Artificial intelligence has another victory notch on its belt. After a marathon 12-day poker session, an artificial intelligence program called Pluribus emerged victorious.
The bot defeated leading professionals in games of no-limit Texas hold'em poker, the most popular form of the game.
During one experiment, Pluribus played five professionals at a time for a total of 10,000 hands and won. Each of these players had previously racked up more than $1 million in prize money.
In another experiment, five copies of Pluribus took on two top professionals, playing around 5,000 hands each time. The AI program defeated them both.
Two years ago a forerunner of Pluribus called Libratus also beat human players, but this program only played one-on-one, not several players at once.
“Pluribus achieved superhuman performance at multi-player poker, which is a recognized milestone in artificial intelligence and in game theory that has been open for decades,” said Professor Tuomas Sandholm, who developed Pluribus with his PhD student Noam Brown at Carnegie Mellon University in Pittsburgh.
“Thus far, superhuman AI milestones in strategic reasoning have been limited to two-party competition.
“The ability to beat five other players in such a complicated game opens up new opportunities to use AI to solve a wide variety of real-world problems.”
Among the real-world problems that Sandholm thinks Pluribus could be applied to are investment banking, deciding how much money political candidates should spend in advertising with various media outlets and negotiation strategies.
The AI poker achievement is described in a research article 'Superhuman AI for Multiplayer Poker' in the current edition of the journal Science.