At Thursday 10 Aug 13:00-13:30 Julien Perolat, research scientist at Google DeepMind will be opening the World Championship with a speech about their research project relating to Stratego and Artificial Intelligence. Perolat is a lead author of the article in Science in December 2022 about “Mastering the game of Stratego with model-free multiagent reinforcement learning“.
Google DeepMind has been working for a long time on an agent that can play Stratego. Due to the imperfect information in Stratego, it is quite difficult to model a working AI player.
DeepMind worked together with the Stratego community to build their agent, DeepNash. Vincent de Boer, considered an expert in the game, is listed as a co-author of the article.
During the 4 days of the World Championship Google DeepMind will offer some time slots for participants of the World Championship to play against the DeepNash agent.
These are aligned with the breaks and are currently set as:
Thursday 10 Aug 15:00-16:00
Friday 11 Aug 14:00-15:00
Saturday 12 Aug 13:45-14:45
Sunday 13 Aug 12:00-13:00
Special rules apply during this event.
Location: Tolhuistuin, IJpromenade 2, 1031 KT Amsterdam, The Netherlands
(6 minutes by Ferry from Amsterdam Central Station)