Rivers Casino Hosts Brains Versus Artificial Intelligence: Upping the Ante Poker Event
The Rivers Casino in Pittsburgh is readying for the Second Bi-Annual Brains Versus Artificial Intelligence poker tournament next week. The 2017 poker event is called Brains Versus Artificial Intelligence: Upping the Ante.
The poker event is the brainchild of computer scientists from the Carnegie Mellon University in Pittsburgh. Starting on January 11, the Rivers Casino will host a 20-day poker tournament. In all, 120 thousand hands of heads-up no-limit Texas hold’em will be played between four poker professionals and the Libratus computer program, which was designed by the Carnegie Mellon computer scientists.
The Brains: Four Poker Professionals
The four poker pros who have volunteered for the combined tourament/experiment are Jason Less, Daniel McAulay, Jimmy Chou, and Dong Kim. Each player is expected to play 30,000 hands of Texas Hold’em apiece against Libratus. The players are vying for ascending shares of a $200,000 prize purse, which lured them to the event.
The AI: Libratus the Software Program
While the poker players have real world cash on the line, the Carnegie Mellon researchers are trying to test and study their artificial intelligence. They seek to set a new benchmark in the advancement of artificial intelligence. While having Libratus play many hands against skilled players, the scientists are trying to use “deep learning” to make their program smarter. Machines have been created to win at Jeopardy, Chess, and Go; apparently, poker is the next step.
Machine learning is based on data. Thus, playing 120,000 hands of a game allows an artificial intelligence like Libratus to collect data on playing habits. Whether that translates to winning poker is another question altogether.
2015: The Claudico Poker Sessions
In the 2015 Brains Versus Artificial Intelligence poker tournament, Carnegie Mellon University designed an AI program named Claudico to play against four poker professionals. Over 80,000 heads-up hands, three of the four players collected more chips than Claudico.
Though the human players won more than the machine, the Carnegie Mellon University designers said that the results were indeterminate. The poker event simply did not have enough hands to produce certain results, so they have added 50% more poker hands this year.
Sponsors of “Upping the Ante” Event
A variety of sponsors are putting up the money for this year’s Brains Versus Artificial Intelligence poker event. GreatPoint Ventures, a venture capital firm, is one of the chief sponsors. The Pittsburgh Supercomputing Center is involved in the Libratus sessions, through a peer-reviewed XSEDE allocation.
Technology companies like Intel, TNG Technology Consulting, Optimized Markets Incorporated, and Avenue4Analytics are other sponsors. Artificial Intelligence, a publication which began in 1970, is also involved in the promotion.
The Carnegie Mellon computer lab, Sandholm’s Electronic Marketplaces Laboratory, is also involved in the Texas hold’em tournament. Of course, Rivers Casino is providing the facilites for the experiment.
Carnegie Mellon University Information
Carnegie Mellon founded the Carnegie Technical Schools in 1900. In 1912, they became the Carnegie Institute of Technology and began offering four-year degrees. In 1967, the institution merged with the Mellon Institution of Industrial Research (founded 1937) and became known as Carnegie Mellon University.
Update on the Libratus Poker Sessions
The event ended on January 30, 2016 with Libratus declared the winner. Over 120,000 hands of Texas hold’em, Libratus collected $1,766,250 in chips.
Tuomas Sandholm said of the competition, “The best AI’s ability to do strategic reasoning with imperfect information has now surpassed that of the best human.”
Tuomas Sandholm is the man for whom the Sandholm’s Electronic Marketplaces Laboratory is named, and also the co-developer of Libratus. Noam Brown, a Ph.D. student in computer science, also co-developed Libratus.
Frank Pfenning on Advancements in Artificial Intelligence
Frank Pfenning, Carnegie Mellon University’s Computer Science Department head, described the ability which put Libratus over the top: bluffing. Pfenning said of the accomplishment, “The computer can’t win at poker if it can’t bluff. Developing an AI that can do that successfully is a tremendous step forward scientifically and has numerous applications. Imagine that your smartphone will someday be able to negotiate the best price on a new car for you. That’s just the beginning.”
Daniel McAuley about Artificial Intelligence
Jason Les and Daniel McAulay both agreed that Libratus was tougher than they expected. McAulay said the software program’s ability to learn from poker professionals was important. McAulay said of the deep learning he encountered, “Whenever you play a top player, you learn from it.“
