Conservative Exploration for Policy Optimization via Off-Policy Policy Evaluation.
Paul DaoudiMathias FormosoOthman GaiziAchraf AzizeEvrard GarcelonPublished in: CoRR (2023)
Keyphrases
- policy evaluation
- least squares
- temporal difference
- policy iteration
- monte carlo
- reinforcement learning
- model free
- markov decision processes
- function approximation
- action selection
- optimal policy
- variance reduction
- optimization algorithm
- partially observable markov decision processes
- semi parametric
- reinforcement learning algorithms
- td learning
- evaluation function
- decision making
- constrained optimization
- statistical inference
- gaussian process
- radial basis function
- markov chain
- markov decision problems
- semi supervised