Decision Making AI for Games
Research into Artificial Intelligence (AI) for non-player characters in games has made significant leaps over the last few years, emphasised by Google DeepMind’s AlphaGo and University of Alberta’s AI poker player beating world champions.
However, the games industry is often reluctant to adopt this technology as it is risky to deploy, requires a deep understanding of recent advances in machine learning to implement and while the expert AI players it produces can be challenging, they may not necessarily be fun.
This project is focused on working with game developers to understand how new AI and machine learning technologies can be introduced into their games and help them to adapt their design processes to incorporate these new approaches.
Our ultimate aim is to deliver AI middleware and, where needed, the associated training or support to make these techniques accessible to any game developer, resulting in games which are both more challenging and, most importantly, more fun. In parallel, as games companies get to grips with this new technology, we expect new game genres to emerge which make use of this new form of AI to develop new game-play mechanics.
Machine learning has already begun to have a significant disruptive effect on a wide range of industries, and will similarly transform the games industry in the near future - by working with DC Labs our games industry partners have the opportunity to become industry leaders in delivering these innovative new experiences to their players.
DC Labs researchers are at the forefront of AI and machine learning research and its applications for games, specialising in the specific areas of Monte Carlo Tree Search and Deep Reinforcement Learning.
Through collaboration with AI Factory the team has developed and integrated improved tools for their card game Spades, delivering high quality AI which took it to #1 on the Google Play store, and has now been downloaded over 6.5 million times.
Later analysis of player data captured through the game has allowed DC Labs to further develop the AI, making it more fun and human-like to play against.