cs.LG（2026-09-17）

📊 共 37 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (18) 支柱九：具身大模型 (Embodied Foundation Models) (15) 支柱一：机器人控制 (Robot Control) (2) 支柱八：物理动画 (Physics-based Animation) (1) 支柱四：生成式动作 (Generative Motion) (1 🔗1)

🔬 支柱二：RL算法与架构 (RL & Architecture) (18 篇)

#题目一句话要点标签🔗⭐
1 Conservation Buys Stability and Factoring Buys Counterfactuals in Physical World Models 提出结构化方法以解决物理世界模型的稳定性与反事实问题 world model world models physically plausible
2 REARL: A Closed-loop Autonomous Driving Simulation Enhancement Framework with Real Traffic Data and Large Language Models 提出REARL框架以解决自动驾驶仿真中的真实交通数据整合问题 PPO large language model
3 Video DeltaNet: A Video-Native Hybrid Attention for Livestream Video Generation 提出Video DeltaNet以解决视频生成中的计算瓶颈问题 linear attention distillation spatiotemporal
4 Score Centering Stabilizes Off-policy Reinforcement Learning 提出得分中心化以稳定离线强化学习 reinforcement learning large language model
5 Learn Your Own Thoughts: Abstract Token Curriculum 提出抽象标记课程以提升大语言模型的思维能力 curriculum learning large language model chain-of-thought
6 MATCH: Model-Aware Tool Learning with Curriculum Scheduling and Hierarchically Gated Rewards 提出MATCH框架以解决工具学习中的课程调度与奖励问题 reinforcement learning curriculum learning large language model
7 When EOS Tokens Disagree: Understanding Length Inflation in On-Policy Distillation 提出终止标记对齐方法以解决OPD中的长度膨胀问题 distillation
8 Seismic Site Response Prediction from Sparse Observations Using Finite-Element-Pretrained Latent Dynamics 提出FLARE-T以改善稀疏观测下的地震场地响应预测 latent dynamics
9 Model-based Bootstrap for Offline Policy Evaluation in Tabular Reinforcement Learning 提出基于模型的自助法以解决离线策略评估中的不确定性量化问题 reinforcement learning
10 EPIG-Tree: Compute-Optimal Branching for Gradient-Efficient Reinforcement Learning 提出EPIG-Tree以解决强化学习中的奖励分配效率问题 reinforcement learning
11 DeliveryGym: An RL Environment for Long-Horizon Embodied Agent Planning with Adaptive Curriculum 提出DeliveryGym以解决长时间规划中的资源管理问题 reinforcement learning multimodal
12 NS3Learn: Transferring 5G NR Mode-2 Reception Realism from ns-3 to the Veins/SUMO Stack for Connected-Vehicle Safety Assessment 提出NS3Learn以解决5G NR Mode-2接收现实性问题 distillation penetration
13 Mitigating Retaliatory Algorithmic Collusion in Repeated Games 提出CURB框架以解决重复博弈中的报复性算法共谋问题 reinforcement learning reward shaping
14 CARE-VI: Conservative Adaptive Reliability Estimation for Value Improvement in Off-Policy Actor-Critic Learning 提出CARE-VI以解决离线策略演员-评论家学习中的可靠性估计问题 SAC TD3
15 ASGARD: Action-Space Guard for UAV Resilience via Reinforcement Learning 提出ASGARD以解决无人机控制中的动作空间攻击问题 reinforcement learning teacher-student privileged information
16 REFINEPPO: Learning Continuous Control Policies by Iterative Action Refinement 提出REFINEPPO以解决连续控制策略迭代改进问题 reinforcement learning deep reinforcement learning DRL
17 Efficient Bayes-Adaptive Reinforcement Learning with Temporal Logic Specifications 提出高效的贝叶斯自适应强化学习以满足时间逻辑规范 reinforcement learning
18 Continuous Delayed-Memory Stochastic Gradient Descent and Continuous-Time Reinforcement Learning from History of Astrophysical Time Series Studies 提出连续延迟记忆随机梯度下降以解决天体物理时间序列建模问题 reinforcement learning

🔬 支柱九：具身大模型 (Embodied Foundation Models) (15 篇)

#题目一句话要点标签🔗⭐
19 SCGFM-ART: Amortized Relational Transport for Structure-Centric Graph Foundation Models 提出SCGFM-ART以解决异构图领域的统一表示问题 foundation model
20 QoS-Aware Federated Learning for Multimodal In-Cabin Interaction in Smart Vehicles 提出FedQoS框架以解决智能车辆中多模态交互的QoS问题 multimodal
21 AURA: Adaptive Uncertainty-Routed Analysis for Email Threat Detection 提出AURA以解决邮件威胁检测中的不确定性问题 large language model multimodal
22 PosteriorBench: From Point Estimates to Posterior Matching in Evaluating Generative Inverse Solvers 提出PosteriorBench以评估生成逆解算器的后验匹配 multimodal
23 Parallelism, critical windows, and separations among diffusion language models 提出精细比较以揭示扩散语言模型的并行性差异 large language model
24 Detecting Deceptive Recruitment: A Signal-theoretic Machine Learning Framework for Early Identification of Labour Exploitation 提出信号理论机器学习框架以检测欺诈性招聘问题 multimodal
25 QUALS: Corpus Equilibrium for Universal Forecasting via Pattern Quantization and Learnability Synchronization 提出QUALS框架以解决时间序列数据多样性问题 foundation model
26 Local Sparsity Enables Unsupervised LLM Safety Detection 提出局部稀疏方法以解决无监督LLM安全检测问题 large language model
27 Past, Future, All at Once: Mitigating Stability-Plasticity Dilemma via Post-hoc JANUS Rectification 提出后处理JANUS修正以解决稳定性-可塑性困境 foundation model
28 The Life of a Token: from Words to Bits on the Wire 探讨语言模型训练中的数据流动与网络需求 large language model
29 ALIBI: Adversarial Legitimacy Injection in Binary Input against LLM Malware Analyzers 提出ALIBI以解决LLM恶意软件分析中的攻击面问题 large language model
30 From Stress to Affect: Multimodal Deep Learning for Physiological Emotion Recognition Across Wearable Sensor Modalities 提出多模态深度学习方法以提升生理情感识别性能 multimodal
31 Complex Problem Solving in Large Language Models: A Statistical Control Survey and Diagnostic Framework 提出统计控制框架以解决大型语言模型的复杂问题解决挑战 large language model
32 When AI Reviews Train AI Reviewers: Scientific-Judgment Collapse and Mitigation 提出TrustReviewer以解决科学评审中的判断崩溃问题 large language model
33 Parallelism, critical windows, and separations among diffusion language models 提出对比分析以揭示扩散语言模型的并行性差异 large language model

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
34 Uni-LaDiR: Latent Diffusion Unifies Multimodal Reasoning 提出Uni-LaDiR以解决多模态推理中的信息整合问题 manipulation vision-language-action VLA
35 Delphi Scanner: efficient and interpretable static malware detection via API sequence modeling 提出Delphi Scanner以解决静态恶意软件检测效率与可解释性问题 manipulation

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#题目一句话要点标签🔗⭐
36 Training Neural Networks to Approach the Optimum Bayes Estimator in Dense Multi-Emitter Localization 提出神经网络训练方法以优化密集发射器定位 spatiotemporal

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
37 Bio-MF: Low-Latency and High-Fidelity EEG-to-fNIRS Cross-Modal Generation for Hybrid Motor-Imagery Brain--Computer Interfaces 提出Bio-MF以解决EEG到fNIRS生成的低延迟与高保真问题 classifier-free guidance ✅

⬅️ 返回 cs.LG 首页 · 🏠 返回主页