The Optimal Approximation Factors in Misspecified Off-Policy Value Function Estimation.
Philip AmortilaNan JiangCsaba SzepesváriPublished in: ICML (2023)
Keyphrases
- lp norm
- closed form
- error bounds
- piecewise linear
- dynamic programming
- constant factor
- factors affecting
- linear approximation
- worst case
- estimation error
- robust estimation
- factors influencing
- update equations
- set of basis functions
- single parameter
- control policy
- approximation methods
- approximation error
- pointwise
- approximation algorithms
- parameter estimation