• Monday, May 28, 2012

Previous

Next

A Robot That Knows How to Check-Raise

August 8, 2006, 1:51 pm

In Las Vegas the World Series of Poker’s main no-limit hold `em event is now whittling its way down to a final table. But just a few time zones away, another high-profile poker competition has already crowned a champion.

A computer program designed by researchers at the University of Alberta has taken top honors at the American Association for Artificial Intelligence’s first-ever Computer Poker Competition, held at the group’s annual conference in Boston. It’s hardly surprising that Alberta’s software would emerge victorious: The university has long been a leader in the specialized field of computer poker, and its artificial intelligence has been used in poker video games.

Computers, it should be noted, have a long way to go before they play poker as well as they play chess or Scrabble. The “bots” competing in this year’s contest played heads-up matches of limit hold `em—a game that involves more raw number-crunching, and less “reading,” than no-limit, its more popular cousin. In order to hang with the world’s best no-limit players, computer programs would have to learn to make crafty plays based on incomplete information. That’s a tantalizing challenge for artificial-intelligence experts, but it’s not one that’s likely to be met any time soon.—Brock Read

This entry was posted in Research. Bookmark the permalink.

  • Print
  • Comment

Comments are closed.