AI learns how to play StarCraft 2 (Video)
Modern computer games often use artificial intelligence (AI) algorithms to control certain elements of the game. This is particularly relevant in genres where decision-making is made over and over again in a single session, such as when playing strategy games.
It is very difficult to match a capable machine learning algorithm. But what if we created another algorithm to compete with the original algorithm created by the game developers?
Bots – or automatic player agents – are nothing new and they have existed since the dawn of video games. But in the past, they were created using a hard-coded logic. What’s the difference when their functionality is enhanced using AI?
The following video shows an experiment in which a machine learning bot was built with the aim of defeating StarCraft 2. This is not an easy task, of course, but a very interesting experiment and well documented. presented here may be useful to those who want to learn the principles of Artificial Intelligence in a fun and intuitive way: