See you again next year!
This years hackathon is now over. The results are in for our inaugural Daisy Intelligence 2017 Hackathon held on January 28-29.
We challenged students from the University of Toronto to solve related parallel computing problems in the field of artificial intelligence and received 23 submissions with 34 teams registered for the event. Read on about the challenge, winners, and their strategies. Movies, photos and other fun stuff will be uploaded on our Facebook page.
Make sure you stay tuned for more information about next year’s event!
Develop an ultimate tic-tac-toe playing bot (program) that was executed on GPU hardware and use a maximum of 5 seconds of analysis time to decide its move. Prior to the head to head double elimination tournament, the students were given 24 hours to develop and optionally pre-train their bots.
Why ultimate tic-tac-toe?
The task represented a miniature version of the A.I. problem that Daisy solves for its corporate clients: “in a defined environment make weekly/daily decisions that our clients will execute that will maximize profitability”. In the hackathon context, this means: “for ultimate tic-tac-toe within 5 seconds decide the moves to make that will lead to a win”.
Both the real world and the ultimate tic-tac-toe environments have practically infinite possibilities, therefore it is not possible to determine an optimal strategy.
The head to head tournament started out with early round matches being a best of 3 games with points penalties for invalid moves or moves that exceeded the 5-second decision limit. If a team made too many invalid or time exceeding moves in a single game, that game would be given to the opponent. Later round games were a best of 8 with a draw being broken by number of squares won.
Scripted logic that first places a priority on winning the centre board, then on the corner squares of the boards, and otherwise uses a tree search with heuristic to find the best move it can. Uses a bigger tree if more move time is available.
Hardcoded logic that uses a scoring system to find the best available move. Has a high priority for squares 1/3/5/7 (non-corner side, non-middle squares).
Uses a deterministic tree search with a heuristic to find the highest scoring move in a set number of iterations.
Deterministic logic that chooses a move from the possible moves based on a coded scoring system.
Deterministic logic that uses a simple algorithm (try to win the game, try to win boards, try to block opponent from doing the same).
total students registered.
total submissions received.
tic-tac-toe matches played.
Thanks for hosting this competition, we had fun and learned a lot.Connal de Souza, Team 0.004184
Want to learn more about Daisy Intelligence? Got a question about the Daisy Intelligence 2017 Hackathon? For more information, contact firstname.lastname@example.org.