PEAC: Unsupervised Pre-training for Cross-Embodiment Reinforcement Learning.
Chengyang YingZhongkai HaoXinning ZhouXuezhou XuHang SuXingxing ZhangJun ZhuPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- supervised learning
- unsupervised learning
- supervised training
- learning process
- training set
- function approximation
- training samples
- supervised methods
- training examples
- training process
- neural network
- labeled data
- data driven
- optimal control
- semi supervised
- supervised classification
- training phase
- temporal difference
- unsupervised manner
- temporal difference learning
- state space
- artificial neural networks
- deep architectures