StarCraft AI Workshop

The IT University of Copenhagen is hosting a two-day StarCraft AI Workshop with sponsored prizes from Google DeepMind.

Computers can beat humans in many games including Chess, Go and most arcade games. The next grand challenge for artificial intelligence (AI) is the real-time strategy game StarCraft. Earlier this year, Google DeepMind and Blizzard released an interface to StarCraft 2 that allows anyone with basic programming skills to build a program that can play the game.

At the workshop, we will explore the capabilities of this interface, how to develop a bot from scratch and how knowledge can be learned from game replays or reinforcement learning.

NOTICE: The deadline for applying to the workshop has passed. You need an acceptance email to attend the workshop.

-- Schedule for the workshop --
Day 1 - Saturday, January 20:
08.45: Doors open
09:00: Welcome by Niels Justesen & Sebastian Risi
09:15: Invited talk - Mike Preuss
09.45: Invited talk - Jonas Busk
10.15: Short break
10.30: Installing and using SC2LE by Niels Justesen
11.00: Loading and processing replays by Niels Justesen
11.30: Mini-games and reinforcement learning by Niels Justesen
12.00: Lunch sponsored by IT Minds
13.00: Group creation and group work
~20.00: End of day 1

Day 2 - Sunday, January 21:
09.00: Doors open - Group work
15.00: Presentations, competitions, and prizes
17.00: End of day 2

You must bring your own laptop. We will most likely do the demonstrations on a Linux machine but the StarCraft 2 Learning Environment (SC2LE) runs on MacOS and Windows as well.

-- Invited talks --
Mike Preuss, Research Associate at University of Muenster: "Lessons learned: StarCraft I in competitions and research since 2010"

"For several years, StarCraft I has been the most prominent test case for research in realtime strategy games (RTS). We report about the early developments and how bots evolved during the first years, including the problems still largely conceived as unsolved. We also provide some insight to strategy selection and build order optimization attempts, before developing some ideas for future directions in RTS research."

Jonas Busk, Ph.D. Student at the Technical University of Denmark: "Deep Learning Primer"