Task-Driven Reinforcement Learning With Action Primitives for Long-Horizon Manipulation Skills.
Hao WangHao ZhangLin LiZhen KanYongduan SongPublished in: IEEE Trans. Cybern. (2024)
Keyphrases
- reinforcement learning
- action primitives
- action recognition
- function approximation
- human activity recognition
- learning algorithm
- reinforcement learning algorithms
- model free
- state space
- activity recognition
- hidden markov models
- temporal difference
- mixture distributions
- markov decision processes
- transfer learning
- video data
- dynamic programming
- spatio temporal
- multi agent