Explore model-based reinforcement learning in this comprehensive lecture, covering key concepts such as introduction, examples, pseudocode, complex models, and comparisons with replay buffers. Delve into practical applications of Dyna, search tree techniques, planning strategies, and Monte Carlo methods. Gain a deep understanding of how model-based RL differs from other approaches and learn to implement these concepts in real-world scenarios.
Overview
Syllabus
Modelbased RL
Introduction
Modelbased Reinforcement Learning
Example
Summary
Pseudocode
Complex models
Modelbased RL vs Replay Buffer
Discussion
Dyna
Dyna in practice
Search tree
Planning
Monte Carlo
Taught by
Pascal Poupart
