Computer Poker Program Challenges Humans

A match between human and computer poker players will test how far artificial intelligence has come at interacting with humans in complex ways.

While programmers have recently mastered the game of checkers by creating a program that can win at every game, the much more complex game of poker is just starting to be unraveled by artificial intelligence. A team of researchers at the University of Alberta in Edmonton, Canada have developed a poker-playing program known as Polaris, which is set to square off with two human opponents on Monday.

Phil Laak and Ali Eslami are both professional poker players who have proved their mettle by raking in winnings from the poker circuit in the past. They will play four games of Texas Hold ‘Em with Polaris, winning $5,000 for each successful round, with an additional $5,000 awarded if they manage to sweep all four games.

A successful poker-playing computer would represent a major milepost for artificial intelligence, since the game depends on emotion and predicting the maneuvers of opponents – rather than sheer mathematical prowess – to win. The showdown will be hosted by a Vancouver hotel as part of the Association for the Advancement of Artificial Intelligence conference. More than 1,000 science-savvy spectators have already registered to attend.