Welcome to the AI 101 series! This collection of written pieces and accompanying videos is aimed at introducing you to how artificial intelligence (AI) works. We do this primarily through video games and this is for two specific reasons:
- We can relate a lot of the theory and practice behind AI to the world of video games.
- AI receives a lot of exposure in video games: either being used in commercial games, or academic research which explores how to play games using AI.
The series is primarily theory-based and we will use games as examples to help relate what we are writing about and put it all into context. In addition, these articles help reinforce some of the theory that is often discussed in brief during the ‘AI & Games’ series, which often provides both a primer to a methodology, followed by a case study of its application within commercial games.
In this first article – a foreword if you will – we want to get you in the right frame of mind. To be thinking about the key issues and also get you prepared for the articles ahead.
An Introduction From Tommy
Hello and welcome to my ‘AI 101’ series: I appreciate you taking the time to come along and read through this work and I hope you find it useful. My goal is rather straightforward: to help educate people as to how AI works, utilising games as a backdrop. In addition, these articles work alongside my ‘AI and Games’ case studies, as well as the ‘AI DIY’ series, to help you see how this comes together: both in developing your own AI and to understand how AI works in commercial and research applications.
The one thing I hope you appreciate before reading further is that this series is not intended to replace more scholarly sources on the subject. I want this to be something fun and a little bit novel: publicly available sources that are accessible and informative for those who are both active students of AI, but also for wanderers of the internet to stumble upon. It’s meant to be fun and will be prone to random Mario pictures to prove my point.
If you are actually studying Artificial Intelligence as a class in undergraduate study, first of all congratulations! I hope you come to enjoy this area of study as much as I do. However, it is not advised to use my articles either as your sole source of information, nor as a citation in a submitted piece of work. With that in mind, many of my articles either directly cite textbooks or research articles such that those who need scholarly citations can find those that are relevant to the topic at hand. Treat these articles in much the same way as you would Wikipedia: don’t cite the page, instead explore and potentially cite the sources it is reliant upon.
There are a large number of high-quality publications that are available courtesy in book stores (both offline and online) that will help you get a grasp of the foundational theory of AI, as well as some indication of how to develop your own AI software. I recommend a handful of books later in this piece, but don’t be afraid to explore on your own.
This is all a little experimental for me. However, I hope that you enjoy this series as it develops. I am always keen to hear feedback (good and bad) to either improve the quality of these pieces or help direct it for the audience. Either way. Enjoy and let’s get to it!
For Those Who Wish to Study More
This series aims to be as open and as accessible as possible. However, it’s important to appreciate that there are other people throughout game development and academia who do just as good a job, if not better, than I will at covering the techincal aspects. This can be for a number of reasons: years of experience, better access to resources or, in some elements of technical detail, because they invented the method. Let’s face it, they know more about this AI system than I ever will!
My goal is accessibility, not necessarily plunging into a tremendous depth of detail. So, I list some important books below for students getting started. While these are the ‘generalist’ books, more will appear throughout the series as we focus on specific topics.
Note: If anyone has suggestions that have not made this list, please contact me with your suggestions!
I don’t think there is a better book to start with than ‘Russell & Norvig’ as it is often referred to. This is arguably ‘the Bible’ of AI reading. If you’re starting out and really want to get a grasp of where AI, as a science, currently stands in the 21 century, this is a must-have.
While ‘Russell and Norvig’ gives a broad overview of theory, ‘AI for Games’ focusses more on the practical applications developers are reliant upon for games. However, I don’t find it good at conveying the theory as well as the previous book. As such, it is often best to have these books sitting close to one another for reference purposes!
“The AI Game Programming Wisdom” series.Edited by Steve Rabin. Charles River Media.
The ‘AI Wisdom’ series are a little old now, given first volume in the series was published in 2002. However, their relevance still carries to date, given that the chapters of the series are often written by AI programmers from the games industry, with the odd researcher or two in there to boot! These have since been succeeded by the ‘Game AI Pro’ series, also edited by Rabin. However, I have yet to read through the first book in the series and will update if necessary.
“But this isn’t about AI!” I hear you say. Yes, that’s true. However, it does provide some really interesting analysis about how games are designed and made based upon a number of characteristics, metrics and historical observations.
AIGameDev.com is a rare website in that it is focussed on talking about the methods used and challenges faced in the commercial games industry. Ran by Alex Champandard, the site often interviews developers working in industry about the AI methods adopted for commercial titles. There is a great range of interviews and other resources available on the site. Though please be aware there is a pay-wall for a some of this material.
AIGameResearch.org is more or less the AIGameDev of research circles. The editorial team is a number of academics (including myself), who aim to highlight great AI applications for games that are being developed within academia. Some of these are just technical demos, others are playable games in their own right. In addition, games added to the site often carry links to the research publications that give more detail on the work behind the scenes.
Shall We Begin?
Now that the introduction and whatnot is out of the way, let’s move onto Part 1, where we begin to cover some basic terminology and explain the principles behind how Artificial Intelligence works in its simplest form.