CoMPS: Continual Meta Policy Search.
Glen BersethZhiwei ZhangGrace ZhangChelsea FinnSergey LevinePublished in: ICLR (2022)
Keyphrases
- policy search
- reinforcement learning
- continuous state
- dynamic programming
- continuous action
- reinforcement learning algorithms
- reward function
- computational complexity
- robot navigation
- partially observable markov decision processes
- markov decision problems
- policy gradient
- decision makers
- markov chain
- action selection