Optimistic Exploration with Learned Features Provably Solves Markov Decision Processes with Neural Dynamics.
Sirui ZhengLingxiao WangShuang QiuZuyue FuZhuoran YangCsaba SzepesváriZhaoran WangPublished in: ICLR (2023)
Keyphrases
- markov decision processes
- model based reinforcement learning
- neural dynamics
- finite state
- dynamic programming
- state space
- reinforcement learning
- transition matrices
- decision theoretic planning
- optimal policy
- finite horizon
- average cost
- interval estimation
- planning under uncertainty
- markov decision process
- policy iteration
- partially observable
- average reward
- reachability analysis
- factored mdps
- action sets
- reward function
- state and action spaces
- infinite horizon
- action recognition
- linear programming
- collaborative filtering