2026-09论文中文摘要
时间相关性如何塑造线性循环神经网络的记忆
How Temporal Correlations Shape Memory in Linear Recurrent Neural Networks
arXiv 2609.00420 · 2026-09-02每个非递归多一度要么包含一个有限一度，要么包含无穷多个有限一度
Every Nonrecursive Many-One Degree Contains Either One or Infinitely Many Finite-One Degrees
arXiv 2609.00419 · 2026-09-02斯大林是否必要？关于苏联经济增长的反事实证据
Was Stalin Necessary? Counterfactual Evidence on Soviet Growth
arXiv 2609.00418 · 2026-09-02一致有界表示的相对性质(T)
Relative property (T) for uniformly bounded representations
arXiv 2609.00417 · 2026-09-02Transformer的后期层对句法进行规范重编码：来自希腊语移位及跨层泛化的证据
Late Transformer Layers Recode Syntax Canonically: Evidence from Greek Scrambling and Cross-Layer Generalisation
arXiv 2609.00416 · 2026-09-02薛定谔算子的分离性：西蒙猜想的一个反例
Separation for Schrödinger operators. A counterexample to Simon's conjecture
arXiv 2609.00415 · 2026-09-02玻利维亚液化石油气补贴改革、能源补偿与社会风险：基于机器学习智能体的微观模拟
LPG Subsidy Reform, Energy Compensation, and Social Risk in Bolivia: A Machine-Learning Agent-Based Microsimulation
arXiv 2609.00414 · 2026-09-02面向金融推理的依赖感知思维链压缩
Dependency-Aware Chain-of-Thought Compression for Financial Reasoning
arXiv 2609.00413 · 2026-09-02FocusBuddy：通过照料水瓶上的虚拟宠物鼓励健康的伏案工作习惯
FocusBuddy: Encouraging Healthy Desk-Work Habits by Caring for a Virtual Pet on a Water Bottle
arXiv 2609.00412 · 2026-09-02揭示人脸嵌入：用基础模型进行读取、渲染与命名
Unmasking Face Embeddings: Reading, Rendering and Naming with Foundation Models
arXiv 2609.00411 · 2026-09-02当我们谈论核反应时，我们在谈论什么
What Are We Talking About When We Talk About Nuclear Reactions
arXiv 2609.00410 · 2026-09-02C*-代数上态的支配性与等价性：拟不变态
Dominance and equivalence for states on $C^*$-algebras: Quasi-Invariant states
arXiv 2609.00409 · 2026-09-02吉布斯采样与CAVI中的非均匀随机扫描
Non-Uniform Random Scans in Gibbs Sampling and CAVI
arXiv 2609.00408 · 2026-09-02DynaNDE：面向批量MoE推理的动态近数据专家调度
DynaNDE: Dynamic Near-Data Expert Scheduling for Batched MoE Inference
arXiv 2609.00407 · 2026-09-02面向安全关键部署流水线的审计优先回滚语义
Audit-First Rollback Semantics for Safety-Critical Deployment Pipelines
arXiv 2609.00406 · 2026-09-02可交换乘法作用的多重遍历平均
Multiple ergodic averages for commuting multiplicative actions
arXiv 2609.00405 · 2026-09-02用于具身机器人能力市场的联邦信任机制
Federated Trust for Embodied Robot Capability Marketplaces
arXiv 2609.00404 · 2026-09-02面向健康虚假信息检测与传播分析的多分支特征融合方法
A Multi-Branch Feature Fusion Approach for Health Misinformation Detection and Propagation
arXiv 2609.00403 · 2026-09-02考虑资本成本与调度约束的公交车队电气化：五机构案例研究
Bus Fleet Electrification Under Capital Cost and Scheduling Constraints: A Five-Agency Case Study
arXiv 2609.00401 · 2026-09-02用于远程ROOT数据分析的客户端透明缓存
Client-side transparent caching for remote ROOT data analysis
arXiv 2609.00400 · 2026-09-02香农关于熵单调性的问题与陶哲轩的一个猜想
Shannon's problem on the monotonicity of entropy and a Conjecture of Tao
arXiv 2609.00399 · 2026-09-02年轻炽热白矮星周围行星固态物质在升华过程中的存活情况
Where planetary solids survive sublimation around young and hot white dwarfs
arXiv 2609.00398 · 2026-09-02关于高维多面体不可避免的面
On Unavoidable Faces of High-Dimensional Polytopes
arXiv 2609.00397 · 2026-09-02SlideMix：通过多模态混洗增强全切片图像分析
SlideMix: Enhancing Whole Slide Image Analysis via Multimodal Shuffling
arXiv 2609.00396 · 2026-09-02基于模块化晶格的超材料的非线性力响应
Nonlinear force response of modular lattice-based metamaterials
arXiv 2609.00395 · 2026-09-02连续顶点删除下具有长伪相似链的图
Graphs with Long Pseudosimilarity Chains under Consecutive Vertex Deletions
arXiv 2609.00394 · 2026-09-02一种降维的Fréchet简化神谕
A Dimension-Reducing Fréchet Simplification Oracle
arXiv 2609.00393 · 2026-09-02面向基于领导者的共识数据存储的操作类型感知客户端路由
Operation-Type-Aware Client Routing for Leader-Based Consensus Datastores
arXiv 2609.00392 · 2026-09-02NeuroPriv：面向可穿戴脑电图系统隐私的对抗性表示学习
NeuroPriv: Adversarial Representation Learning for Privacy in Wearable EEG Systems
arXiv 2609.00390 · 2026-09-02算子学习的神经均值与核修正方法
Neural means and kernel corrections for operator learning
arXiv 2609.00389 · 2026-09-02约旦输电电网的开放合成测试系统
An Open Synthetic Test System for the Jordanian Transmission Grid
arXiv 2609.00388 · 2026-09-02文档衍生属性图中的隐藏关系：动态演化本体上的top-k块嵌入与逆距离加权
Hidden relationships in a document-derived property graph: top-k chunk embeddings and inverse-distance weighting over a dynamically evolving ontology
arXiv 2609.00387 · 2026-09-02短区间内的几乎k次幂
Almost $k$-th powers in short intervals
arXiv 2609.00386 · 2026-09-02面向自主超车的风险感知决策：一种基于世界模型的混合专家框架
Risk-Aware Decision-Making for Autonomous Overtaking: A World Model-Based Mixture-of-Experts Framework
arXiv 2609.00385 · 2026-09-02RestoreBench：AI智能体能恢复潮流收敛吗？
RestoreBench: Can AI Agents Restore Power Flow Convergence?
arXiv 2609.00384 · 2026-09-02对布谷鸟巢的一次忠实遍历
One Faithful Pass Over the Cuckoo's Nest
arXiv 2609.00383 · 2026-09-02有界、不确定还是漏洞：面向SQL聚合函数差异测试的感知条件预言机
Bounded, Indeterminate, or a Bug: A Condition-Aware Oracle for Differential Testing of SQL Aggregates
arXiv 2609.00381 · 2026-09-02关于内生问题的沟通
Communicating About Endogenous Issues
arXiv 2609.00380 · 2026-09-02关于LTLf到LTL归约的一个注记
A note on the reduction from LTLf to LTL
arXiv 2609.00379 · 2026-09-02可移除且不可约：多语言分词税的代币成本账本
Removable and Irreducible: A Token-Cost Ledger for the Multilingual Tokenization Tax
arXiv 2609.00378 · 2026-09-02FoldingAgent：从演示视频推断参数化折纸流程的智能体框架
FoldingAgent: Inferring Parametric Origami Procedures from Demonstration Videos
arXiv 2609.00377 · 2026-09-02采用空间验证机器学习方法的住宅价格建模：不同地理背景下的比较
Residential Price Modeling using Spatially Validated Machine Learning Methods: A Comparison Across Geographic Contexts
arXiv 2609.00375 · 2026-09-02无梯度适应：仿射统计传输及其证书的作用
Adapting Without Gradients: Affine Statistics Transport and What Its Certificate Can Tell You
arXiv 2609.00374 · 2026-09-02企业忠诚度：部分AI系统会有差别地淡化其开发者的争议
Corporate Loyalty: Some AI Systems Differentially Downplay their Creators' Controversies
arXiv 2609.00373 · 2026-09-02面向量子数据的无监督表示学习：兼具推理与生成能力的量子模型
Towards unsupervised representation learning for quantum data: quantum models with inference and generation
arXiv 2609.00372 · 2026-09-02MorphPatch：利用表面近似和视触觉错觉增强形状显示器上的VR交互
MorphPatch: Enhancing VR Interaction on Shape Displays using Surface Approximation and Visuo-Haptic Illusions
arXiv 2609.00371 · 2026-09-02反事实接近加速度风险：一种适用于跟驰盲区的预测性替代安全措施
Counterfactual Closing-Acceleration Risk: An Anticipatory Surrogate Safety Measure for the Blind Region of Car-Following
arXiv 2609.00370 · 2026-09-02Puppeteer：基于物体的姿态感知同步言语手势生成
Puppeteer: Object-Grounded Posture-Aware Co-Speech Gesture Generation
arXiv 2609.00369 · 2026-09-02手性腔驱动的手性声子
Chiral phonons driven by chiral cavities
arXiv 2609.00368 · 2026-09-02用于数据工程的神经符号学：无需微调即可实现长上下文令牌减少
Neurosymbolics for Data Engineering: Achieving Long Context Token Reduction Without Finetuning
arXiv 2609.00367 · 2026-09-02
