2026-10论文中文摘要
互补特征域：信息保持并不意味着预测贡献保持
Complementary Feature Domains: Information Preservation Does Not Imply Predictive-Contribution Preservation
arXiv 2610.07565 · 2026-10-07基于曲率的 $p$-容量界与容量--面积不等式
Curvature-Based Bounds for $p$-Capacity and Capacity--Area Inequalities
arXiv 2610.07564 · 2026-10-07HouseholdBench：评估大型语言模型作为家庭经济行为预测器的性能
HouseholdBench: Evaluating Large Language Models as Predictors of Household Economic Behavior
arXiv 2610.07563 · 2026-10-07学习GFlowNets的混合模型
Learning a Mixture of GFlowNets
arXiv 2610.07562 · 2026-10-07LARK：一种低成本、精确、抗遮挡、卡尔曼滤波辅助的图像引导手术跟踪系统
LARK: A Low-Cost, Accurate, Occlusion-Resilient, Kalman Filter-Assisted Tracking System for Image-Guided Surgery
arXiv 2610.07561 · 2026-10-07导航路线潜在空间用于可合成分子设计
Navigating Route Latent Space for Synthesizable Molecular Design
arXiv 2610.07560 · 2026-10-07TAFFY：具有上下文多样性的任务自适应表格基础模型
TAFFY: A Task-Adaptive Tabular Foundation Model with In-Context Diversity
arXiv 2610.07559 · 2026-10-07看见不可见：物理引导的视觉提示用于温度和辐射感知的VLA导航
Seeing the Invisible: Physics-Guided Visual Prompting for Temperature- and Radiation-Aware VLA Navigation
arXiv 2610.07558 · 2026-10-07CheckerBench：长时程智能体能否合成静态分析检查器？
CheckerBench: Can Long-Horizon Agents Synthesize Static-Analysis Checkers?
arXiv 2610.07557 · 2026-10-07全局传输耦合用于无分类器引导流
Global Transport Couplings for Classifier-Free Guided Flows
arXiv 2610.07555 · 2026-10-07CGMY tempered 稳定过程下期权定价的 B 样条伽辽金方法
A B-Spline Galerkin Method for Option Pricing under the CGMY Tempered Stable Process
arXiv 2610.07554 · 2026-10-07何时接纳与接纳什么：面向数据为中心的小语言模型微调的梯度接纳方法
Which and When to Admit: Gradient Admission for Data-Centric Small Language Model Finetuning
arXiv 2610.07553 · 2026-10-07AI/ML 磁性材料计算搜索的可靠性：数据库、验证与可合成性
Reliability of AI/ML Computational Searches for Magnetic Materials: Databases, Validation, and Synthesizability
arXiv 2610.07552 · 2026-10-07高维中梯度EM学习高斯混合模型是否必须满足$\sqrt{d}$的分离条件？
Is $\sqrt{d}$ Separation Necessary for Gradient EM to Learn Gaussian Mixtures in High Dimensions?
arXiv 2610.07551 · 2026-10-07基础模型辅助的多智能体强化学习用于无线随机接入网络优化
Foundation Model-Aided Multi-Agent Reinforcement Learning for Wireless Random Access Network Optimization
arXiv 2610.07550 · 2026-10-07基本超几何级数的正性与 MacMahon 恒等式
Positivity of the basic hypergeometric series and the MacMahon identity
arXiv 2610.07549 · 2026-10-07AdS/CFT 与 Carroll 全息中的自旋-1 体态局域态
Spin-1 Bulk Local States in AdS/CFT and Carrollian Holography
arXiv 2610.07548 · 2026-10-07基于体物理性质的数据驱动预测聚合物中原子氧侵蚀产率
Data-driven prediction of atomic-oxygen erosion yield in polymers from bulk physical properties
arXiv 2610.07547 · 2026-10-07$K_n$ 的公平 $(s,p)$-边着色的嵌入
Embedding equitable (s,p)-edge-colorings of $K_n$
arXiv 2610.07546 · 2026-10-07边缘设备上的质量感知自校正语音翻译
Quality-Aware Self-Correcting Speech Translation on an Edge Device
arXiv 2610.07545 · 2026-10-07大型语言模型在广告相关性判断中的偏见系统性研究
A Systematic Investigation of Bias in Large Language Models for Advertising Relevance
arXiv 2610.07544 · 2026-10-07索引对数 Cartier 变换
Indexed Logarithmic Cartier Transform
arXiv 2610.07543 · 2026-10-07己二酸生产过程中一氧化二氮排放的卫星观测
Satellite observations of nitrous oxide emissions from adipic acid production
arXiv 2610.07542 · 2026-10-07在JEPA世界模型中通过逆动力学保留不稳定模态
Preserving Unstable Modes Through Inverse Dynamics in JEPA World Models
arXiv 2610.07540 · 2026-10-07电路网络与辛不变量
Electrical Networks and Symplectic Invariants
arXiv 2610.07539 · 2026-10-07有限域中的本原正规等差数列
Primitive normal progressions in finite fields
arXiv 2610.07538 · 2026-10-07面向数值相对论应用的人工粘性间断伽辽金方法
A discontinuous Galerkin method with artificial viscosity geared toward numerical relativity applications
arXiv 2610.07536 · 2026-10-07在异构LLM模拟中解耦模型与人格
Disentangling Models from Personas in Heterogeneous LLM Simulations
arXiv 2610.07535 · 2026-10-07EPIC框架：面向规范驱动的开发
The EPIC Framework for Spec-Driven Development
arXiv 2610.07534 · 2026-10-07SkillFormer：音频语言模型的技能分解适配
SkillFormer: Skill-Decomposed Adaptation for Audio Language Models
arXiv 2610.07533 · 2026-10-07通过暗知识中的模型无关潜在安全信号保护大语言模型
Safeguarding LLMs via Model-Agnostic Latent Safety Signals from Dark Knowledge
arXiv 2610.07532 · 2026-10-07BVI：轻量级、以数据为中心的基于区块链的身份声明验证
BVI: Lightweight, Data-Centric Blockchain-Based Verification of Identity Claims
arXiv 2610.07531 · 2026-10-07四维凸环面域辛嵌入问题中的障碍
Obstructions in Symplectic Embedding Problems for Four-Dimensional Convex Toric Domains
arXiv 2610.07530 · 2026-10-07定向搜索表明随机器件测试低估了模拟波神经算子中的最坏情况误差
Targeted search shows that random-device testing underestimates worst-case error in a simulated wave-based neural operator
arXiv 2610.07529 · 2026-10-07有限全纯拉回、Segre 数与一般Lê数
Finite holomorphic pullbacks, Segre numbers, and generic Lê numbers
arXiv 2610.07528 · 2026-10-07任务空间模仿引导用于高效强化学习
Task-Space Imitation Guidance for Efficient Reinforcement Learning
arXiv 2610.07527 · 2026-10-07Szarek不等式最佳常数的稳定性
Stability of Szarek's inequality with best constant
arXiv 2610.07526 · 2026-10-07ReDex: 通过手指级柔顺交互修复仿真到现实的灵巧策略
ReDex: Repairing Sim-to-Real Dexterous Policies by Finger-Level Compliant Interaction
arXiv 2610.07525 · 2026-10-07区域感知掩码用于口音鲁棒的跨语言文本到语音合成
Region-Aware Masking for Accent-Robust Cross-Lingual Text-to-Speech
arXiv 2610.07524 · 2026-10-07PAIR：实时多模态对话智能体中的感知情感推断与调节
PAIR: Perceptual Affective Inference and Regulation in a Real-Time Multimodal Conversational Agent
arXiv 2610.07523 · 2026-10-07激活去噪：并行与串行LLM量化的鲁棒性视角
Activation Denoising: A Robustness View on Parallel vs Sequential LLM Quantization
arXiv 2610.07522 · 2026-10-07锚定规划之形：重新思考自动驾驶中物理智能的接地性
Grounding What Shapes the Plan: Rethinking Groundedness for Physical Intelligence in Autonomous Driving
arXiv 2610.07521 · 2026-10-07关于参数模 Hecke 代数的进一步结果
Further results on the paramodular Hecke algebra
arXiv 2610.07520 · 2026-10-07非儿童所表达：面向儿童AI的手语到文本安全接口审计
Not What a Child Expressed: Auditing the Sign-to-Text Safety Interface in Child-Facing AI
arXiv 2610.07519 · 2026-10-07有害SFT在LLM检查点更新中留下连续痕迹
Harmful SFT Leaves a Continuous Trace in LLM Checkpoint Updates
arXiv 2610.07518 · 2026-10-07速率理论计算什么：从阿伦尼乌斯到机器学习势的化学动力学
What Rate Theories Compute: Chemical kinetics from Arrhenius to machine-learned potentials
arXiv 2610.07517 · 2026-10-07NCCL M2N：一种面向分布式张量重分片的布局与拓扑感知集合通信原语
NCCL M2N: A Layout- and Topology-Aware Collective for Distributed Tensor Resharding
arXiv 2610.07516 · 2026-10-07从局部证据到安全判定：视觉-语言模型中的因果追踪
From Local Evidence to Safety Verdicts: Causal Tracing in Vision-Language Models
arXiv 2610.07514 · 2026-10-07结构化线性算子的高效近似：使用随机单侧查询
Efficient Approximation of Structured Linear Operators using Randomized One-Sided Queries
arXiv 2610.07513 · 2026-10-07X2C启发的多小波基下Dirac耦合算符近似
X2C-inspired approximation of the Dirac coupling operator in a multiwavelet basis
arXiv 2610.07512 · 2026-10-07
