TL;DR This project will make use of the Nervana DevCloud to build a multinode reinforcement and / or supervised learner for StarCraft II using Intel optimized deep learning and machine learning libraries.
The next big thing after AlphaGo (Zero) is an artificial intelligence (AI) algorithm that can beat humans in the real-time strategy game, StarCraft II. Recently, Google's DeepMind and the video game developer company, Blizzard, have open sourced both the API of StarCraft II https://github.com/deepmind/pysc2 and a rich data repository of past human plays for this game. These developments have generated a lot of excitement in the AI community as the game of StarCraft II has higher complexity than the game of go. Furthermore, the lessons learned from building an AI for this StarCraft II can be applied to search-and-rescue missions among other real-world uses.
This project is started by the Portland Reinforcement Learning Group. https://www.meetup.com/meetup-group-PmeikSbV