Mastering Reinforcement Learning with Python
Mastering Reinforcement Learning with Python was published by Packt in December 2020. It's 18 chapters and roughly 500 pages, covering the full spectrum of reinforcement learning from bandits and dynamic programming through deep RL, multi-agent systems, and model-based methods.
The book is designed for practitioners. Every chapter includes Python implementations using TensorFlow, Ray/RLlib, and OpenAI Gym, with examples drawn from real-world domains: robotics, supply chain, finance, marketing, and cybersecurity. It's been an Amazon bestseller in its category.
I'm currently working on a second edition that will cover the RL developments since 2020, including RLHF, LLM alignment, and modern policy optimization methods.
Table of Contents
- Introduction to Reinforcement Learning
- Multi-Armed Bandits
- Contextual Bandits
- Tick-Tack-Toe and Dynamic Programming
- Tabular Reinforcement Learning
- Function Approximation
- Deep Q-Networks
- REINFORCE and Policy Gradient Methods
- Actor-Critic Methods
- Derivative-Free Optimization
- Model-Based Methods
- Multi-Agent Reinforcement Learning
- Parallel Methods
- Automated Reinforcement Learning
- RL in Robotics
- RL in Supply Chain and Finance
- RL in Marketing
- RL in Cybersecurity