Название: Reinforcement Learning: Theory and Python Implementation
Автор: Zhiqing Xiao
Издательство: Springer/China Machine Press
Год: 2024
Страниц: 574
Язык: английский
Формат: pdf (true)
Размер: 10.1 MB

Reinforcement Learning: Theory and Python Implementation is a tutorial book on reinforcement learning, with explanations of both theory and applications. Starting from a uniform mathematical framework, this book derives the theory of modern reinforcement learning systematically and introduces all mainstream reinforcement learning algorithms such as PPO, SAC, and MuZero. It also covers key technologies of GPT training such as RLHF, IRL, and PbRL. Every chapter is accompanied by high-quality implementations, and all implementations of deep reinforcement learning algorithms are with both TensorFlow and PyTorch. Codes can be found on GitHub along with their results and are runnable on a conventional laptop with either Windows, macOS, or Linux.

Reinforcement Learning (RL) is a type of Artificial Intelligence (AI) that changes our lives: RL players have defeated human in many games such as the game of Go and StarCraft; RL controllers are driving varied robots and unmanned vehicles; RL traders are making tons of money in financial markets, and the large language model with RL such as ChatGPT have been used in many business applications. Since the same RL algorithm with the same parameter setting can solve very different tasks, RL is also regarded as an important way to general AI. Here I sincerely invite you to learn RL to surf in these AI waves.

This book is a tutorial on RL, with explanation of theory and Python implementation.
It consists of the following three parts:

• Chapter 1: Introduce the background of RL from scratch, and introduce the environment library Gym.
• Chapters 2–14: Introduce the mainstream RL theory and algorithms. Based on the most influential RL model–discounted return discrete-time Markov decision process, we derive the fundamental theory mathematically. Upon the theory we introduce algorithms, including both classical RL algorithms and deep RL algorithms, and then implement those algorithms in Python.
• Chapters 15–16: Introduce other RL models and extensions of RL models, including average-reward, continuous-time, non-homogenous, semi-Markov, partial observability, preference-based RL, and imitation learning, to have a complete understanding of the landscape of RL and its extension.

This book is intended for readers who want to learn reinforcement learning systematically and apply reinforcement learning to practical applications. It is also ideal to academical researchers who seek theoretical foundation or algorithm enhancement in their cutting-edge AI research.