cs.LG(2026-07-05)

📊 共 64 篇论文

🎯 兴趣领域导航

支柱二:RL算法与架构 (RL & Architecture) (32) 支柱九:具身大模型 (Embodied Foundation Models) (24) 支柱一:机器人控制 (Robot Control) (5) 支柱八:物理动画 (Physics-based Animation) (2) 支柱七:动作重定向 (Motion Retargeting) (1)

🔬 支柱二:RL算法与架构 (RL & Architecture) (32 篇)

#题目一句话要点标签🔗
1 MKGR: Multimodal Knowledge-Graph Representation Learning for Cold-Start Protein-Protein Interaction Prediction 提出MKGR以解决冷启动蛋白质相互作用预测问题 representation learning multimodal
2 A Survey of Circuit Foundation Model: Foundation AI Models for VLSI Circuit Design and EDA 提出电路基础模型以提升VLSI电路设计效率 representation learning large language model foundation model
3 When Do Conservation Laws Survive Learned Representations? Certified Horizons for Latent World Models 提出物理不变性认证方法以解决学习表示中的守恒法则问题 world model world models latent dynamics
4 Trust Region Inverse Reinforcement Learning: Explicit Dual Ascent using Local Policy Updates 提出信任区域逆强化学习以解决稳定性与效率问题 reinforcement learning imitation learning inverse reinforcement learning
5 Certified World Models as Sensing Clocks: Drift-Aware Deadlines for Active Perception 提出认证世界模型作为感知时钟以解决主动感知中的漂移问题 world model world models JEPA
6 Denser $\neq$ Better: Limits of On-Policy Self-Distillation for Continual Post-Training 提出自蒸馏策略优化以解决持续后训练中的遗忘问题 reinforcement learning policy learning distillation
7 Generalization in offline RL: The structure is more important than the amount of pessimism 提出对称性价值函数以优化离线强化学习中的泛化能力 reinforcement learning offline RL offline reinforcement learning
8 Composite Reward Design in PPO-Driven Adaptive Filtering 提出基于PPO的复合奖励设计以解决自适应滤波问题 reinforcement learning PPO reward design
9 Certified World Models: Predictability Across Configuration, Horizon, and Resolution 提出可认证的世界模型以解决预测可靠性问题 world model world models
10 Neuron-Aware Data Selection for Annotation-Free LLM Self-Distillation 提出Neuron-OPSD以解决无标注LLM自蒸馏问题 offline RL distillation large language model
11 DemoPSD: Disagreement-Modulated Policy Self-Distillation 提出DemoPSD以解决特权信息泄露与探索能力不足问题 distillation privileged information large language model
12 Evolutionary Feature Engineering for Structured Data 提出进化特征工程框架以优化结构化数据处理 MAE large language model foundation model
13 DecompRL: Solving Harder Problems by Learning Modular Code Generation 提出DecompRL以解决复杂问题的模块化代码生成 reinforcement learning world model world models
14 ThreadWeaver: Adaptive Threading for Efficient Parallel Reasoning in Language Models 提出ThreadWeaver以解决大语言模型推理效率问题 reinforcement learning large language model chain-of-thought
15 On the Sample Efficiency of Inverse Dynamics Models for Semi-Supervised Imitation Learning 提出逆动力学模型以提升半监督模仿学习的样本效率 policy learning imitation learning behavior cloning
16 Disentangled Latent Dynamics Manifold Fusion for Solving Parameterized PDEs 提出DLDMF以解决参数化PDEs的泛化与外推问题 latent dynamics spatiotemporal
17 Learning the Supports for Categorical Critic in Reinforcement Learning 提出动态学习支持区间的方法以解决值函数估计问题 reinforcement learning deep reinforcement learning
18 Mean Field Reinforcement Learning 提出均值场强化学习以解决大规模随机控制问题 reinforcement learning deep reinforcement learning
19 When Does Predictive Inverse Dynamics Outperform Behavior Cloning? 提出预测逆动力学模型以解决行为克隆样本效率低下问题 imitation learning behavior cloning
20 On the Role of Computation in Reinforcement Learning 提出计算受限策略以提升强化学习的学习能力 reinforcement learning offline RL
21 Liquid Latent State Dynamics for Interpretable Turbofan Degradation Modeling 提出液态潜在状态动态模型以解决涡扇发动机退化建模问题 world model world models latent dynamics
22 The Rollout Infrastructure Tax in Coding-Agent Reinforcement Learning 提出执行基础设施优化以提升编码代理强化学习效率 reinforcement learning
23 Wind-Aware Reinforcement Learning Control of a Small Quadrotor Using Learned Onboard Wind Estimation in Simulated Atmospheric Turbulence 提出基于风感知的强化学习控制方法以应对小型四旋翼在气流中的挑战 reinforcement learning
24 WARP: Weight-Space Analysis for Recovering Training Data Portfolios 提出WARP框架以恢复训练数据组合 MAE foundation model
25 Role-Aware Neural Convex Divergence Heads for Asymmetric Representation Learning 提出角色感知神经凸散度头以解决非对称表示学习问题 representation learning
26 Many Voices, One Reward: Multi-Role Rubric Generation for LLM Judging and Reward Modeling 提出多角色评分生成方法以解决单一评估者的盲点问题 reinforcement learning large language model
27 Rank-Then-Act: Reward-Free Control from Frame-Order Progress 提出Rank-Then-Act框架以解决无奖励控制问题 reinforcement learning policy learning reward design
28 One More Time: Revisiting Neural Quantum States from a Reinforcement Learning Perspective 提出PWO算法以优化神经量子态训练 reinforcement learning
29 WattGPU: Predicting Inference Power and Latency on Unseen GPUs and LLMs 提出WattGPU以解决LLM推理能耗与延迟预测问题 predictive model large language model
30 Staleness-Learning Rate Scaling Laws for Asynchronous RLHF 提出时效性学习率缩放法则以优化异步RLHF系统 RLHF
31 Is One Layer Enough? Training A Single Transformer Layer Can Match Full-Parameter RL Training 提出单层Transformer训练方法以优化强化学习效果 reinforcement learning large language model
32 Exact equivariance, kept through training, buys zero-shot generalisation across the symmetry group 提出对称性保持的潜在世界模型以实现零-shot泛化 world model world models

🔬 支柱九:具身大模型 (Embodied Foundation Models) (24 篇)

#题目一句话要点标签🔗
33 Zeus: Towards Tuning-Free Foundation Model for Time Series Analysis 提出Zeus以解决时间序列分析中的调优问题 foundation model
34 Probabilistic Low-Voltage Peak Load Forecasting with Time Series Foundation Models Evaluated on Application-Oriented Metrics 提出基于时间序列基础模型的低压负荷预测方法以解决不确定性问题 foundation model
35 Bayesian Sparse Low-Rank Adaptation for Large Language Model Uncertainty Estimation 提出DALorRA以解决大语言模型的不确定性估计问题 large language model
36 Scaling to Multimodal and Multichannel Heart Sound Classification with Synthetic and Augmented Biosignals 提出基于增强数据集的多模态心音分类方法以提高心血管疾病检测准确性 multimodal
37 Geometric Signatures of Reasoning: A Spectral Perspective on Task Hardness 提出几何特征分析以解决推理任务难度问题 large language model chain-of-thought
38 Program-as-Weights: A Programming Paradigm for Fuzzy Functions 提出模糊函数编程以解决传统编程任务的局限性 large language model foundation model
39 BOUNDARY_SYNC: Measuring Communication-Induced Representational Coupling in Multi-Agent LLM Systems 提出BOUNDARY_SYNC以测量多智能体LLM系统中的通信引发的表征耦合 large language model
40 SINA: A Fully Automated Circuit Schematic Image to Netlist Generator Using Artificial Intelligence 提出SINA以解决电路原理图自动转换为网表的问题 large language model
41 DeadPool: Resilient LLM Training with Hot-Swapping via Zero-Overhead Checkpoint 提出DeadPool以解决大规模语言模型训练中的故障恢复问题 large language model
42 SCAPE: Accurate and Efficient LLM Training with Extreme Sparse Communication 提出SCAPE以解决大规模语言模型训练中的通信效率问题 large language model
43 kNNGuard: Turning LLM Hidden Activations into a Training-Free Configurable Guardrail 提出kNNGuard以解决LLM安全性检测问题 large language model
44 Ask the Right Comparison:Bias-Aware Bayesian Active Top-$k$ Ranking with LLM Judges 提出偏见感知的贝叶斯主动Top-k排名方法以解决LLM评判偏差问题 large language model
45 QFedAgent: Quantum-Enhanced Personalized Federated Learning for Multi-Agent Activity Recognition 提出QFedAgent以解决多代理活动识别中的隐私与效率问题 multimodal
46 A global predicted-fMRI drive signal from TRIBE does not predict YouTube replay heatmaps 提出TRIBE模型以预测YouTube视频的观众重播行为 multimodal
47 Lynx: Progressive Speculative Quantization for accelerating KV Transfer in Long-Context Inference 提出Lynx以解决长上下文推理中的KV传输瓶颈问题 large language model
48 Bringing Agentic Search to Earth Observation Data Discovery 提出代理搜索系统以提升地球观测数据发现效率 large language model
49 LEFT: Learnable Fusion of Tri-view Tokens for Unsupervised Time Series Anomaly Detection 提出LEFT框架以解决无监督时间序列异常检测问题 TAMP
50 QTALE: Quantization-Robust Token-Adaptive Layer Execution for LLMs 提出QTALE以解决大语言模型的高效部署问题 large language model
51 Enabling KV Caching of Shared Prefix for Diffusion Language Models 提出双向前缀缓存技术以解决扩散语言模型的KV缓存问题 large language model
52 ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research 提出ResearchClawBench以评估自主科学研究能力 multimodal
53 From Lab to Reality: A Practical Evaluation of Deep Learning Models and LLMs for Vulnerability Detection 评估深度学习模型与LLMs在漏洞检测中的实际应用 large language model
54 FlexServe: A Fast and Secure LLM Serving System for Mobile Devices with Flexible Resource Isolation 提出FlexServe以解决移动设备LLM推理的安全与效率问题 large language model
55 FlexServe: A Fast and Secure LLM Serving System for Mobile Devices with Flexible Resource Isolation 提出FlexServe以解决移动设备LLM推理的安全与效率问题 large language model
56 Sample Complexities of Estimating Gumbel--Max Watermark Proportions with and without Reduction to Pivotal Statistics 提出水印比例估计方法以解决LLM生成文档的复杂性问题 large language model

🔬 支柱一:机器人控制 (Robot Control) (5 篇)

#题目一句话要点标签🔗
57 Predicting Closed-Loop Performance of Latent World Models: Offline Checkpoint Selection for MPC and Model-Based RL Under Non-Markovian Rewards in LunarLander 提出复合奖励可观测性分数以优化闭环性能预测 MPC world model world models
58 Gaming Consensus: Coordinated Manipulation in Crowdsourced Fact-Checking 提出协调操控机制以解决众包事实核查中的操控问题 manipulation
59 Privacy-Preserving and Verifiable Approximate Distributed Coded Computing 提出一种隐私保护与可验证的分布式编码计算框架以应对安全挑战 manipulation
60 Scalable and Distributed Silhouette Approximation 提出高效算法以解决大规模数据集的轮廓估计问题 MPC
61 Meta-Representational Predictive Coding: Neuroscience-Informed Self-Supervised Learning 提出元表征预测编码以解决自监督学习中的生物不合理性问题 MPC

🔬 支柱八:物理动画 (Physics-based Animation) (2 篇)

#题目一句话要点标签🔗
62 Do Physics Foundation Models Learn Generalizable Physics? A Bias-Aware Benchmark Across Physical Regimes and Distribution Shifts 构建偏差感知基准以评估物理基础模型的可泛化性 spatiotemporal foundation model
63 Quantifying the Uncertainty of Blindly Estimated Room Embeddings Using a Dispersion-Calibrated Score 提出一种框架以量化盲估房间嵌入的不确定性 PULSE

🔬 支柱七:动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗
64 SA-HGNN: Sample-Adaptive Hyperbolic Graph Neural Network for EEG-Based Depression Recognition 提出SA-HGNN以解决EEG抑郁症识别中的层次结构捕捉问题 spatial relationship

⬅️ 返回 cs.LG 首页 · 🏠 返回主页