Incorporating Explicit Uncertainty Estimates into Deep Offline Reinforcement Learning.
David BrandfonbrenerRemi Tachet des CombesRomain LarochePublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- sequential decision problems
- function approximation
- partial observability
- confidence intervals
- markov decision processes
- machine learning
- model free
- uncertain data
- reinforcement learning algorithms
- measurement error
- optimal policy
- state space
- inherent uncertainty
- estimation error
- belief functions
- decision theory
- temporal difference
- learning capabilities
- temporal difference learning
- learning process
- real time