Tutorials for Delta Academy: Introduction to Reinforcement Learning.
Preface: what these tutorials are and who they are for
These tutorials are designed for Delta Academy. It’s a 4-week cohort-based course where every week you compete in an AI-building competition against the rest of the cohort.
The focus of the tutorials is therefore practical. There are interactive coding exercises in Python, with solutions to compare against. The necessary theoretical background is provided, without being too onerous.
The tutorials are appropriate for someone with introductory coding ability and basic maths, who wants to dive into Reinforcement Learning.
They are best studied as part of the Delta Academy cohort, where weekly competitions and discussions with a cohort of others keep you motivated and on track.
1 - Motivation, States, Actions and Rewards
2 - Return, Value Functions & Bellman Equations
3 - Learning from Experience, TD-Learning, $\epsilon$-Greedy
4 - Generalised Policy Iteration
5 - Curse of Dimensionality, Function Approximation & Parameters