2026-10论文中文摘要
一个工具，一种品味？氛围编码如何以集体多样性换取个人创造力
One Tool, One Taste? How Vibe Coding Trades Collective Diversity for Individual Creativity
arXiv 2609.38183 · 2026-10-01EmAvatar：通过冲突消解与表现力引导的多模态共情回复生成
EmAvatar: Multimodal Empathetic Response Generation via Conflict Resolution and Expressive Guidance
arXiv 2609.38182 · 2026-10-01大型语言模型是近似生存估计器
Large Language Models are Approximate Survival Estimators
arXiv 2609.38181 · 2026-10-01VideoLoop：循环工作记忆对抗长视频智能体中的语义抖动
VideoLoop: Looped Working Memory Against Semantic Thrashing in Long-Form Video Agents
arXiv 2609.38119 · 2026-10-01学习量子对称性
Learning quantum symmetries
arXiv 2609.38085 · 2026-10-01RS-OPSD：面向超高分辨率遥感视觉问答的可靠特权在线策略自蒸馏
RS-OPSD: Reliable Privileged On-Policy-Self-Distillation for Ultra-High-Resolution Remote Sensing VQA
arXiv 2609.38072 · 2026-10-01VADER：具有声明性召回率的过滤向量搜索
VADER: Filtered Vector Search with Declarative Recall
arXiv 2609.38053 · 2026-10-01性别偏见在大语言模型中普遍存在且高度异质
Gender bias across LLMs is common and highly heterogeneous
arXiv 2609.38036 · 2026-10-01KV-Kaizen：学习上下文自适应的缓存压缩选择
KV-Kaizen: Learning Context-Adaptive Cache Compression Choices
arXiv 2609.37988 · 2026-10-01SelfSearch：面向自我改进智能体的无奖励搜索
SelfSearch: Reward-Free Search for Self-Improving Agents
arXiv 2609.37968 · 2026-10-01看得更细：面向AI生成图像检测的块级监督
Look Closer: Patch-wise Supervision for AI-Generated Image Detection
arXiv 2609.37937 · 2026-10-01从像素到按键：探索游戏逆向动力学中的空间与运动线索
Pixels to Keys: Exploring Spatial and Motion Cues in Gameplay Inverse Dynamics
arXiv 2609.37907 · 2026-10-01视觉分支是CLIP类增量学习所需的关键
Visual Branch is What You Need for CLIP-based Class-Incremental Learning
arXiv 2609.37888 · 2026-10-01主动预算会损害灵敏度：诊断与修复TopK稀疏自编码器的可靠性
Active Budget Can Kill Sensitivity: Diagnosing and Repairing TopK Sparse Autoencoder Reliability
arXiv 2609.37857 · 2026-10-01圆里德伯原子间的长时自旋交换相互作用
Spin-exchange interactions between circular Rydberg atoms over long times
arXiv 2609.37766 · 2026-10-01反向缠绕S形Aharonov-Bohm干涉仪中的旋转对称性与共模相位漂移
Rotational symmetry and common-mode phase drift in a counter-wound S-shaped Aharonov--Bohm interferometer
arXiv 2609.37753 · 2026-10-01空间形式乘积中的等参子流形
Isoparametric Submanifolds in Products of Space Forms
arXiv 2609.37739 · 2026-10-01PAIQ：通过残差旋转进行补丁对齐的语义注入
PAIQ: Patch-Aligned Semantic Injection via Residual Rotation
arXiv 2609.37685 · 2026-10-01针对对比视觉-语言模型的目标视觉反事实解释
Targeted Visual Counterfactual Explanations for Contrastive Vision-Language Model
arXiv 2609.37638 · 2026-10-01约化子空间上范数达到算子的稳定性与稠密性
The Stability and Density of Norm Attaining Operators on Reducing Subspaces
arXiv 2609.37621 · 2026-10-01金刚石中NV$^-$系综相干性的氮同位素效应的半经典自旋浴计算
Semiclassical spin-bath calculation of the nitrogen-isotope effect on NV$^-$ ensemble coherence in diamond
arXiv 2609.37596 · 2026-10-01RoboHarn-Evo：演化分层物理知识以实现自改进的机器人操作
RoboHarn-Evo: Evolving Hierarchical Physical Knowledge for Self-Improving Robotic Manipulation
arXiv 2609.37583 · 2026-10-01MERGE：通过生成式增强实现多LLM检索集成
MERGE: Multi-LLM Ensemble for Retrieval via Generative Enrichment
arXiv 2609.37574 · 2026-10-01三层命名测量框架及其在代币发行平台上的普查应用
A Three-Layer Framework for Measuring Names and Its Census Application on a Token Launchpad
arXiv 2609.37572 · 2026-10-01推荐反馈如何演化智能体记忆？
How Can Recommendation Feedback Evolve Agent Memory?
arXiv 2609.37544 · 2026-10-01率无关生物声学：基于连续滤波器组和傅里叶神经算子的异质多分类
Rate-Agnostic Bioacoustics: Heterogeneous Multi-Taxa Classification with Continuous Filterbanks and Fourier Neural Operators
arXiv 2609.37540 · 2026-10-01Bad：用蝙蝠活动检测器驯服生物声学数据洪流
Bad: Taming the Bioacoustic Data Deluge with a Bat Activity Detector
arXiv 2609.37518 · 2026-10-01流式 Clifford+T 编译中的记忆-魔法交换律
A Memory-Magic Exchange Law in Streaming Clifford+T Compilation
arXiv 2609.37368 · 2026-10-01可旋转天线赋能的天-空-地一体化网络：机遇与挑战
Rotatable Antenna-Enabled Space-Air-Ground Integrated Networks: Opportunities and Challenges
arXiv 2609.37338 · 2026-10-01全息方法中的量子湍流主动正常流体反馈
Active normal-fluid feedback in quantum turbulence from holography
arXiv 2609.37269 · 2026-10-01HELIX：净化与统一——重新思考大规模推荐中的特征交互与序列建模
HELIX: Purified and Unified - Rethinking Feature Interaction and Sequence Modeling for Large-Scale Recommendation
arXiv 2609.37183 · 2026-10-01工程化双曲异质结构中的真空涨落
Engineering Vacuum Fluctuations in Hyperbolic Heterostructures
arXiv 2609.37182 · 2026-10-01SimpleEvol：一种用于LLM驱动的自动化启发式设计且最小化人类先验的智能体循环框架
SimpleEvol: An Agent-Loop Framework for LLM-Driven Automated Heuristic Design with Minimal Human Priors
arXiv 2609.37172 · 2026-10-01代理何时应检查外部状态？为存储意图的观察进行预算分配
When Should Agents Check External State? Budgeting Observations for Stored Intentions
arXiv 2609.37125 · 2026-10-01从可行失败前缀中学习：面向长程LLM智能体的里程碑可行性潜力策略优化
Learning from Viable Failure Prefixes: Milestone Viability Potential Policy Optimization for Long-Horizon LLM Agents
arXiv 2609.37111 · 2026-10-01Waypoint-1.5：面向消费级硬件的实时视频世界模型
Waypoint-1.5: A Real-Time Video World Model for Consumer Hardware
arXiv 2609.37107 · 2026-10-01V2X-WAM：一种用于端到端自动驾驶的协同世界动作模型
V2X-WAM: A Cooperative World Action Model for End-to-End Autonomous Driving
arXiv 2609.37098 · 2026-10-01Traverse：学习何时记忆、重置与重定向以进行长时程网络搜索
Traverse: Learning When to Remember, Reset, and Redirect for Long-Horizon Web Search
arXiv 2609.37082 · 2026-10-01具有有效界面条件的均质化模型用于斯托克斯流通过多孔膜的数值研究
Numerical investigation of a homogenized model with effective interface conditions for Stokes flow through porous membranes
arXiv 2609.37068 · 2026-10-01actr：对齐推理大语言模型中的思维与响应以实现多语言安全
ACTR: Aligning Thoughts and Responses for Multilingual Safety in Reasoning LLMs
arXiv 2609.37054 · 2026-10-01NHO：一种用于锚点区域定位与稠密对应的神经哈密顿算子
NHO: A Neural Hamiltonian Operator for Anchor-based Region Localization and Dense Correspondence
arXiv 2609.37048 · 2026-10-01多层深度时间融合的前馈、局部训练脉冲神经网络
Multi-Depth Temporal Fusion for Feedforward, Locally Trained Spiking Neural Networks
arXiv 2609.37047 · 2026-10-01从思考模式优势中学习：基于策略的蒸馏
Learning from Think-Mode Advantage via On-Policy Distillation
arXiv 2609.37044 · 2026-10-01RBF-GNN：用于基于伪坐标的图卷积的有理基函数
RBF-GNN: Rational Basis Functions for Pseudo-Coordinate based Graph Convolutions
arXiv 2609.37015 · 2026-10-01World2Motion：将视频世界模型转化为3D人体运动生成器
World2Motion: Turning Video World Models into 3D Human Motion Generators
arXiv 2609.37004 · 2026-10-01视觉并行搜索：通过并行瓦片检查与自适应缩放学习搜索高分辨率图像
Visual Parallel Search: Learning to Search High-Resolution Images with Parallel Tile Inspection and Adaptive Zoom
arXiv 2609.37002 · 2026-10-01JudgeCast：基于经验信息协变量判断的时间序列预测
JudgeCast: Time Series Forecasting with Experience-Informed Covariate Judgements
arXiv 2609.36966 · 2026-10-01有效量子场论与SMEFT
Effective Quantum Field Theory and SMEFT
arXiv 2609.36960 · 2026-10-01基于能量神经网络的安全设计学习
Safe-by-Design Learning via Energy-based Neural Networks
arXiv 2609.36942 · 2026-10-01CoEM：利用基于证据的承诺记忆增强长上下文推理
CoEM: Empowering Long-Context Reasoning with Commit-on-Evidence Memory
arXiv 2609.36935 · 2026-10-01
