According to computer scientists at Carnegie Mellon University, artificial intelligence is very close to yet another milestone. Although the CMU team’s Claudico poker program technically lost this weekend’s big showdown with four human poker professionals, the results were close enough to statistically call it a tie (three out of four humans ended up with more cash than Claudico after 80,000 hands).
Despite coming in fourth place out of five players, the developers of Claudico were ecstatic with the results of the big humans versus artificial intelligence poker showdown. They say it shows their poker software can compete with the best human poker players, and they are very confident they can fine-tune the poker program to where it can beat even the best human beings.
Statement from poker program developers
“We knew Claudico was the strongest computer poker program in the world, but we had no idea before this competition how it would fare against four Top 10 poker players,” commented Tuomas Sandholm, a CMU professor of computer science who led the development of Claudico. “It would have been no shame for Claudico to lose to a set of such talented pros, so even pulling off a statistical tie with them is a tremendous achievement.”
“Beating humans isn’t really our goal; it’s just a milestone along the way,” Sandholm continued. “What we want to do is create an artificial intelligence that can help humans negotiate or make decisions in situations where they can’t know all of the facts.”
More on humans versus AI poker showdown
Despite the fact that three of the four professional poker players won more than Claudico, their $732,713 collective lead over the AI poker program was not large enough to reach statistical significance, which means the results are not scientifically reliable. Of note, $170 million was “bet” during the two-week AI versus humans poker showdown.
The final results were Bjorn Li with an individual chip total of $529,033, Doug Polk with $213,671 and Dong Kim with $70,497. Jason Les lost to Claudico by $80,482. Keep in mind that all of the human players are ranked among the world’s top 10 professionals in heads-up No-limit Texas Hold’em.
The showdown involved Claudico playing 20,000 hands with each human. No real money was bet during the poker showdown, though the professional players will receive appearance fees from a prize pool of $100,000 contributed by Rivers Casino and Microsoft Research.
Poker is a greater challenge for artificial intelligence because it is an incomplete information game. Incomplete information in the sense that players don’t know what cards other players have, and all players try to mislead their opponents.
Software researchers point out that Claudico’s strategy was based on algorithms not rules based human poker expertise. The algorithms ran on the Pittsburgh Supercomputing Center’s Blacklight computer with no input except rules of poker. Experts note the same sort of algorithms could also be used to create strategies for applications involving cybersecurity, business transactions and medicine. An AI might, for example, help develop sequential treatment plans for a patient, or design antibiotics that are less prone to resistance. An AI might even be able to help people negotiate when purchasing a house or a car or making an investment.