​Carnegie Mellon artificial intelligence victorious in heads-up poker tournament

Libratus, an AI system from Carnegie Mellon University, has beat out four poker professionals in a 20-day heads-up Texas Hold'em tournament.

An artificial intelligence (AI) system developed by Carnegie Mellon University (CMU) has defeated four professional card sharks in a 20-day heads-up poker tournament in Pittsburgh.

The AI, Libratus, possesses the ability to perform strategic reasoning and the compute power to process the 10^160 possible information sets a game of heads-up no-limit Texas Hold'em poker has.

Libratus was developed by professor of computer science Tuomas Sandholm and computer science PhD student Noam Brown, and hosted on the Pittsburgh Supercomputing Center's Bridges computer.

According to both Sandholm and Brown, Libratus' victory was not the result of luck.

"The best AI's ability to do strategic reasoning with imperfect information has now surpassed that of the best humans," Sandholm said.

In order to beat the four professionals, Dong Kim, Jimmy Chou, Daniel McAulay, and Jason Les, Libratus had to make decisions without knowing all of the cards in play, while also trying to catch its opponent bluffing.

Throughout the competition, CMU said Libratus recruited the raw power of approximately 600 of Bridges' 846 compute nodes. Bridges' total speed is 1.35 petaflops -- approximately 7,250 times as fast as a high-end laptop -- with its memory coming in at 274 terabytes.

"After play ended each day, a meta-algorithm analysed what holes the pros had identified and exploited in Libratus' strategy," Sandholm said.

"It then prioritised the holes and algorithmically patched the top three using the supercomputer each night -- this is very different than how learning has been used in the past in poker.

"Typically researchers develop algorithms that try to exploit the opponent's weaknesses. In contrast, here the daily improvement is about algorithmically fixing holes in our own strategy."

According to Frank Pfenning, head of the Computer Science Department in CMU's School of Computer Science, Libratus has applications in any area in which information is incomplete and opponents sow misinformation, pointing to business negotiation, military strategy, cybersecurity, and medical treatment planning as areas that could all benefit from automated decision-making.

"The computer can't win at poker if it can't bluff," Pfenning said. "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.

"It has been very exciting to watch the progress of poker-playing programs that have finally surpassed the best human players. Each one of these accomplishments represents a major milestone in our understanding of intelligence."

CMU was previously instrumental in the development of Watson, the AI from IBM that beat out human competitors in Jeopardy! and also IBM's Deep Blue that beat world chess master Garry Kasparov in 1997.

With AI development a key priority for the university, CMU announced in November that it was receiving a $10 million endowment from global law firm K&L Gates LLP to establish a new research centre focused on the ethics of AI.

At the time, CMU president Subra Suresh said that it is not just technology that will shape the next century.

"Our future will also be influenced strongly by how humans interact with technology, how we foresee and respond to the unintended consequences of our work, and how we ensure that technology is used to benefit humanity, individually and as a society," he said.

He also added that Carnegie Mellon is "uniquely positioned" to examine such questions given its history and current leadership in areas including AI, brain science, cybersecurity, and robotics.

Similarly, researchers behind a poker-playing AI system called DeepStack claimed earlier this month it was the first algorithm to have ever beaten poker professionals in a game of heads-up no-limit Texas Hold 'em.

And in March last year, Google subsidiary DeepMind witnessed its AI AlphaGo beat grandmaster Lee Se-dol over five Go matches between man and machine.

Newsletters

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All