Feb. 3, 2017
Last weekend saw dozens of student teams huddled over computers in the Bahen Centre. They were competing in the first ever Daisy Intelligence Hackathon, a 24-hr contest in parallel computing for artificial intelligence and machine learning.
Students were asked to develop an ultimate tic-tac-toe playing bot (program) that executed on GPU hardware and had only 5 seconds to decide each move. A total of 34 teams took on the challenge, working day and night to develop and pre-train their bots before a head-to-head double elimination tournament.
Ultimate tic-tac-toe is complex enough that there are practically infinite possibilities, making it impossible to determine an optimal strategy. The task represented a miniature version of the A.I. problem that Daisy Intelligence solves for its corporate clients – making fast, profitable decisions in a complex environment. “This hackathon is a great example of how even a rudimentary parallel A.I. method can defeat a human developed rules-based strategy,” said Gary Saarenvirta (8T8 Aerospace), Founder and CEO of Daisy Intelligence. “In the business world, A.I. and parallel computing can develop strategies and tactics better than the best human brains.”
The winning team of Connal de Souza, Jerry Zhou, and Kathryn Yu (all Year 3 ECE) used a tree search and were undefeated in the tournament. They received cash prizes and a chance at an internship with Daisy Intelligence.
“We were positively surprised at what the students were able to accomplish in the allotted time. We didn’t expect every team to come up with a GPU enabled parallel A.I. strategy but a few came close,” said Saarenvirta.