Propagating Uncertainty in Reinforcement Learning via Wasserstein Barycenters.
Alberto Maria MetelliAmarildo LikmetaMarcello RestelliPublished in: NeurIPS (2019)
Keyphrases
- reinforcement learning
- partial observability
- function approximation
- sequential decision problems
- temporal difference
- learning algorithm
- inherent uncertainty
- uncertain data
- incomplete information
- markov decision processes
- optimal policy
- supervised learning
- learning problems
- state space
- optimal control
- artificial intelligence
- real time
- conditional probabilities
- decision theory
- multi agent
- probability theory
- reinforcement learning algorithms
- robust optimization
- uncertain information
- knowledge base