Beating your friends at board games using Python and machine learning

by Daniel Kats

Machine Learning & Data Science

Do you hate losing at board games? Do you have too many friends and too few enemies? Then this talk is for you! I will demonstrate rapid prototyping of board games in Python, training simple yet powerful game AI, and extracting their strategies in a human-understandable way to ultimately defeat all your board game foes. This talk will cover helpful newly-introduced Python features for rapid prototyping, using keras for games, and Monte Carlo Tree Search. You will see the effectiveness of these techniques on popular games such as Settlers of Catan and Machi Koro.

About the Author

I’m a rock climber, board game geek, and obsessive coffee drinker. I’m also a principal researcher at Symantec Research Labs. I attended the University of Toronto, interspersed with internships at Yelp and IBM. I graduated with a B.Sc., followed by an M.Sc. studying under Prof. Eyal de Lara. In my free time I read The Atlantic, go on anger-fueled morning jogs, and show up to events 5 minutes after doors close.

Talk Details

Date: Sunday Nov. 17

Location: Concert Hall

Begin time: 12:45

Duration: 30 minutes