Book Image

Hands-On Reinforcement Learning for Games

By : Micheal Lanham
Book Image

Hands-On Reinforcement Learning for Games

By: Micheal Lanham

Overview of this book

With the increased presence of AI in the gaming industry, developers are challenged to create highly responsive and adaptive games by integrating artificial intelligence into their projects. This book is your guide to learning how various reinforcement learning techniques and algorithms play an important role in game development with Python. Starting with the basics, this book will help you build a strong foundation in reinforcement learning for game development. Each chapter will assist you in implementing different reinforcement learning techniques, such as Markov decision processes (MDPs), Q-learning, actor-critic methods, SARSA, and deterministic policy gradient algorithms, to build logical self-learning agents. Learning these techniques will enhance your game development skills and add a variety of features to improve your game agent’s productivity. As you advance, you’ll understand how deep reinforcement learning (DRL) techniques can be used to devise strategies to help agents learn from their actions and build engaging games. By the end of this book, you’ll be ready to apply reinforcement learning techniques to build a variety of projects and contribute to open source applications.
Table of Contents (19 chapters)
1
Section 1: Exploring the Environment
7
Section 2: Exploiting the Knowledge
15
Section 3: Reward Yourself

Advancing RL with ML-Agents

The ML-Agents toolkit, the part that allows you to train DRL agents, is considered one of the more serious and top-end frameworks for training agents. Since the framework was developed on top of Unity, it tends to perform better on Unity-like environments. However, not unlike many others who spend time training agents, the Unity developers realized early on that some environments present such difficult challenges as to require us to assist our agents.

Now, this assistance is not so much direct but rather indirect and often directly relates to how easy or difficult it is for an agent to find rewards. This, in turn, directly relates to how well the environment designer can build a reward function that an agent can use to learn an environment. There are also the times when an environment's state space is so large and not obvious that creating a typical...