MER: Modular Element Randomization for robust generalizable policy in deep reinforcement learning.
Yihan LiJinsheng RenTianren ZhangYing FangFeng ChenPublished in: Knowl. Based Syst. (2023)
Keyphrases
- reinforcement learning
- optimal policy
- action selection
- policy search
- machine learning
- state space
- computationally efficient
- privacy preserving
- approximate dynamic programming
- data sets
- markov decision processes
- policy gradient
- markov decision process
- exploration exploitation tradeoff
- markov decision problems
- control policies
- asymptotically optimal
- partially observable
- infinite horizon
- learning process
- multi agent