Facebook and Carnegie Mellon University have built their version of artificial intelligence bots that managed to beat some of the top poker players in one-on-one competition. According to Facebook, this is the first time a bot has been able to beat top pros in any major benchmark game.
As per the study published in Science Magazine, artificial intelligence-based system named Pluribus bested professional players in no-limit Texas Hold’em in a number of formats including five AI bots and one human, and one bot and five real-life players.
The team behind the development of Pluribus claimed that a multiplayer poker bot is undoubtedly a recognized AI milestone (swatch the below featured video for a demo).
Poker is a challenging game whereas other games like Chess and Go have everything laid out in the open. But in poker, there is lots of hidden information that calls for the players to apply different and complex strategies such as bluffing.
In fact, bluffing is a strategy that can dramatically change the game – either in your favor or in opposition. And having a bot balance the art of bluffing is definitely an achievement for the AI community.
A bot developed by Carnegie Mellon also beat pros in heads up play a couple of years ago. However, Pluribus is a more advanced version of the former and equipped with better technologies that enable it to learn more quickly and deal with hidden information more efficiently.
In fact, Facebook claims that it uses less than 128 GB of memory and runs on just two GPUs while playing. It is also twice as fast as the pro players are and take an average of 20 seconds per hand when it played copies of itself.
During the 10,000 hands of poker in over 12 days, Pluribus played against several experts including World Series of Poker Maini event champions and World Poker Tour winners. These include Chris Ferguson, Greg Merson, Darren Elias, and Jimmy Chou. They have all won at least $1 million during professional games previously.
Even the pros were intrigued by the strategies applied by Pluribus and found the whole experience fascinating. Chris Ferguson even claimed that Pluribus is a challenging opponent and it was quite difficult to pin him down on any kind of hand.
"Thanks to Tuomas Sandholm and the team at CMU who have been working on strategic reasoning technologies over the last 16 years. Sandholm has founded two companies in this work — Strategic Machine Inc., Strategy Robot Inc. — that have exclusively licensed the technologies developed in his Carnegie Mellon laboratory. Strategic Machine is applying the technologies to poker, gaming, business, and medicine, and Strategy Robot is applying them to defense and intelligence.", explained Facebook AI Research Scientist, Noam Brown, in a blog post. Adding further, "Pluribus builds on and incorporates large parts of that technology and code. It also includes poker-specific code, written as a collaboration between Carnegie Mellon and Facebook for the current study, that will not be applied to defense applications."
Read next: Tencent, Sony, Microsoft, Apple, Google: Which Companies Earn the Most Money From Games?
As per the study published in Science Magazine, artificial intelligence-based system named Pluribus bested professional players in no-limit Texas Hold’em in a number of formats including five AI bots and one human, and one bot and five real-life players.
The team behind the development of Pluribus claimed that a multiplayer poker bot is undoubtedly a recognized AI milestone (swatch the below featured video for a demo).
Poker is a challenging game whereas other games like Chess and Go have everything laid out in the open. But in poker, there is lots of hidden information that calls for the players to apply different and complex strategies such as bluffing.
In fact, bluffing is a strategy that can dramatically change the game – either in your favor or in opposition. And having a bot balance the art of bluffing is definitely an achievement for the AI community.
A bot developed by Carnegie Mellon also beat pros in heads up play a couple of years ago. However, Pluribus is a more advanced version of the former and equipped with better technologies that enable it to learn more quickly and deal with hidden information more efficiently.
In fact, Facebook claims that it uses less than 128 GB of memory and runs on just two GPUs while playing. It is also twice as fast as the pro players are and take an average of 20 seconds per hand when it played copies of itself.
During the 10,000 hands of poker in over 12 days, Pluribus played against several experts including World Series of Poker Maini event champions and World Poker Tour winners. These include Chris Ferguson, Greg Merson, Darren Elias, and Jimmy Chou. They have all won at least $1 million during professional games previously.
Even the pros were intrigued by the strategies applied by Pluribus and found the whole experience fascinating. Chris Ferguson even claimed that Pluribus is a challenging opponent and it was quite difficult to pin him down on any kind of hand.
"Thanks to Tuomas Sandholm and the team at CMU who have been working on strategic reasoning technologies over the last 16 years. Sandholm has founded two companies in this work — Strategic Machine Inc., Strategy Robot Inc. — that have exclusively licensed the technologies developed in his Carnegie Mellon laboratory. Strategic Machine is applying the technologies to poker, gaming, business, and medicine, and Strategy Robot is applying them to defense and intelligence.", explained Facebook AI Research Scientist, Noam Brown, in a blog post. Adding further, "Pluribus builds on and incorporates large parts of that technology and code. It also includes poker-specific code, written as a collaboration between Carnegie Mellon and Facebook for the current study, that will not be applied to defense applications."
Read next: Tencent, Sony, Microsoft, Apple, Google: Which Companies Earn the Most Money From Games?
