Introduction
Google's Artificial Intelligence research group, DeepMind recently released a python API, pySC2 for the popular Real Time Strategy (RTS) computer game, StarCraftII. After successfully conquering the boardgame, Go, with their AlphaGo program, DeepMind has set their sights on the next big challenge for AI systems in attempting to train these systems to learn how to compete with world champions in the increadibly complex world of StarCraft. In this post, we use the pySC2 API to collect gameplay data from replays of human played games with the aim of discoverying some macro elements of the game, such as the different technology progression trees players use. Ultimately, the hope is to build up some intuition on what data is availible from the pySC2 API and how it might be useful in building systems capable of playing the game, but that work will be outside the scope of...
Continue Reading...