2026-09论文中文摘要

• EasyPPO：稳定评论家是关键

  EasyPPO: Stabilizing the Critic Is Key

  arXiv 2609.36802 · 2026-09-30

• 场景重定向：利用类比迁移学习物体布局

  Scene Retargeting: Learning Object Placement with Analogical Transfer

  arXiv 2609.36801 · 2026-09-30

• AI 作为编译器：无需 Triton 编译器即可编译 Triton 内核

  AI as a Compiler: Compiling Triton kernels without the Triton compiler

  arXiv 2609.36800 · 2026-09-30

• 一类用于网络化双病毒SIS模型的分散式反馈控制器

  On a Class of Decentralized Feedback Controllers for the Networked Bivirus SIS Model

  arXiv 2609.36799 · 2026-09-30

• 听见应被看见的：全模态大语言模型中跨模态捷径的诊断与修复

  Seeing What Should Be Heard: Diagnosing and Repairing Cross-Modal Shortcuts in Omni-Modal LLMs

  arXiv 2609.36798 · 2026-09-30

• 随机性在神经细胞自动机中何处起作用？

  Where Does Randomness Matter in Neural Cellular Automata?

  arXiv 2609.36797 · 2026-09-30

• 冷暗物质晕中的动力学摩擦与次晕加热

  Dynamical friction vs. subhalo heating in Cold Dark Matter haloes

  arXiv 2609.36796 · 2026-09-30

• 认知类型作为 PostgreSQL 表访问方法：置信度伪造与 Sybil 协调下的对抗性冲突消解

  Epistemic Typing as a PostgreSQL Table Access Method: Adversarial Conflict Resolution Under Confidence Forgery and Sybil Coordination

  arXiv 2609.36795 · 2026-09-30

• RAE-PPG：面向PPG基础模型的持续时间锚定保留与扩展预训练

  RAE-PPG: Duration-Grounded Retain-and-Extend Pretraining for PPG Foundation Models

  arXiv 2609.36794 · 2026-09-30

• 耦合Stokes-Brinkman-Darcy流动的硬迹一阶残差学习

  Hard-Trace First-Order Residual Learning for Coupled Stokes-Brinkman-Darcy Flow

  arXiv 2609.36793 · 2026-09-30

• 社交网络中通过添加链接提升影响力排名

  Influence Ranking Improvement via Link Addition in Social Networks

  arXiv 2609.36791 · 2026-09-30

• 基于TDOA的同时传感器配对与重定位在线目标跟踪

  TDOA-Based Online Target Tracking with Simultaneous Sensor Pairing and Relocation

  arXiv 2609.36790 · 2026-09-30

• GitHarness：为持续的用户需求初始化你的工作记忆

  GitHarness: Git Init Your Harness Working Memory for Perpetual User Requirements

  arXiv 2609.36789 · 2026-09-30

• 利用大型语言模型将任务相关上下文编译进贝叶斯优化

  Harnessing Large Language Models to Compile Task-Relevant Context into Bayesian Optimisation

  arXiv 2609.36788 · 2026-09-30

• 正则化策略梯度与学习的高斯混合用于连续动作博弈

  Regularized policy gradient with learned mixtures of Gaussians for games with continuous actions

  arXiv 2609.36787 · 2026-09-30

• 基于球形随机场单一实现的对角各向异性检验

  Testing Diagonal Anisotropy Based on Single Realisation of Spherical Random Field

  arXiv 2609.36786 · 2026-09-30

• TaRL：从触觉演示中学习通用与物理奖励

  TaRL: Learning General and Physical Rewards from Tactile Demonstrations

  arXiv 2609.36785 · 2026-09-30

• 跨异构机械臂的尺度不变可操作性形状跟踪

  Scale-Invariant Manipulability Shape Tracking Across Heterogeneous Manipulators

  arXiv 2609.36784 · 2026-09-30

• 解码情感细微差别：通过层级情感推理与对比判别剪枝增强多模态大语言模型

  Decoding Affective Nuances: Enhancing MLLMs via Hierarchical Emotion Reasoning and Contrastive Discriminative Pruning

  arXiv 2609.36782 · 2026-09-30

• Aperture：压缩令牌的合并一致性旋转状态

  Aperture: Merge-Consistent Rotary States for Compressed Tokens

  arXiv 2609.36781 · 2026-09-30

• 化学复杂合金中分布型层错能量学的证据及其对变形的影响

  Evidence for Distributed Fault Energetics and Their Impact on Deformation in a Chemically Complex Alloy

  arXiv 2609.36780 · 2026-09-30

• DRHeC：基于RGB梯度的可微渲染手眼标定

  DRHeC: Differentiable Rendering for Hand-Eye Calibration with RGB-Based Gradients

  arXiv 2609.36779 · 2026-09-30

• Code4Scene：用于构建和编辑3D场景的编码智能体基准

  Code4Scene: Benchmarking Coding Agents for Constructing and Editing 3D Scenes

  arXiv 2609.36777 · 2026-09-30

• Causal-EVC：通过时空基础与反事实干预打破情感虚假因果

  Causal-EVC: Breaking Emotional Spurious Causality via Spatiotemporal Grounding and Counterfactual Intervention

  arXiv 2609.36776 · 2026-09-30

• DSPO：面向稳健情感推理的多样性感知主观策略优化

  DSPO: Diversity-aware Subjective Policy Optimization for Robust Emotional Reasoning

  arXiv 2609.36775 · 2026-09-30

• LexiconVLA：为未见任务学习可复用的原子动作码本

  LexiconVLA: Learning Reusable Atomic Action Codebooks for Unseen Tasks

  arXiv 2609.36774 · 2026-09-30

• NeuroDyn-EEG：基于神经动力学的可解释脑电图预训练模型

  NeuroDyn-EEG: An Interpretable Pre-trained Model for EEG Based on Neural Dynamics

  arXiv 2609.36773 · 2026-09-30

• 纽结的单值化特征的数量有限性

  Quantitative finiteness of monic characters of knots

  arXiv 2609.36772 · 2026-09-30

• 超越条件独立性：基于深度因果模型的根因分析

  Beyond Conditional Independence: Root Cause Analysis with Deep Causal Models

  arXiv 2609.36771 · 2026-09-30

• 无共享门控或跨智能体梯度的自监督协作视觉专家群体中的涌现特化

  Emergent Specialization in Populations of Self-Supervised Collaborative Vision Experts Without a Shared Gate or Cross-Agent Gradients

  arXiv 2609.36770 · 2026-09-30

• 毛细管液体闪烁体中子探测器的研制与性能研究

  Development and Performance Study of a Capillary Liquid Scintillator Neutron Detector

  arXiv 2609.36769 · 2026-09-30

• 测试KMTNet--PRIME光学--近红外源颜色约束在KMT-2024-BLG-0211和KMT-2024-BLG-1522中的应用

  Testing KMTNet--PRIME Optical--Near-Infrared Source-Color Constraints in KMT-2024-BLG-0211 and KMT-2024-BLG-1522

  arXiv 2609.36768 · 2026-09-30

• Schwarzschild时空中闭合测地线的多叶拓扑与热点图像时空自相关的对应关系

  Correspondence between multileaf topology of closed geodesics and spatiotemporal autocorrelations of hotspot images in Schwarzschild spacetime

  arXiv 2609.36767 · 2026-09-30

• 前缀何时能编译 LoRA？冻结注意力的精确资源上限测试

  When Can Prefixes Compile LoRA? Exact Resource-Capped Tests for Frozen Attention

  arXiv 2609.36766 · 2026-09-30

• 图谱流匹配用于多元时间序列异常检测

  Graph-Spectral Flow Matching for Multivariate Time Series Anomaly Detection

  arXiv 2609.36765 · 2026-09-30

• AdS$_4$ 中的 Dirichlet 引力子与 $\mathcal{L}_\Lambda w_{1+\infty}$ 楔形代数

  Dirichlet gravitons in AdS$_4$ and $\mathcal{L}_Λw_{1+\infty}$ wedge algebra

  arXiv 2609.36764 · 2026-09-30

• SCORAS-MoE：LEO卫星网络中MoE-VLMs的联合压缩与资源自适应部署

  SCORAS-MoE: Joint Compression and Resource-Adaptive Deployment of MoE-VLMs in LEO Satellite Networks

  arXiv 2609.36763 · 2026-09-30

• 联邦聚类中的未知局部与全局簇数量

  Federated Clustering with Unknown Local and Global Cluster Cardinalities

  arXiv 2609.36762 · 2026-09-30

• 超图中阿贝尔覆盖的范围

  The Reach of Abelian Covers in Hypergraphs

  arXiv 2609.36761 · 2026-09-30

• 双模式低秩学习器与桥接原型集成用于视觉-语言类增量学习

  Dual-Mode Low-Rank Learner with Bridge-Prototype Ensemble for Vision-Language Class-Incremental Learning

  arXiv 2609.36759 · 2026-09-30

• NaSICON 薄固体电解质的离子束合成与改性

  Synthesis and modification of thin NaSICON solid electrolytes using ion beams

  arXiv 2609.36758 · 2026-09-30

• FastVR：基于一步扩散的高效流式视频修复

  FastVR: Efficient Streaming Video Restoration with One-Step Diffusion

  arXiv 2609.36757 · 2026-09-30

• 拖拽作为证据：基于运动锚定的潜在重组用于拖拽编辑

  Drag as Evidence: Motion-Grounded Latent Recomposition for Drag-Based Editing

  arXiv 2609.36755 · 2026-09-30

• 韵律训练表示是否有助于超越可训练融合？一项与冻结HuBERT的参数匹配研究

  Does a prosody-trained representation help beyond trainable fusion? A parameter-matched study with frozen HuBERT

  arXiv 2609.36754 · 2026-09-30

• 用于LiDAR SLAM的简并正交几何约束

  Degeneracy-Orthogonal Geometric Constraints for LiDAR SLAM

  arXiv 2609.36753 · 2026-09-30

• cktFormer：基于Transformer的模拟电路自动设计方法

  cktFormer: Transformer-Based Approach for Automated Analog Circuit Design

  arXiv 2609.36752 · 2026-09-30

• 澳大利亚能源部门可靠性风险度量

  Measures of Reliability Risk for the Australian Energy Sector

  arXiv 2609.36751 · 2026-09-30

• 分组边缘化自奖励强化学习驱动零标签自我进化

  Group-Marginalized Self-Rewarding RL Drives Zero-Label Self-Evolving

  arXiv 2609.36750 · 2026-09-30

• LAGP涂层聚乙烯隔膜用于抑制薄型锂硫电池中多硫化物穿梭效应的研究

  Study of LAGP coating on polyethylene separator for polysulfide suppression in thin Li-S batteries

  arXiv 2609.36749 · 2026-09-30

• 可泛化的终身模型编辑：基于偏好优化

  Generalizable Lifelong Model Editing via Preference Optimization

  arXiv 2609.36748 · 2026-09-30