When AlphaGo, the AI system built by Google’s DeepMind, beat one of the world’s best Go player, Lee Sedol, last year, it marked yet another classic game where computers now outclass humans. But machines have often struggled when it comes to besting professional poker players - until now.
Libratus, an artificial intelligence program developed at Carnegie Mellon University, has just beaten four of the world’s best poker players - Dong Kim, Jimmy Chou, Daniel McAulay and Jason Les – in a 20-day Heads-up, No-Limit Texas Hold’em tournament.
The event was held at Rivers Casino in Pittsburgh, and saw Libratus up by $1,776,250 after over 120,000 hands had been played. Games went on for 11 hours a day during the entire twenty days.
Created by professor of computer science Tuomas Sandholm and computer science PhD student Noam Brown, Libratus’ “historic” victory has been hailed as a milestone for artificial intelligence systems.
"Heads-up no limit Texas hold 'em is in a way the last frontier of all the games," said Sandhome, "Othello, Chess, Go, Jeopardy have all been conquered, but this remained elusive: this is a landmark in AI gameplay."
Figuring out human behavior when it comes to bluffing and other poker plays has proved to be a problem for machines in the past, but even with the four pros conferring on how to beat Libratus, the AI, using the Pittsburgh Supercomputing Center's 1.35 petaflop Bridges computer, dominated throughout.
The University said Libratus used around 600 of Bridges’ 846 compute nodes. "After play ended each day, a meta-algorithm analyzed what holes the pros had identified and exploited in Libratus' strategy," said Sandholm.
“It then prioritized 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,” added the professor.
CMU just made history: AI beats top humans at Texas Hold'em poker. A stunning accomplishment, comparable to Deep Blue & AlphaGo!— Andrew Ng (@AndrewYNg) January 31, 2017
More details on Libratus’ workings will be revealed once researchers analyze the results and publish the findings in a paper.
Beating Poker pros is just a first step in what could lead other applications for the AI. "The computer can't win at poker if it can't bluff," said Frank Pfenning, head of the Computer Science Department in CMU's School of Computer Science. "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.”
The four players involved in the tournament may have lost out to Libratus, but they still shared a $200,000 prize fund, ranked in order of how well they played.