Another Win for AI Over Humans in Poker

Back in January [3], a team of four poker pros lost decisively to poker playing GAI software named Libratus. That program was developed by Prof. Toumas Sandholm and Noam Brown from Carnegie Mellon University, who put an improved version of Libratus to the test against six Chinese players this time.

Libratus' "big brother" is called Lengpudashi, which translates to "cold poker master", and it lived up to his name: it beat a team of humans by a wider margin than Libratus, 793.327 chips to be exact over the course of 36.000 hands. The means that the Carnagie U developers got take home the $290,000 prize.

The human side was lead by Chinese venture capitalist and 2016 WSOP bracelet winner Yue Du, but the team also featured engineers who tried to use their understanding of AI technology during the game. 

“I am very excited to take this new kind of AI technology to China, I want to explore various commercial opportunities for this in poker and a host of other application areas, ranging from recreational games to business strategy to strategic pricing to cybersecurity and medicine."  - said the co-creator of Libratus and Lengpundashi, Prof. Sandholm before the match-up.

Also prior to the game, he pointed out something that takes a little away from the machine's victory:

[4]

“This is an exhibition, not a match, challenge or competition,” he added. “We are running a relatively small number of hands, so this is not a scientific experiment like the Brains Vs. AI competition in January.” 

Either way, heads-up No limit Texas Hold'em seems to be on its way to be solved completely by computers.