Reinforcement Learning Cheat Sheet
Starting a PhD means diving into an overwhelming amount of information. One of the main areas I’ll be focusing on over the coming years is Reinforcement Learning (RL). As a PhD student, I need to be ready to explain the key concepts of RL at a moment’s notice so, keeping my knowledge organized and fresh is essential.
To help with this, I created this cheat sheet. It’s my personal quick-reference guide, something I can pin up on the wall or pull out of my notebook whenever I need to recall the core components and algorithms of RL.
There are a few RL cheat sheets available online, but none exactly met my needs for fast, structured explanations — so I made my own. Starting from left to right and top to bottom, I’ve outlined:
- The basics of Reinforcement Learning: I include a simple system architecture, inspired by Sutton’s book and various scientific articles.
- Reinforcement Learning subgroups: The cheat sheet breaks RL into three main categories — Model Representation, Optimization Strategy, and Policy Type. I also provide concise explanations and figures to clarify how these subgroups function.
- Key RL algorithms: I highlight the primary algorithms used in RL, like Q-Learning, Deep Q-Learning, and Actor-Critic models, and summarize how they are applied.
This cheat sheet isn’t perfect or comprehensive — nor is it meant to be. It’s simply a tool that helps me explain the core concepts of RL in five minutes or less. It’s a living document, and I welcome any suggestions or improvements.
Feel free to share your thoughts! 🙂
Idea and Conceptualization: Rui Fernandes
Design: Eva Nizon