A Sharp Characterization of Linear Estimators for Offline Policy Evaluation.
Juan C. PerdomoAkshay KrishnamurthyPeter L. BartlettSham M. KakadePublished in: CoRR (2022)
Keyphrases
- policy evaluation
- least squares
- matrix inversion
- monte carlo
- temporal difference
- semi parametric
- variance reduction
- model free
- reinforcement learning
- markov decision processes
- policy iteration
- function approximation
- linear model
- statistical inference
- neural network
- optimal policy
- function approximators
- finite state
- fixed point
- decision making