SIRL: Self-Imitation Reinforcement Learning for Single-step Hitting Tasks.
Yongle LuoYuxin WangKun DongYu LiuZhiyong SunQiang ZhangBo SongPublished in: ICARM (2023)
Keyphrases
- single step
- reinforcement learning
- transfer learning
- multi step
- function approximation
- markov decision processes
- machine learning
- state space
- multiple tasks
- optimal control
- complex domains
- real robot
- neural network
- multi agent environments
- markov chain
- mobile robot
- multi agent
- action selection
- learning algorithm
- reinforcement learning agents
- transferring knowledge