2026-09论文中文摘要
用户离开后自主深度研究期间会发生什么？
What Happens During Autonomous Deep Research After the User Steps Away?
arXiv 2609.33509 · 2026-09-29纠缠物理态的完全退极化噪声猜想：二十年的视角
The Fully Depolarizing Noise Conjecture for Entangled Physical States: A Twenty-Year Perspective
arXiv 2609.33508 · 2026-09-29Alperin-McKay-Navarro 猜想的一个约化定理
A Reduction Theorem for the Alperin-McKay-Navarro Conjecture
arXiv 2609.33507 · 2026-09-29当证据改变主题：用于学习路由的主题类型声明许可
When Evidence Changes the Subject: Subject-Typed Claim Licensing for Learned Routing
arXiv 2609.33505 · 2026-09-29反德西特三维空间中常平均曲率与常高斯曲率类时曲面（基于圈群方法）
Timelike surfaces of constant mean and constant Gaussian curvature in anti-de Sitter 3-space via loop groups
arXiv 2609.33504 · 2026-09-29RelaxKV：基于查询的稀疏上下文注意力重计算指导，用于高效KV缓存复用
RelaxKV: Recomputation Guided by the Query with Sparse Context Attention for Efficient KV Cache Reuse
arXiv 2609.33503 · 2026-09-29H-三元组的块关系与Clifford理论
Block Relations and Clifford Theory for H-Triples
arXiv 2609.33502 · 2026-09-29Pulseflow: 通过潜在传输实现 PPG 反事实生成
Pulseflow: PPG Counterfactual Generation Via Latent Transport
arXiv 2609.33501 · 2026-09-29一种用于耦合剖面与平衡演化的快速集成托卡马克建模方法
A Rapid Integrated Tokamak Modelling Approach for Coupled Profile and Equilibrium Evolution
arXiv 2609.33500 · 2026-09-29有用自然梯度更新的代价
The cost of useful natural gradient updates
arXiv 2609.33499 · 2026-09-29Hamiltonian JEPA：具有继承控制状态的行动条件世界模型
Hamiltonian JEPA: Action-Conditioned World Models with an Inherited Control State
arXiv 2609.33497 · 2026-09-29Chameleon：用于高效扩散模型的动态格式适配器
Chameleon: Dynamic Format Adapter for Efficient Diffusion
arXiv 2609.33496 · 2026-09-29LLMs Trust Their Own: 多智能体系统中身份依赖的从众行为
LLMs Trust Their Own: Identity-Dependent Conformity in Multi-Agent Systems
arXiv 2609.33495 · 2026-09-29多电子原子中的非偶极效应
Non-dipole effects in a multielectron atom
arXiv 2609.33494 · 2026-09-29坐标下降、单值性和带标记根映射的有限正规化
Coordinate descents, monodromy, and finite normalization of marked-root maps
arXiv 2609.33491 · 2026-09-29域自适应扩散模型用于条件独立性检验
Domain-Adapted Diffusion Models for Conditional Independence Testing
arXiv 2609.33490 · 2026-09-29通过交叉曲率预测块坐标性能
Predicting Block-Coordinate Performance via Cross-Curvature
arXiv 2609.33489 · 2026-09-29环面Deligne--Mumford栈上的等变主丛
Equivariant principal bundles over toric Deligne--Mumford stacks
arXiv 2609.33488 · 2026-09-29哪种掩码几何结构最适合 EEG 基础模型？
What masking geometry works best for EEG foundation models?
arXiv 2609.33487 · 2026-09-29掩蔽引起的音频XAI位移：通过Logit轨迹分解
Mask-Induced Displacement in Audio XAI via Logit Trajectory Decomposition
arXiv 2609.33486 · 2026-09-29DISCO：通过接地-推理分离实现分布式长上下文扩展
DISCO: Distributed Long Context Scaling with Grounding-Reasoning Disaggregation
arXiv 2609.33485 · 2026-09-29AMBIT：面向人形机器人双臂跟踪的预期性多模态身体招募
AMBIT: Anticipatory Multimodal Body Recruitment for Bimanual Tracking on a Humanoid
arXiv 2609.33484 · 2026-09-29连续激光驱动α粒子源
Continuous Laser-Driven α Source
arXiv 2609.33483 · 2026-09-29合成标签如何改进共形预测：条件覆盖率的视角
How Synthetic Labels Improve Conformal Prediction: A Perspective on Conditional Coverage
arXiv 2609.33482 · 2026-09-29忘记一个人意味着什么？视觉-语言模型中的个体级遗忘
What Does It Mean to Forget a Person? Individual-Level Unlearning in Vision-Language Models
arXiv 2609.33481 · 2026-09-29二项式系数与operadic蝴蝶图
Binomial coefficients and the operadic butterfly
arXiv 2609.33480 · 2026-09-29具有两个调制系数和对数型增长的向量值双相泛函的正则性
Regularity for vectorial double phase functionals with two modulating coefficients and logarithmic-type growth
arXiv 2609.33479 · 2026-09-29迈向催化剂介导的金刚石和立方氮化硼线状晶体生长
Toward catalyst-mediated growth of diamond and cubic boron nitride wires
arXiv 2609.33478 · 2026-09-29让线性状态遗忘遥远的过去：面向混合LLM的基于后缀重放的前缀缓存
Just Let Linear States Forget the Distant Past: Prefix Caching via Suffix Replay for Hybrid LLMs
arXiv 2609.33477 · 2026-09-29双平衡移位与可变规范
Bi-balanced shifts and variable specification
arXiv 2609.33476 · 2026-09-29一个分子在失去相干性之前能散射多少光子？
How many photons can a molecule scatter before its coherence is lost?
arXiv 2609.33475 · 2026-09-29贝叶斯调度中的间隙与增强
Gaps and Augmentations in Bayesian Scheduling
arXiv 2609.33474 · 2026-09-29三次二分Holant问题带复代数权重的二分法
A Dichotomy for Cubic Bipartite Holant Problems with Complex Algebraic Weights
arXiv 2609.33473 · 2026-09-29预测-然后-优化学习中的几何识别
Geometric Identification in Predict-Then-Optimize Learning
arXiv 2609.33472 · 2026-09-29一个单鞍点1/5近似算法用于公共核双矩阵博弈：识别、精确段优化、尖锐选择器界与认证鲁棒性
A One-Saddle 1/5 Approximation Algorithm for Common-Kernel Bimatrix Games: Recognition, Exact Segment Optimization, Sharp Selector Bounds, and Certified Robustness
arXiv 2609.33471 · 2026-09-29LiveOption：评估具有非线性收益的结构化期权交易中的LLM智能体
LiveOption: Evaluating LLM Agents in Structured Option Trading with Nonlinear Payoffs
arXiv 2609.33470 · 2026-09-29诺特形式的转移理论
Transfer theory for noetherian forms
arXiv 2609.33469 · 2026-09-29基于DDPG的STAR-RIS辅助6G网络智能切换
DDPG-Based Intelligent Handover For STAR-RIS-Assisted 6G Networks
arXiv 2609.33468 · 2026-09-29廉价验证器足矣：LLM 后训练对错误奖励具有鲁棒性
A Cheap Verifier is Good Enough: LLM Post-training is Robust to Erroneous Rewards
arXiv 2609.33467 · 2026-09-29再感染的生命期模式：作为行波的流行病队列
Life-time patterns of reinfections: epidemic cohorts as traveling waves
arXiv 2609.33466 · 2026-09-29GSM：利用共享全局状态的高效语言建模
GSM: Efficient Language Modeling with Shared Global State
arXiv 2609.33465 · 2026-09-29VIDEAS：通过先验引导仿真从演示视频中为世界模型提炼显式动作语义
VIDEAS: Distilling Explicit Action Semantics from Demonstration Videos for World Models via Prior-Guided Simulation
arXiv 2609.33464 · 2026-09-29重新思考SFT的令牌加权：抑制、反转和外推已学习特征
Rethinking Token Reweighting for SFT: Suppress, Reverse, and Extrapolate Learned Features
arXiv 2609.33463 · 2026-09-29受保护核心并不足够：认证AI提出的时序规范修订
Protected Cores Are Not Enough: Certifying AI-Proposed Revisions of Temporal Specifications
arXiv 2609.33461 · 2026-09-29裸片反铁磁计算
Bare-Die Antiferromagnetic Computing
arXiv 2609.33460 · 2026-09-29求解多项式方程中幂级数合并的形式体系：主级数与主数
A Formalism for Merging Power Series in Solving Multinomial Equations: Master Series and Master Numbers
arXiv 2609.33459 · 2026-09-29“狗”的概念何时在音频大语言模型中涌现？
When Does the Concept of "Dog" Emerge in an Audio LLM?
arXiv 2609.33458 · 2026-09-29一个自由的旋钮：在基于阈值的评估中解耦校准与预测技能
A Free Knob: Decoupling Calibration and Predictive Skill in Threshold-Based Evaluation
arXiv 2609.33457 · 2026-09-29协变量偏移下共形预测的锐利训练条件覆盖
Sharp training-conditional coverage for conformal prediction under covariate shift
arXiv 2609.33456 · 2026-09-29Kitaev-Heisenberg模型在方形-六边形-十二边形格子上的研究
Kitaev-Heisenberg model on the square-hexagon-dodecagon lattice
arXiv 2609.33454 · 2026-09-29
