Artificial Intelligence and Deep Learning in Video Games A Brief Review

Abstract

Artificial Intelligence has been used in many fields since the term was coined six decades ago. Artificial Intelligence in video games was introduced with Atari 2600’s early titles such as ‘Computer Space’ and ‘Pong’ and has come a long way since. However, it has become expected of developers to have flawless AI which is as human-like as possible. Neural Networks are allowing AI systems to become smarter and used in ways such as OpenAI does with their game-playing bots. Many game-playing Neural Networks use reinforcement learning to train as it allows for more efficient training for larger games where the number of possible positions and combinations of game mechanics are extremely large. This paper provides an introductory literature review of these core fields of research as they applied within a video game context.