效应：编排设计如何设定企业智能体人工智能的代币经济学
The Harness Effect: How Orchestration Design Sets the Token Economics of Enterprise Agentic AI
arXiv 2607.06906 · 2026-10-06用于推理泛化的几何自蒸馏
Geometric Self-Distillation for Reasoning Generalization
arXiv 2607.06855 · 2026-10-06UI2App：可执行Web应用程序生成中视觉交互推理的基准测试
UI2App: Benchmarking Visual Interaction Inference in Executable Web Application Generation
arXiv 2607.06306 · 2026-10-06通过量子相干实现经典可逆计算
Classical Reversible Computation by Quantum Coherence
arXiv 2607.06219 · 2026-10-06GEM-Occ：从视觉几何证据到具身语义占用记忆
GEM-Occ: From Visual Geometry Evidence to Embodied Semantic Occupancy Memory
arXiv 2607.05543 · 2026-10-06利用过程挖掘从软件工程过程记录生成人工智能代理
Using Process Mining to Generate AI Agents from Software Engineering Process Records
arXiv 2607.04948 · 2026-10-06由具有快速振荡噪声的物理布朗运动驱动的随机微分方程的逼近
Approximation of Random Differential Equations Driven by Physical Brownian Motion with Fast Oscillating Noise
arXiv 2607.04943 · 2026-10-06AgenticPD：用于物理设计QoR优化的阶段感知智能体框架
AgenticPD: A Stage-Aware Agentic Framework for Closed-Loop Physical Design Optimization
arXiv 2607.04758 · 2026-10-06LP-SFT：通过多模态熵结构进行局部保持监督微调
LP-SFT: Keeping Plausible Alternatives Alive in Supervised Fine-Tuning
arXiv 2607.04733 · 2026-10-06GitHub上的AI代理拉取请求：频率、结构和合并冲突率
AI Agent Pull Requests on GitHub: Frequency, Structure, and Merge Conflict Rates
arXiv 2607.04697 · 2026-10-06Mask2Real-WM：作为可控灵巧世界模型的模拟到现实桥梁的分割掩码
Mask2Real-WM: Controllable Dexterous World Models via Segmentation Masks as a Sim-to-Real Bridge
arXiv 2607.04546 · 2026-10-06三思而后行：将树搜索提炼为冻结视觉语言动作模型的动作评估
Look Before You Leap: Distilling Tree Search into Action Evaluation for Frozen VLA Models
arXiv 2607.03751 · 2026-10-06Xe~II、Rn~II和Og~II离子的能级和超精细结构计算
Calculation of the energy levels and hyperfine structure for Xe~II, Rn~II, and Og~II ions
arXiv 2607.03175 · 2026-10-06城市重建与人口再分布：来自关东大地震后东京的证据
Urban Reconstruction and Population Redistribution: Evidence from Tokyo after the Great Kanto Earthquake
arXiv 2607.02978 · 2026-10-06通过拉格朗日奖励增强实现安全推理时对齐
Safe Inference-Time Alignment via Lagrangian Reward Augmentation
arXiv 2607.02781 · 2026-10-06有限量子记忆下的最优稳定子测试与学习
Optimal Stabilizer Testing and Learning with Limited Quantum Memory
arXiv 2607.02444 · 2026-10-06以对象为中心的LeJEPA
Mask-supervised Object-centric Representation Learning with LeJEPA
arXiv 2607.02404 · 2026-10-06AnyGroundBench: 视觉语言模型中视频定位的专业领域基准
AnyGroundBench: A Multi-Domain Adaptation Benchmark for Video Grounding in VLMs
arXiv 2607.02269 · 2026-10-06潜意识时钟：扩散语言模型中的潜在时间建模
Subliminal Clocks: Latent Time Modelling in Diffusion Language Models
arXiv 2607.01774 · 2026-10-06Mastermind: 基于策略学习的仓库级漏洞复现
Mastermind: Strategy-grounded Learning for Repository-Scale Vulnerability Reproduction
arXiv 2607.01764 · 2026-10-06如何用量子环面的模糊环面逼近平坦谱三元组：一个扭曲的故事
How to approximate the flat spectral triple of a quantum torus by fuzzy tori : a twisted tale
arXiv 2607.01681 · 2026-10-06一种用于高台阶攀爬和转向的可重构摇臂-转向架机器人
A Reconfigurable Rocker-Bogie Robot for High Step Climbing and Turning
arXiv 2607.01554 · 2026-10-06语言模型真的能进行上下文检索吗？在百万Token规模的文档中淹没
Can Language Models Actually Retrieve In-Context? Drowning in Documents at Million Token Scale
arXiv 2607.01538 · 2026-10-06辐射过滤统一了活动星系核中的宽线现象学
Radiative filtering unifies broad-line phenomenology in active galactic nuclei
arXiv 2607.01479 · 2026-10-06办公室理解基准
Office Comprehension Benchmark
arXiv 2607.01245 · 2026-10-06消息传递实现高效推理
Message Passing Enables Efficient Reasoning
arXiv 2607.01077 · 2026-10-06带有厚吸积流的旋转Konoplya-Zhidenko黑洞的视界尺度强度与偏振图像
Horizon-scale intensity and polarization images of rotating Konoplya-Zhidenko black holes with thick accretion flows
arXiv 2607.01017 · 2026-10-06从角色到情节：基于角色的多智能体长篇小说生成
From Personas to Plot: Character-Grounded Multi-Agent Story Generation for Long-Form Narratives
arXiv 2607.00918 · 2026-10-06毫米波波束对齐的元迁移学习
Meta-Transfer Learning for mmWave Beam Alignment
arXiv 2607.00860 · 2026-10-06三量子比特非局域性悖论：超越GHZ
Three-qubit nonlocality paradoxes: beyond GHZ
arXiv 2607.00795 · 2026-10-06KineticXGPU：暗扇区自散射的张量化碰撞算子
KineticXGPU: A Tensorized Collision Operator for Dark-Sector Self-Scattering
arXiv 2607.00755 · 2026-10-06标题瓶颈模型
Caption Bottleneck Models
arXiv 2607.00578 · 2026-10-06EgoSafetyBench：用于评估具身VLM作为运行时安全卫士的诊断性自我中心视频基准
EgoSafetyBench: A Diagnostic Egocentric Video Benchmark for Evaluating Embodied VLMs as Runtime Safety Guards
arXiv 2607.00218 · 2026-10-06用物理振幅表示分波的左切割
The left-cut for partial waves in terms of physical amplitudes
arXiv 2607.00136 · 2026-10-06gCICY 三维流形的镜像对称 (I)：一个 $[3,-1]$ 块与非 Gorenstein 环面相
Mirror symmetry for gCICY threefolds (I): a $[3,-1]$ block and a non-Gorenstein toric phase
arXiv 2610.03719 · 2026-10-05暗物质束缚态简单化
Dark matter bound-states made easy
arXiv 2610.03718 · 2026-10-05少解码器，多编码器：从新视角合成中学习几何表示
Less Decoder is More Encoder: Geometric Representation Learning from Novel View Synthesis
arXiv 2610.03717 · 2026-10-05MoSE3：学习每个像素的世界空间SE(3)
MoSE3: Learning World-Space SE(3) at Every Pixel
arXiv 2610.03716 · 2026-10-054DCodeBench：动态场景逆向图形智能体基准测试
4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes
arXiv 2610.03715 · 2026-10-05SpiderCSS：可扩展的容错CSS态制备
SpiderCSS: Scalable Fault-Tolerant CSS State Preparation
arXiv 2610.03714 · 2026-10-05
