2026-09论文中文摘要
将AI引入自主系统——从认知到集体智能
Bringing AI to Autonomous Systems -- From Cognition to Collective Intelligence
arXiv 2609.30291 · 2026-09-28生产环境文本到SQL流水线中LLM作为评判者的失败审计与修复
Auditing and Repairing LLM-as-Judge Failures in a Production Text-to-SQL Pipeline
arXiv 2609.30290 · 2026-09-28并非所有记忆都同等重要：面向LLM智能体有效性感知检索的层级协作记忆
Not All Memories Are Equal: Hierarchical Collaborative Memory for Validity-Aware Retrieval in LLM Agents
arXiv 2609.30289 · 2026-09-28流形投影与迭代自编码器精化用于掩码语言建模
Manifold Projection and Iterative Autoencoder Refinement for Masked Language Modeling
arXiv 2609.30288 · 2026-09-28冻结BERT中AI文本检测神经元的机制研究：基于RAID的稀疏探针与激活修补
A Mechanistic Study of AI-Text Detection Neurons in Frozen BERT: Sparse Probing and Activation Patching on RAID
arXiv 2609.30287 · 2026-09-28平流感知图神经网络何时有助于预测？——基于自监督云运动估计器的分布式光伏爬坡受控研究
When Does Advection-Aware Graph Nowcasting Help? A Controlled Study of Distributed Solar Ramp Forecasting with a Self-Supervised Cloud-Motion Estimator
arXiv 2609.30286 · 2026-09-28冯·诺依曼的两个量子熵
Von Neumann's Two Quantum Entropies
arXiv 2609.30284 · 2026-09-28通过层级辅助量子位控制子空间投影限制等变量子网络中的可训练李代数增长
Restricting Trainable Lie-Algebra Growth in Equivariant Quantum Networks via Hierarchical Ancilla-Controlled Subspace Projections
arXiv 2609.30283 · 2026-09-28超越Floquet-Volkov路径的阿秒拓扑干涉
Attosecond topological interference beyond Floquet-Volkov paths
arXiv 2609.30282 · 2026-09-28中子监测器时间序列预测的季节性与量子启发模型
Seasonal and Quantum-inspired Models for Neutron Monitor Time Series Forecasting
arXiv 2609.30281 · 2026-09-28图像诱导哈密顿谱的噪声稳定性与旋转等变性用于鲁棒无监督分割
Noise Stability and Rotational Equivariance of Image-Induced Hamiltonian Spectra for Robust Unsupervised Segmentation
arXiv 2609.30280 · 2026-09-28神经理想与神经代码：用于神经网络分类与特征解释的代数框架
Neural Ideals and Neural Codes: An Algebraic Framework for Neural Network Classification and Feature Interpretation
arXiv 2609.30279 · 2026-09-28SatoyamaCT：用于日本农林复合生态系统中农作物损害野生动物监测的多轴夜间红外相机陷阱基准
SatoyamaCT: A Multi-Axis Night-IR Camera-Trap Benchmark for Monitoring Crop-Damaging Wildlife in Japanese Agroforestry
arXiv 2609.30278 · 2026-09-28为什么裁剪对AdaGrad至关重要？——广义光滑性下的高概率理论
Why Clipping Matters in AdaGrad? Toward a High-Probability Theory under Generalized Smoothness
arXiv 2609.30276 · 2026-09-28余弦相似度不是证据：量化下可解释性迁移的噪声基底测量
Cosine Similarity Is Not Evidence: Measuring the Noise Floor of Interpretability Transfer Under Quantization
arXiv 2609.30275 · 2026-09-28$\mathbb{R}^d$ 上耗散随机动力系统命中时间的分布及其在随机梯度下降中的应用
Distribution of hitting times for dissipative random dynamical systems on $\mathbb{R}^d$, with application to stochastic gradient descent
arXiv 2609.30274 · 2026-09-28离线策略评估作为设计自适应实验的决策支持工具
Offline Policy Evaluation as a decision support tool for designing Adaptive Experiments
arXiv 2609.30273 · 2026-09-28ENAS：面向资源受限微控制器上TinyML的高效硬件感知神经架构搜索框架
ENAS: An Efficient Hardware-Aware Neural Architecture Search Framework for TinyML on Resource-Constrained Microcontrollers
arXiv 2609.30272 · 2026-09-28当预条件指数变为负值时：学习率耦合与跨环境泛化
When the Preconditioning Exponent Turns Negative: Learning-Rate Coupling and Cross-Environment Generalization
arXiv 2609.30271 · 2026-09-28HybridInfer：面向设备端、边缘和云端LLM推理的热感知强化学习分层路由
HybridInfer: Thermal-Aware Reinforcement-Learning Tier Routing for On-Device, Edge, and Cloud LLM Inference
arXiv 2609.30270 · 2026-09-28当少数异类触发数字采纳级联：企业主体基模型中网络可达性、阈值异质性与绩效条件复杂传染
When a Few Misfits Trigger Digital Adoption Cascades: Network Reach, Threshold Heterogeneity, and Performance-Conditioned Complex Contagion in an Agent-Based Model of Firms
arXiv 2609.30269 · 2026-09-28AD-WM：用于反事实模型预测控制的动作判别世界模型
AD-WM: Action-Discriminative World Models for Counterfactual Model Predictive Control
arXiv 2609.30264 · 2026-09-28AI引导的高通量发现无铱无钌的钯氧化物催化剂用于耐久的酸性析氧反应
AI-guided high-throughput discovery of iridium- and ruthenium-free palladium-oxide catalysts for durable acidic oxygen evolution
arXiv 2609.30133 · 2026-09-28扭转MoS$_2$中100 K以上的Pomeranchuk-like电子局域化
Pomeranchuk-like electronic localization above 100 K in twisted MoS$_{2}$
arXiv 2609.30128 · 2026-09-28C$_6$H 与 Na$^+$、Mg$^+$ 和 Al$^+$ 金属阳离子通过辐射缔合形成加合物
Formation of adducts of C$_6$H with Na$^+$, Mg$^+$ and Al$^+$ metal cations by radiative association
arXiv 2609.30126 · 2026-09-28PrivDrift：在活跃LLM对话中审计主题漂移下的用户秘密泄露
PrivDrift: Auditing User-Secret Leakage Under Topic Drift in Active LLM Conversations
arXiv 2609.30094 · 2026-09-28关于理想的反向族的约化模与协约化模
Reduced and coreduced modules with respect to inverse families of ideals
arXiv 2609.30091 · 2026-09-28PK/PD整合的贝叶斯平台设计用于II期剂量方案优化
PK/PD-integrated Bayesian platform design for phase II dose regimen optimization
arXiv 2609.30072 · 2026-09-28多样几何，冻结权重：通过因果专家集成实现稳健的异质性处理效应估计
Diverse Geometries, Frozen Weights: Robust Heterogeneous Treatment-Effect Estimation via Causal Expert Ensembles
arXiv 2609.29974 · 2026-09-28Galton-Watson树上偏置随机游走速度单调性在已知范围之外的计算机辅助证明
A Computer-Assisted Proof of Speed Monotonicity for the Biased Random Walk on a Galton-Watson Tree Beyond the Known Range
arXiv 2609.29894 · 2026-09-28两组分Becker-Döring系统的长时间行为
Long Time Behaviour of the Two-Component Becker-Döring System
arXiv 2609.29891 · 2026-09-28GPT-6-Astra 点亮具身导航：连续环境中零样本视觉-语言导航评估
GPT-6-Astra Lights Up Embodied Navigation: Evaluation in Zero-Shot Vision-and-Language Navigation in Continuous Environments
arXiv 2609.29861 · 2026-09-28双组分Becker-Döring系统：基本性质
The Two-Component Becker-Döring System: Basic Properties
arXiv 2609.29842 · 2026-09-28PUBG Ally：作为AI队友的对话式具身智能体
PUBG Ally: A Conversational Embodied Agent as an AI Teammate
arXiv 2609.29837 · 2026-09-28广义Saito准则：从线排列到曲线
A generalized Saito criterion, from line arrangements to curves
arXiv 2609.29824 · 2026-09-28连续音频编码器中潜在维度与帧率的联合分析
Joint Analysis of Latent Dimensionality and Frame Rate in Continuous Audio Encoders
arXiv 2609.29780 · 2026-09-28DAMSEP：基于多RIR估计的距离感知单声道源分离
DAMSEP: Distance-Aware Monaural Source Separation using Multi-RIR Estimation
arXiv 2609.29749 · 2026-09-28二聚体模型与一般权重的随机格点置换
Dimer model and random lattice permutations with general weights
arXiv 2609.29671 · 2026-09-28超越Harnack刚性
Beyond Harnack Rigidity
arXiv 2609.29655 · 2026-09-28关于 $E_0$-类泛型等价的两个问题
Two questions on $E_0$-like generic equivalence
arXiv 2609.29641 · 2026-09-28每个固定维度 $d\ge 4$ 中通过双着陆根包实现的周期长时间 Boltzmann--Grad 极限
A Periodic Long-Time Boltzmann--Grad Limit in Every Fixed Dimension $d\ge 4$
arXiv 2609.29566 · 2026-09-28ToCo-Mesh：通过自适应细分和表面对齐2DGS实现拓扑一致的动态网格重建
ToCo-Mesh: Topology-Consistent Dynamic Mesh Reconstruction via Adaptive Tessellation and Surface-Aligned 2DGS
arXiv 2609.29529 · 2026-09-28具有非标准增长的单相拟线性方程中的分岔现象
Bifurcation phenomena in one-phase quasilinear equations with nonstandard growth
arXiv 2609.29438 · 2026-09-28基于不确定性感知视觉Transformer的多族裔近视与非近视人群青光眼检测：一项多中心模型开发与验证研究
Detecting Glaucoma Across Multi-ethnic Myopic and Non-Myopic Populations Using an Uncertainty-Aware Vision Transformer: A Multicentre Model Development and Validation Study
arXiv 2609.29433 · 2026-09-28Rufus-Air：一种开放的大语言模型后训练配方
Rufus-Air: An Open LLM Post-Training Recipe
arXiv 2609.29421 · 2026-09-28全被动单片硅基量子光子电路用于纠缠光子对产生
Fully passive monolithic silicon quantum photonic circuit for entangled photon-pair generation
arXiv 2609.29420 · 2026-09-28似然排序在LLM中不像提示那样随规模扩展
Likelihood Ranking doesn't Scale Like Prompting in LLMs
arXiv 2609.29390 · 2026-09-28Free-Init：用于多普勒激光雷达-惯性系统的无扫描、无运动、无对应初始化
Free-Init: Scan-Free, Motion-Free, and Correspondence-Free Initialization for Doppler LiDAR-Inertial Systems
arXiv 2609.29375 · 2026-09-28FMCW-LIO：一种多普勒激光雷达-惯性里程计
FMCW-LIO: A Doppler LiDAR-Inertial Odometry
arXiv 2609.29374 · 2026-09-28带控制约束的Hamilton-Jacobi-Isaacs方程的策略迭代及其与Hamilton-Jacobi-Bellman方程的比较
Policy iteration for Hamilton-Jacobi-Isaacs equations with control constraints and comparison with Hamilton-Jacobi-Bellman equations
arXiv 2609.29368 · 2026-09-28
