Preface

As we enter the 21st century, it is quickly becoming apparent that AI and machine learning technologies will radically change the way we live our lives in the future. We now experience AI daily, from conversational assistants to smart recommendations in a search engine, and the average user/consumer now expects a smarter interface in anything they do. This most certainly includes games, and is likely one of the reasons why you, as a game developer, are considering reading this book.

This book will provide you, with a hands-on approach to building deep learning models for simple encoding for the purpose of building self-driving algorithms, generating music, and creating conversational bots, finishing with an in-depth discovery of deep reinforcement learning (DRL). We will begin with the basics of reinforcement learning (RL) and progress to combining DL and RL in order to create DRL. Then, we will take an in-depth look at ways to optimize reinforcement learning to train agents in order to perform complex tasks, from navigating hallways to playing soccer against zombies. Along the way, we will learn the nuances of tuning hyperparameters through hands-on trial and error, as well as how to use cutting-edge algorithms, including curiosity learning, Curriculum Learning, backplay, and imitation learning, in order to optimize agent training.