Strategy games such as chess have long been considered important ways to measure artificial intelligence. But A.I. researchers at Carnegie Mellon University chose a different, and in some ways, more challenging game: poker. Susan Koeppen of CBS Pittsburgh station KDKA reports on what happens when the chips are down.
Doug Polk, 26, is considered the best heads up, or one on one, no limit Texas hold 'em player in the world. He's defeated countless opponents and won millions of dollars.
Polk bet his reputation that he could beat Claudico, Carnegie Mellon's artificial intelligence super computer.
"You're playing a cold-blooded killer because when he goes all in and you snap him off and win his stack, he's not scared now, he's just computing, right?" Polk said.
For the past two weeks, Polk and three other professional poker players each played 20,000 hands against Claudico at Rivers Casino in Pittsburgh.
Viewers from more than 100 countries watched online, but nobody paid closer attention than the man responsible for Claudico, professor Tuomas Sandholm.
"The computer definitely bluffs and does all sorts of other tricks that human poker players know, but the key is that we don't program in the bluffing," Sandholm said. "So the algorithms themselves figure out the strategy, how to bluff, when to bluff, in what situations and so forth.*
In 1997, the world watched in wonder when IBM's Deep Blue, whose research originated at Carnegie Mellon, defeated the world's best chess player, Garry Kasparov. And again in 2011 when Watson bested "Jeopardy" champion Ken Jennings.
So why is poker a better gauge of A.I. than playing "Jeopardy" or chess?
"In chess it's a game of complete information, so when it's your turn to move you know exactly what the state of the world is, what the state of the game is," Sandholm said. "In poker, you don't.
"This is really to be able to assist humans and companies in interacting, let's say in negotiation," Sandholm said. "Wouldn't it be nice if you had an agent that helped you strategize in the world when you're buying a car or buying insurance?"
Jason Les studied computer science in college before becoming an online poker pro. It turns out his education wasn't much of a help. But he was still happy he signed up to play a computer.
"I thought this was a historic event and a big landmark in poker and artificial intelligence," he said. "And I'm happy that I came up here and I was able to be a part of the winning team."
Well, not exactly. According to Carnegie Mellon, the pros' combined $732,000 lead in fake money makes it a statistical tie. The university plans to rewrite Claudico's algorithms.