A Free, Short Book on Reinforcement Learning — From Monte Carlo to PPO

Developer alxndrTL has released The Little Book of Reinforcement Learning, a compact introduction that takes readers from RL fundamentals through to applied algorithms. The companion GitHub repository pairs the text with runnable PyTorch implementations of the methods it covers, spanning Monte Carlo approaches up to Proximal Policy Optimization (PPO), so learners can move directly from theory to working code.

Beyond the core book, the repo ships supplementary material, including detailed explanations and formal proofs for the dynamic programming algorithms that the main text only touches on briefly — a document the author originally wrote in 2021. A print-your-own option is offered, and the maintainer notes that additional material will be added over time. Version 1 is dated June 2026.

The project is released under the non-commercial Creative Commons CC BY-SA 4.0 license, putting it alongside the growing set of free, open educational resources aimed at making machine learning theory more approachable. For engineers wanting a concise, proof-backed on-ramp to RL with accompanying reference implementations, it offers a self-contained starting point.