2026-10论文中文摘要
一个反向里斯估计与谱隙相结合蕴含庞加莱不等式
A reverse Riesz estimate combined with a spectral gap implies a Poincaré inequality
arXiv 2607.08322 · 2026-10-05AutoPersonas：用于开放式角色演变的多时间尺度循环引擎
A Multi-Timescale Recursive Self-Improvement Engine for Open-Ended Persona Growth
arXiv 2607.08252 · 2026-10-05超越条件数的更快量子线性系统求解器
Faster quantum linear system solver beyond the condition number
arXiv 2607.07691 · 2026-10-05基于生物信息神经网络的可靠机械算子恢复：架构与优化设计原理
Reliable mechanistic operator recovery with biologically-informed neural networks: principles for architecture and optimisation design
arXiv 2607.07425 · 2026-10-05朗道准粒子的低能有效哈密顿量：II. 应用于接触费米气体
A low-energy effective Hamiltonian for Landau quasiparticles: II. Application to the contact Fermi gas
arXiv 2607.07041 · 2026-10-05矩阵加权网络的子空间一致性
Subspace Consensus
arXiv 2607.06970 · 2026-10-05ELSA3D：用于统一3D理解与生成的弹性语义锚定
ELSA3D: Elastic Semantic Anchoring for Unified 3D Understanding and Generation
arXiv 2607.06565 · 2026-10-05量子信道多项式处理
Quantum Channel Polynomial Processing
arXiv 2607.06557 · 2026-10-05计算多项式系统的奇异解：迈向无收缩的超线性收敛
A superlinear arclength endgame for corank-one singular polynomial systems
arXiv 2607.06329 · 2026-10-05SovereignPA-Bench：在演化意图、平台中介与同意约束下评估用户自有个人智能体
SovereignPA-Bench: Evaluating User-Owned Personal Agents under Evolving Intent, Platform Mediation, and Consent Constraints
arXiv 2607.05363 · 2026-10-05用于优化的混合精度牛顿法
Mixed precision Newton's method for optimization
arXiv 2607.04828 · 2026-10-05规定高斯曲率方程反问题中的规范刚性
Gauge rigidity in an inverse problem for the prescribed Gaussian curvature equation
arXiv 2607.04657 · 2026-10-05通过扰动的基于接地感知敏感性检测检索增强生成中的幻觉（GASP）
Sentence-Level Context Sensitivity as a Training-Free Detector of Unsupported Content, Evaluated Against Trained Verifiers
arXiv 2607.04223 · 2026-10-05主权谈判基准：在隐私、同意、证据和制度压力下评估委托谈判中用户拥有的个人代理
SovereignNegotiation-Bench: Evaluating User-Owned Personal Agents In Delegated Bargaining Under Privacy, Consent, Evidence, And Institutional Pressure
arXiv 2607.02814 · 2026-10-05被叙事诱惑：评估半开放文本沙盒中的规则遵守情况
Assessing Rule Adherence of LLM Adjudicators in Call of Cthulhu TRPG
arXiv 2607.02802 · 2026-10-05双强子碎裂框架下质子-质子碰撞中的横向自旋依赖的能量-能量关联子
Transverse-spin dependent energy-energy correlators in proton-proton collisions within the dihadron fragmentation framework
arXiv 2607.02441 · 2026-10-05更密集 ≠ 更好：持续后训练中同策略自蒸馏的局限性
Denser $\neq$ Better: Limits of On-Policy Self-Distillation for Continual Post-Training
arXiv 2607.01763 · 2026-10-05面向样本生成模型的决策感知训练
Decision-Aware Training for Sample-Based Generative Models
arXiv 2607.01171 · 2026-10-05超流氦中量子涡旋晶格上的惯性波和开尔文波可视化
Visualization of Inertial and Kelvin Waves on the Quantum Vortex Lattice in Superfluid Helium
arXiv 2607.00821 · 2026-10-05ClarifyCodeBench: 评估大语言模型在代码生成中澄清模糊需求的能力
ClarifyCodeBench: Evaluating LLMs on Clarifying Ambiguous Requirements for Code Generation
arXiv 2607.00711 · 2026-10-05理解语言模型为何产生幻觉：针对先验知识的推理测试
Understanding Why Language Models Hallucinate: Testing Reasoning Against Priors
arXiv 2607.00447 · 2026-10-05耦合复制-突变动力学中的实现滤波器与延迟预算不稳定性
Same Mean Lag, Different Stability: Periodic Implementation in Adaptive Competition
arXiv 2607.00227 · 2026-10-05摩尔、埃舍尔、彭罗斯：共形黄金辫
Moore, Escher, Penrose: A Conformal Golden Braid
arXiv 2610.02210 · 2026-10-02德西特空间中算符和态的领先引力修饰
Leading gravitational dressing of operators and states in de Sitter space
arXiv 2610.02209 · 2026-10-02一个基元驱动所有：高斯混合形状蒸馏用于实时虚拟化身
One Basis to Animate Them All: Gaussian Blendshape Distillation for Real-Time Avatars
arXiv 2610.02207 · 2026-10-02KaliBench：面向Kali Linux网络安全工具使用的细粒度基准，具有免运行时可验证奖励
KaliBench: A Fine-Grained Benchmark for Cybersecurity Tool Use on Kali Linux with Runtime-Free Verifiable Rewards
arXiv 2610.02206 · 2026-10-02重建、练习、实战：具身智能体的引导式自我改进
Reconstruct, Practice, Go Real: Guided Self-Improvement for Embodied Agents
arXiv 2610.02204 · 2026-10-02嵌入预测有助于图像生成
Embedding Prediction Helps Image Generation
arXiv 2610.02203 · 2026-10-02ScholarCatalyst：一个用于检索激发新研究论文的基准
ScholarCatalyst: A Benchmark for Retrieving Papers That Inspire New Research
arXiv 2610.02202 · 2026-10-02SILSA：用于保持拓扑的高分辨率3D生成的滑动窗口切片潜变量
SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation
arXiv 2610.02201 · 2026-10-02VISTA：交互世界中的视觉推理驾驭框架
VISTA: A Visual Harness for Reasoning in an Interactive World
arXiv 2610.02200 · 2026-10-02FERPO：前向熵正则化策略优化
FERPO: Forward Entropy-Regularized Policy Optimization
arXiv 2610.02198 · 2026-10-02HiPhy：面向物理合理多原理视频生成的层次对齐方法
HiPhy: Hierarchical Alignment for Physically-Plausible Multi-Principle Video Generation
arXiv 2610.02197 · 2026-10-02图上的代价增强薛定谔桥是精确可解的：Feynman-Kac 倾斜替代学习控制
Cost-augmented Schrödinger bridges on graphs are exactly solvable: a Feynman-Kac tilt replaces learned control
arXiv 2610.02195 · 2026-10-02SW1D：用于一维弹性介质中地震面波解的Python软件包
SW1D: A Python package for seismic surface wave solutions in 1-D elastic media
arXiv 2610.02194 · 2026-10-02分层连续扩散语言模型
Hierarchical Continuous Diffusion Language Models
arXiv 2610.02193 · 2026-10-02具有质量无序的双曲晶格：相与相变
Hyperbolic lattices with mass disorder: Phases and phase transitions
arXiv 2610.02192 · 2026-10-02缺失的原语：诊断与修复大语言模型中的数学推理
The Missing Primitive: Diagnosing and Repairing Mathematical Reasoning in Large Language Models
arXiv 2610.02191 · 2026-10-02信任方向，搜索步长：用于大语言模型微调的零阶与一阶混合方法
Trust the Direction, Search the Step: Zero-and-First-Order Methods for LLM Fine-Tuning
arXiv 2610.02190 · 2026-10-02通过强化稀疏自编码器特征的内在无序蛋白区域生成建模
Generative modeling of intrinsically disordered protein regions by reinforcing sparse autoencoder features
arXiv 2610.02189 · 2026-10-02DMAD：分布匹配作为对抗蒸馏用于快速视觉生成
DMAD: Distribution Matching as Adversarial Distillation for Fast Visual Generation
arXiv 2610.02188 · 2026-10-02鲁棒稳定的最小实验：信息、谱几何与持续时间
Minimal Experiments for Robust Stabilization: Information, Spectral Geometry, and Duration
arXiv 2610.02187 · 2026-10-02用于化学中生成模型和基础模型的高阶分子文法
Higher-Order Molecular Grammars for Generative and Foundation Models in Chemistry
arXiv 2610.02186 · 2026-10-02更好地解码循环Transformer（几乎免费）
Decoding Looped Transformers Better for (Almost) Free
arXiv 2610.02185 · 2026-10-02反应式系统的充分理由与解释
Sufficient Reasons and Explanations for Reactive Systems
arXiv 2610.02184 · 2026-10-02单粒子谱估计
Single-Particle Spectral Estimation
arXiv 2610.02183 · 2026-10-02SoftServe：一种用于深度学习的高可扩展拟牛顿方法
SoftServe: A Scalable Quasi-Newton Method for Deep Learning
arXiv 2610.02182 · 2026-10-02OmniSeek：面向多轮音视频推理的原生工具集成
OmniSeek: Native Tool Integration for Multi-turn Audio-Visual Reasoning
arXiv 2610.02181 · 2026-10-02生成式电影摄影师：在3D中编排相机与物体运动
Generative Cinematographer: Composing Camera and Object Motion in 3D
arXiv 2610.02180 · 2026-10-02从梯度到能力：理解多教师同策略蒸馏
From Gradients to Capabilities: Understanding Multi-Teacher On-Policy Distillation
arXiv 2610.02179 · 2026-10-02
