Pessimistic Bootstrapping for Uncertainty-Driven Offline Reinforcement Learning.
Chenjia BaiLingxiao WangZhuoran YangZhihong DengAnimesh GargPeng LiuZhaoran WangPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- expected utility
- possibility theory
- markov decision processes
- information extraction
- function approximation
- partial observability
- decision theory
- data driven
- uncertain data
- multi agent
- state space
- real time
- temporal difference
- sequential decision problems
- optimal policy
- robotic control
- action space
- multi agent reinforcement learning
- uncertain information
- decision theoretic
- minimally supervised
- markov decision process
- partially observable
- reinforcement learning algorithms
- belief functions
- incomplete information
- utility function
- decision makers
- machine learning