Doubly Robust Estimator for Off-Policy Evaluation with Large Action Spaces.
Tatsuhiro ShimizuLaura ForastierePublished in: CoRR (2023)
Keyphrases
- action space
- markov decision processes
- policy iteration
- reinforcement learning
- temporal difference
- action selection
- state space
- markov decision process
- model free
- finite state
- optimal policy
- reinforcement learning algorithms
- real valued
- markov decision problems
- function approximation
- infinite horizon
- dynamic programming
- random sampling
- variance reduction
- partially observable
- np hard
- decision making
- transition probabilities
- linear programming
- least squares
- reward function
- image segmentation
- learning algorithm