A team of computing scientists who are from the University of Alberta’s Computer Poker Research Group is once again capturing the world’s collective fascination with Artificial Intelligence (AI).
Over the historic result for the case of flourishing AI research community, team has developed AI system which is called DeepStack.
Teams of researchers are including researchers from the Charles University in Prague and Czech Technical University. This newer AI has defeated professional poker players in December 2016.
Landmark findings have just been published in Science.
DeepStack: Bridging the Gap
DeepStack is bridging gap between approaches which are used for games of perfect information. It include those which are used in checkers, chess and Go. It is with those which were used for imperfect information games.
It was with reasoning while it plays using “intuition” which is honed through deep learning for reassessing its strategy with each of the decision.
Micheal Bowling who is a professor in the University of Alberta’s Faculty of Science and principal investigator of the study says, “Poker has been a longstanding challenge problem in artificial intelligence. It is the quintessential game of imperfect information in the sense that the players don’t have the same information or share the same perspective while they’re playing.”
Don’t let the name fool you, imperfect information games are quite serious business.
These games are the general mathematical model which describes how decision makers are interacting.
Artificial intelligence research has a storied history of using parlor games for studying these models.
Yet the attention has been focused primarily over perfect information games.
Bowling explaining developing techniques for solving information games will possess applications which are far beyond the poker table.
Applications of AI:
He adds, “We need new AI techniques that can handle cases where decision-makers have different perspectives. Think of any real world problem. We all have a slightly different perspective of what’s going on, much like each player only knowing their own cards in a game of poker.”
It immediate applications will include making a robust medical treatment recommendations, strategic defense planning and negotiation.
This latest discovery is building upon already an impressive body of research findings regarding artificial intelligence and imperfect information games.
This will further stretch back to the creation of the University of Alberta’s Computer Poker Research Group in 1996.
Bowling who has became the group’s principal investigator in 2006, has led the group to several milestones for artificial intelligence.
He and his colleagues has developed Polaris in 2008 while beating top poker players at heads-up limit Texas hold’em poker.
They even went over solving heads-up limit hold’em with Cepheus which was published in 2015 in Science.
DeepStack is extending the ability to think about each of the situation during play. It has been famously successful in games which include checkers, chess and Go.
It is to imperfect information games while using the technique which is called continual re-solving.
This allows DeepStack for determination of correct strategy for a particular poker situation without ever thinking about the entire game by using its “intuition”.
It was to evaluate how the game might play out in the near future.
While thinking about each of the situation as they arise are important for complex problems which includes heads-up no-limit hold’em.
It is having vastly more unique situation than there are atoms in the universe.
It is largely due to the ability of players ability for wagering amounts which are including the dramatic “all-in”.
Despite the complexity of game, DeepStack is taking action at human speed. It is having average of only three seconds of “thinking” time.
It is running over a simple gaming laptop with an Nvidia graphics processing unit.
For testing the approach, DeepStack has played against a pool of professional poker players in December 2016.
It was recruited by the International Federation of Poker. Thirty three players from 17 countries were recruited.
Each of them has asked to play a 3,000 hand match over a period of four weeks. DeepStack beat each of the 11 players who finished their match.
It is only one outside the margin of statistical significance while making it the first computer program for beating professional players in heads up no-limit texas hold’em poker.