ShareThis Page

Rematch of brains vs. poker bots under way at Rivers Casino

Aaron Aupperlee
| Wednesday, Jan. 11, 2017, 5:54 p.m.
Professional poker player Jason Les squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Nate Smallwood | Tribune-Review
Professional poker player Jason Les squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Professional poker player Jason Les squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Nate Smallwood | Tribune-Review
Professional poker player Jason Les squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Carnegie Mellon University’s School of Computer Science Professor Tuomas Sandholm speaks at the beginning of the competition between Libratus, an artificial intelligence(AI) and professional poker players playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Nate Smallwood | Tribune-Review
Carnegie Mellon University’s School of Computer Science Professor Tuomas Sandholm speaks at the beginning of the competition between Libratus, an artificial intelligence(AI) and professional poker players playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Professional poker player Daniel McAulay squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Nate Smallwood | Tribune-Review
Professional poker player Daniel McAulay squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Professional poker player Jason Les speaks with Carnegie Mellon University’s School of Computer Science Professor Tuomas Sandholm while Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Nate Smallwood | Tribune-Review
Professional poker player Jason Les speaks with Carnegie Mellon University’s School of Computer Science Professor Tuomas Sandholm while Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Professional poker player Daniel McAulay squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Nate Smallwood | Tribune-Review
Professional poker player Daniel McAulay squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Professional poker player Jason Les squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.
Nate Smallwood | Tribune-Review
Professional poker player Jason Les squares off against Libratus, an artificial intelligence(AI) program playing Heads-Up, No-Limit Texas Hold’em at Rivers Casino on Jan. 11, 2017. Libratus was developed at Carnegie Mellon University’s School of Computer Science.

Dan McAulay sat with an ace and a king, both hearts.

He started the betting, growing the pot. The computer countered, raising and checking McAulay's bets.

The flop yielded a nine, six and jack. No help. The turn was a two. The pot approached $2,000, and McAulay folded, cutting into his healthy chip lead.

Much more than money is at stake as four of the best Heads-Up No-Limit Texas Hold'em poker players challenge an artificially intelligent poker bot created at Carnegie Mellon University.

For poker players — and humanity at large, in many ways — it's a chance to retain supremacy over machines in what many see as the final challenge. Computers already have bested the best humans in checkers, chess, "Jeopardy" and recently Go, a popular strategy game in Asia.

"This is similar to a new guy coming and trying to pick us off," said Jason Les, one of the four professional poker players challenging Libratus, the poker bot. "Heads-Up No-Limit is important to us. It means something to me to know that a computer cannot defeat us."

For the AI — and those who have worked decades to further artificial intelligence — it's a chance to make a statement that computers are ready to handle some of the most complex human interactions.

"Measuring performance against top humans has really been a goal of AI since the onset," said Tuomas Sandholm, a CMU professor who built Libratus. "In a way, this is the last frontier for the foreseeable future."

Play started Wednesday in a rematch of the Brains vs. Artificial Intelligence poker tournament at Rivers Casino on Pittsburgh's North Shore. The brains won the first match in August 2015, but Sandholm has returned with what he believes is a more powerful and more clever opponent.

Libratus already was impressing Les. The computer played a incredibly complicated pre-flop strategy, the betting scheme after the first two cards are dealt to the players but no cards are flipped over for the table.

"It's not something any human could do with just a human brain," Les said.

The pros — Les, McAulay, Dong Kim and Jimmy Chou — each will play 30,000 hands against Libratus over the course of 20 days for a total of 120,000 hands. That's 1,500 hands per person per day for a share of a $200,000 pot at the end of it.

Les said that's pretty common for a professional player. But playing poker from 11 a.m. to 7 p.m. every day for 20 days and mixing in time to study and debrief with the other pros makes for a long tournament.

"It's a lot of fun, like for the first few days," said Doug Polk, a professional poker player who participated in the August 2015 tournament. "But then it just became such a grind."

Polk watched the early hands of this tournament on the video gaming streaming service Twitch. He said he saw Libratus use many of the same strategies that Claudico, the AI he faced 18 months ago, employed. The game seemed even at first, but Polk was confident the humans would come out on top again.

"They would have had to substantially increase its capability in the last year," Polk said of the transition from Claudico to Libratus. "And how much they would have to increase it by to beat the humans is just not that likely."

Sandholm said Libratus was an underdog on some betting websites. He played against Libratus on Tuesday night.

"I got destroyed," he said.

Libratus, a Latin word meaning balance and powerful, packs serious computing power. Claudico ran off a single computer that Sandholm brought to the casino for the tournament. Libratus runs off the Bridges supercomputer housed at the Pittsburgh Supercomputing Center. The AI uses about half of the supercomputer's capability, about the same computing power as 10,000 computers working in tandem, said Ralph Roskies, scientific director of the Pittsburgh Supercomputing Center.

Andrew Moore, dean of CMU's School of Computer Science, said the "AI world" would be watching the tournament closely, not because of the poker but because of what an AI advanced enough to beat the best at poker could mean. Poker is an example of a game with hidden information. One player purposefully conceals information from another. Players get two cards they keep secret and try to make the best poker hand with five others dealt face up.

As we expect more from artificial intelligence, we will expect it to handle hidden information scenarios in business, health, the military and other situations, Moore said. Imagine a cellphone app designed to buy a car for you. You tell the app you are willing to pay $5,000 for the car, but you don't want the sales person to know that. The app goes in ready to negotiate and concealing that $5,000 is an acceptable price.

That's what an AI that can beat the four best Heads-Up No-Limit Texas Hold'em poker players could do, Moore said.

"I don't feel like if we lose that it's a bad thing. Poker is a benchmark they are using for their research. They're not trying to beat poker," Les said, explaining that Libratus or similar AIs could be used in medicine, finance, security and more. "Either I'm defending humanity, or I lose to something that is going to help humanity."

A win by the computer could send shock waves through the online poker community. Polk said if Libratus wins, fewer people might play online poker because they would know it's possible to face a computer they can't beat. Les said it's unlikely an AI as sophisticated at Libratus would end up in an online game, but he said poker bots are already a problem in online poker.

"A lot of the pros are going back to live poker because they don't know who they are playing with online. They could be playing against the computer," said Craig Clark, Rivers Casino general manager.

Clark, who is rooting for the humans, said a win by Libratus could persuade online players to take a seat in Rivers' poker room.

And while the pros stared at computer screens, tapping keys or clicking the mouse, real, live humans sat around tables in the poker room to face off against each other. Dave Schneider, 28, of North Huntingdon, wandered from his table over to where Les and McAulay played.

"It's wild," Schneider said. "I'm rooting for the humans, which is kind of a scary thought."

TribLIVE commenting policy

You are solely responsible for your comments and by using TribLive.com you agree to our Terms of Service.

We moderate comments. Our goal is to provide substantive commentary for a general readership. By screening submissions, we provide a space where readers can share intelligent and informed commentary that enhances the quality of our news and information.

While most comments will be posted if they are on-topic and not abusive, moderating decisions are subjective. We will make them as carefully and consistently as we can. Because of the volume of reader comments, we cannot review individual moderation decisions with readers.

We value thoughtful comments representing a range of views that make their point quickly and politely. We make an effort to protect discussions from repeated comments either by the same reader or different readers

We follow the same standards for taste as the daily newspaper. A few things we won't tolerate: personal attacks, obscenity, vulgarity, profanity (including expletives and letters followed by dashes), commercial promotion, impersonations, incoherence, proselytizing and SHOUTING. Don't include URLs to Web sites.

We do not edit comments. They are either approved or deleted. We reserve the right to edit a comment that is quoted or excerpted in an article. In this case, we may fix spelling and punctuation.

We welcome strong opinions and criticism of our work, but we don't want comments to become bogged down with discussions of our policies and we will moderate accordingly.

We appreciate it when readers and people quoted in articles or blog posts point out errors of fact or emphasis and will investigate all assertions. But these suggestions should be sent via e-mail. To avoid distracting other readers, we won't publish comments that suggest a correction. Instead, corrections will be made in a blog post or in an article.