cs.LG（2023-12-08）

📊 共 15 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) (6) 支柱九：具身大模型 (Embodied Foundation Models) (5 🔗1) 支柱一：机器人控制 (Robot Control) (2) 支柱五：交互与反应 (Interaction & Reaction) (1) 支柱八：物理动画 (Physics-based Animation) (1)

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

#	题目	一句话要点	标签	🔗	⭐
1	Understanding Community Bias Amplification in Graph Representation Learning	提出RGCCL模型，缓解图表示学习中的社区偏见放大问题	representation learning contrastive learning
2	Disentangled Latent Representation Learning for Tackling the Confounding M-Bias Problem in Causal Inference	提出DLRCE框架，解决因变量同时导致混淆偏差和M偏差的因果推断难题	representation learning
3	Modeling Risk in Reinforcement Learning: A Literature Mapping	构建安全强化学习风险模型：文献综述与风险特征分析	reinforcement learning
4	Backward Learning for Goal-Conditioned Policies	提出基于逆向学习的目标条件策略，实现无奖励强化学习。	reinforcement learning imitation learning world model
5	Pruning Convolutional Filters via Reinforcement Learning with Entropy Minimization	提出基于熵最小化的强化学习卷积滤波器剪枝方法，实现高效网络部署。	reinforcement learning
6	StructComp: Substituting Propagation with Structural Compression in Training Graph Contrastive Learning	提出StructComp，通过结构压缩加速图对比学习训练并提升性能	contrastive learning

🔬 支柱九：具身大模型 (Embodied Foundation Models) (5 篇)

#	题目	一句话要点	标签	🔗	⭐
7	Large-scale Training of Foundation Models for Wearable Biosignals	利用大规模可穿戴设备生物信号训练PPG/ECG基础模型，助力健康监测。	foundation model
8	EE-LLM: Large-Scale Training and Inference of Early-Exit Large Language Models with 3D Parallelism	EE-LLM：通过3D并行实现大规模早退大语言模型的训练与推理	large language model	✅
9	DeltaZip: Efficient Serving of Multiple Full-Model-Tuned LLMs	DeltaZip：高效服务多个全参数微调LLM，压缩模型增量达10倍。	large language model
10	SparQ Attention: Bandwidth-Efficient LLM Inference	SparQ Attention：通过选择性历史缓存，提升LLM推理带宽效率。	large language model
11	Make Them Spill the Beans! Coercive Knowledge Extraction from (Production) LLMs	提出模型审讯方法，通过操纵logits从LLM中强制提取隐藏的有害信息。	large language model

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
12	Learning 3D Particle-based Simulators from RGB-D Videos	提出Visual Particle Dynamics，从RGB-D视频中学习三维粒子模拟器	sim-to-real privileged information
13	DiSK: A Diffusion Model for Structured Knowledge	提出DiSK：一种用于结构化知识的扩散模型，提升表格数据建模、合成和补全性能。	manipulation

🔬 支柱五：交互与反应 (Interaction & Reaction) (1 篇)

#	题目	一句话要点	标签	🔗	⭐
14	AI Competitions and Benchmarks: Competition platforms	综述AI竞赛平台，分析其功能、经济模型和社区动态的差异。	ReMoS

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

#	题目	一句话要点	标签	🔗	⭐
15	Neural Spectral Methods: Self-supervised learning in the spectral domain	提出神经谱方法，通过谱域自监督学习高效求解参数化偏微分方程	spatiotemporal

⬅️ 返回 cs.LG 首页 · 🏠 返回主页