EDMONTON 含羞草研究社� A computer program has learned to win at one of the most complex poker games by copying a very human impulse 含羞草研究社� trusting its gut.
含羞草研究社淚 think there含羞草研究社檚 a lot of similarities to real human intuition,含羞草研究社� said Michael Bowling of the University of Alberta含羞草研究社檚 Computer Poker Research Group.
Teaching poker to computers has been a popular tool in the artificial intelligence community for years.
Unlike games such as chess, no poker player knows what cards other players hold. Having to deal with incomplete information makes poker programs useful in everything from improving public security to helping doctors treat patients with diabetes.
Bowling含羞草研究社檚 lab has long worked on poker and attracted worldwide attention in 2015 for developing Cepheus, a program that was unbeatable in two-handed, fixed-bet Texas Hold 含羞草研究社楨m. The lab含羞草研究社檚 latest achievement, revealed Thursday in the journal Science, went after the much more complex version of Texas Hold 含羞草研究社楨m in which there is no limit on bets.
Cepheus worked by allowing the computer to learn from mistakes. After billions of hands, it developed a 10-terabyte table of probabilities that made it unbeatable.
But that wasn含羞草研究社檛 going to work for no-limit Hold 含羞草研究社楨m.
The fixed-bet version of the game has decision points equal to the number 10 with 14 zeros after it. The no-limit equivalent is 10 with 160 zeros after it.
含羞草研究社淭hat含羞草研究社檚 more than there are atoms in the universe,含羞草研究社� said Bowling 含羞草研究社� way too many to simply crunch through probabilities.
Bowling含羞草研究社檚 team found the answer with a program called Deep Stack.
含羞草研究社淒eep Stack doesn含羞草研究社檛 compute the whole strategy beforehand,含羞草研究社� Bowling said.
含羞草研究社淚t含羞草研究社檚 going to compute how it含羞草研究社檚 going to play on line, as it含羞草研究社檚 playing. It含羞草研究社檚 going to only worry about the decision points it reaches while it plays, and figure out how to play those on the fly, in the middle of the game, much more like a human player.含羞草研究社�
The key is in developing what Bowling calls intuition. Like a human player, Deep Stack trains its instincts through repetition 含羞草研究社� in this case, 10 million poker hands played against itself.
含羞草研究社淒eep Stack will play against itself over and over again until it figures out, 含羞草研究社業 think this is how much it含羞草研究社檚 worth, being in this poker situation,含羞草研究社櫤卟菅芯可鐫 Bowling said.
That data gets fed into a program that recognizes patterns, which in turn allows Deep Stack to deal with new situations.
含羞草研究社淚f it does a good job, what I should be able to do is feed it a poker situation that含羞草研究社檚 not any of the situations it含羞草研究社檚 seen before, and it should still give me a good answer for how valuable that situation is.
含羞草研究社淚t含羞草研究社檚 going to generalize from all its past experience.含羞草研究社�
Remarkably, Deep Stack doesn含羞草研究社檛 take that much horsepower. While Cepheus required vast amounts of computing ability, Deep Stack can be run on an off-the-shelf gaming laptop.
And it seems to work.
In December, Deep Stack played a tournament of 3,000 hands with 30 poker professionals from around the world. The program beat all 11 professionals who completed all 3,000 hands.
It was the first time a computer beat humans in a two-handed, no-limit game.
含羞草研究社淭here were some players that, even when you showed them that they were losing, they were convinced they were not,含羞草研究社� Bowling recalled. 含羞草研究社漈his is the nature of poker. You can always convince yourself that you just got unlucky.含羞草研究社�
含羞草研究社� Bob Weber, The Canadian Press
