A look at the latest AI poker bot
Poker has been always regarded as a mammoth challenge in the field of Artificial Intelligence (AI). However, in July, Facebook and Carnegie Mellon researchers created an (AI) bot that can beat top human professional poker players in a six-player game of Texas hold’em poker.
A game of poker has a lot of complex permutations and combinations and involves hidden information. In order to be successful, the game calls for tactics and a skill set that does not apply to many other games. Therefore making it harder to have the same breakthrough results in AI techniques when it comes to poker. While researchers did develop an AI that can beat just one other player in a no-limit Texas hold’em game, multiplayer was a bigger hurdle to cross.
Titled as Pluribus, Facebook’s poker-playing AI is the brainchild of Facebook AI research scientist Noam Brown and Carnegie Mellon University professor Tuomas Sandholm. Pluribus mastered multiplayer Texas hold’em by playing against previous versions of itself. This self-play learning method means no prior data has been fed from humans and it has not observed games played by other AI systems either. The AI starts from scratch and gradually improves its learning curve comparing it’s outcomes to earlier versions of its strategy.
For all you poker fans out there thinking Pluribus is not going to catch those human tells; the AI beat top-ranked professional players in a five AIs plus one human player format and a one AI plus five poker pros format. Among the professionals were Chris Ferguson and Darren Elias. The bot won an average of $5 per hand and it was estimated to make $1000 per hour playing humans if each chip was worth a dollar!
Pluribus is a supercharged version of another AI bot called Libratus, which beat human pros in two-player Texas hold’em games in 2017. Unlike Libratus, Pluribus contains a new online search algorithm that can evaluate its options by searching a few moves ahead, as well as faster self-play algorithms. These two factors coming together made it possible to train Pluribus using relatively little processing power and memory.
AI researchers have been using games as a testbed for their AI agents for decades now. Thanks to the advances in computing, better data sets, and sophisticated AI techniques, the field has been seeing a number of breakthroughs! Tech giants are investing heavily into space with the hope that gaming breakthroughs will lead to breakthroughs in other important areas such as healthcare, science, and energy.