Active Exploration in Markov Decision Processes.
Jean TarbouriechAlessandro LazaricPublished in: AISTATS (2019)
Keyphrases
- markov decision processes
- active exploration
- reinforcement learning
- state space
- finite state
- active learning
- optimal policy
- small sample
- dynamic programming
- transition matrices
- policy iteration
- problem based learning
- decision theoretic planning
- partially observable
- markov decision process
- reachability analysis
- planning under uncertainty
- infinite horizon
- model based reinforcement learning
- reward function
- average cost
- action space
- state and action spaces
- average reward
- action sets
- programming concepts
- case study
- markov chain