Facebook, CMU’s AI Pluribus Beats Pros in NL Texas Hold’em
Artificial Intelligence (AI) has once again approved that it is capable of matching up to what humans can do or even beating them at some of those things. Pluribus, an AI program designed by Carnegie Mellon University (CMU) and Facebook’s AI lab researchers has recently been able to best some of the world’s top professional poker players in a series of games of 6-person No-Limit Texas Hold’em poker.
While this is not the first time that AI and bots have been able to beat professional poker players in one-on-one competitions, according to Facebook, this is the very first time that an AI system has been able to beat top poker professionals in “any major benchmark game” when there are multiple opponents. As such, this marks a huge milestone for the field of AI research especially with regards to the game of poker. Unlike games like Go and Chess where everything is openly laid out, poker involved a lot of hidden information and complex strategies that have proven to be very challenging for AI systems to account for and act on.
Moreover, there is also the issue of bluffing which is quite an interesting challenge for both players and AI. This technique can dramatically change a poker game in one’s favor but if used too much, it may make the player predictable. Pluribus, which is an advanced version of CMU’s bot Libratus, somehow manages to balance bluffing with betting on legitimately strong hands just like the top poker players do. The AI utilizes a new online search that lets it look at the available options for some moves ahead and not just the end of the game.
“We think of bluffing as this very human trait. But what we see is that bluffing is actually mathematical behavior. When the bot ‘bluffs’, it doesn’t view it as deceptive or dishonest, it’s just the way to make the most money,” explained Noam Brown, the lead researcher from Facebook’s AI team
In addition to that, Pluribus also has “faster self-play algorithms for games with hidden information,” which means that it is more efficient when it comes to dealing with hidden information in games it played against copes of itself.
What Now?
Well, it goes without saying that there have been concerns that the advancement of AI technology has raised a few concerns especially with regards to fraud and unfair play in online poker rooms. However, even though it is still not completely clear why tech companies like Facebook are interested in the building advanced AI algorithms to beat human players, researchers have confirmed that the intention is not to ruin the games themselves. Instead, the goal is to broaden the applications of Artificial Intelligence in the real world.