The Little Book of Reinforcement Learning
A short, freely available introduction to reinforcement learning, from the basics through applied algorithms, with a PyTorch implementation of each algorithm covered and supplementary proofs for the dynamic-programming material.
The Little Book of Reinforcement Learning is a short book introducing reinforcement learning from first principles through to applied algorithms. Its GitHub repository holds the book itself along with supplementary material: PyTorch implementations of the algorithms it covers, from Monte Carlo methods to PPO, and detailed explanations and proofs for the dynamic-programming algorithms it covers only briefly in the text.
The repository is expected to grow with additional material over time, and a printable version of the book is available.
It is distributed under a non-commercial Creative Commons license (CC BY-SA 4.0).