Starcraft2 is a complex game where two (or more) players compete to collect resources, build armies, and overcome the opponent to win the game. One of the major complexities in Starcraft compared to other strategy games is the real-time aspect; units and building must be accurately controlled live on top of executing high-level strategies. Therefore, one promising direction is to utilize deep learning. Neural nets can encode previous experience in their weights and perform fast reactive queries.
In this project we’ll focus on base-building \ army-raising aspect (also called “macroing”). We’ll implement an algorithm that learns both from self-play and from a demonstrator. The project will require learning about reinforcement learning (RL), search methods, and probabilistic models.