cs.LG（2024-06-28）

📊 共 15 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (7 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (6) 支柱一：机器人控制 (Robot Control) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (7 篇)

#题目一句话要点标签🔗⭐
1 CHASE: A Causal Hypergraph based Framework for Root Cause Analysis in Multimodal Microservice Systems 提出CHASE框架，利用因果超图解决多模态微服务系统中的根因分析问题 multimodal
2 InfiniGen: Efficient Generative Inference of Large Language Models with Dynamic KV Cache Management InfiniGen：通过动态KV缓存管理实现大语言模型的高效生成式推理 large language model
3 Enhancing Stability for Large Language Models Training in Constrained Bandwidth Networks 提出改进的ZeRO++算法，解决低带宽网络下大语言模型训练的收敛性问题 large language model
4 ScaleBiO: Scalable Bilevel Optimization for LLM Data Reweighting 提出ScaleBiO，通过可扩展双层优化实现LLM数据重加权，显著提升指令跟随和数学推理能力。 large language model instruction following
5 ProgressGym: Alignment with a Millennium of Moral Progress 提出ProgressGym框架，用于学习和模拟人类道德进步，解决AI对社会价值观的潜在负面影响。 large language model ✅
6 LLMEasyQuant: Scalable Quantization for Parallel and Distributed LLM Inference LLMEasyQuant：面向并行和分布式LLM推理的可扩展量化框架 large language model
7 A Survey on Data Quality Dimensions and Tools for Machine Learning 综述机器学习数据质量评估与改进工具，并展望LLM的应用前景 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
8 Edge-DIRECT: A Deep Reinforcement Learning-based Method for Solving Heterogeneous Electric Vehicle Routing Problem with Time Window Constraints 提出Edge-DIRECT，基于深度强化学习解决带时间窗约束的异构电动汽车路径优化问题 reinforcement learning deep reinforcement learning DRL
9 Operator World Models for Reinforcement Learning 提出基于算子世界模型的强化学习算法POWR，解决策略镜像下降法在强化学习中的应用难题。 reinforcement learning world model
10 Towards Stable and Storage-efficient Dataset Distillation: Matching Convexified Trajectory 提出匹配凸化轨迹(MCT)方法，解决数据集蒸馏中训练轨迹匹配的不稳定性和存储效率问题。 distillation large language model
11 TabSketchFM: Sketch-based Tabular Representation Learning for Data Discovery over Data Lakes TabSketchFM：提出基于草图的表格表示学习方法，用于数据湖中的数据发现。 representation learning
12 Reinforcement Learning for Efficient Design and Control Co-optimisation of Energy Systems 提出基于强化学习的能源系统设计与控制协同优化框架，提升可再生能源利用率。 reinforcement learning
13 LLM Critics Help Catch LLM Bugs 利用LLM评论员辅助发现LLM代码缺陷，提升人工评估准确性 reinforcement learning RLHF

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
14 Modeling the Real World with High-Density Visual Particle Dynamics 提出高密度视觉粒子动力学模型，用于模拟真实场景物理动态 bi-manual world model linear attention
15 Model Predictive Simulation Using Structured Graphical Models and Transformers 提出基于Transformer和概率图模型的模型预测模拟方法，提升多智能体交互场景的安全性。 MPC model predictive control

⬅️ 返回 cs.LG 首页 · 🏠 返回主页