Can I read Model-Free Control in RL: Lecture 5 on EtoBox?
Model-Free Control in RL: Lecture 5 by Fawaz Parto is a document available to read on EtoBox.
What is Model-Free Control in RL: Lecture 5 about?
This document discusses model-free reinforcement learning techniques for control problems with unknown models. It introduces on-policy Monte Carlo control which uses Monte Carlo returns to evaluate a policy and then improves the policy greedily. It also describes epsilon-greedy exploration to ensure continued exploration. Off-policy temporal difference learning is discussed as an alternative to Monte Carlo control that has lower variance and can operate online from incomplete sequences.
- Author
- Fawaz Parto
- Language
- EN