AI goes Star Craft

Google und Blizzard haben gestern auf der BlizzCon eine offizielle Coop angekündigt, die StarCraft (also wohl den Code und damit die Game Mechaniken) für AI-Wissenschaftler und speziell Machine Learning zugänglich zu machen. Zerg Rush!

Aus GoogleDeepMinds Blogpost dazu: DeepMind and Blizzard to release StarCraft II as an AI research environment.

StarCraft is an interesting testing environment for current AI research because it provides a useful bridge to the messiness of the real-world. The skills required for an agent to progress through the environment and play StarCraft well could ultimately transfer to real-world tasks. […] StarCraft’s high-dimensional action space is quite different from those previously investigated in reinforcement learning research; to execute something as simple as “expand your base to some location”, one must coordinate mouse clicks, camera, and available resources. This makes actions and planning hierarchical, which is a challenging aspect of Reinforcement Learning.

Vorgestern erst hatten Wissenschaftler in einem Paper TorchCraft vorgestellt, „a Library for Machine Learning Research on Real-Time Strategy Games“: AI boffins turn to StarCraft to train future neural networks TorchCraft to build machine-learning agents from cosmic battles.