The matches between AI and pro players have been ensuing this year in Pittsburgh and this time it seems that the machines have an advantage against the humans.
Two robot players named Libratus and DeepStack are currently engaged in daily games of Texas Hold’em and other variants against pro human players. DeepStack is an AI program that has beaten the professionals in a no limit Texas hold’em game. This was proclaimed by researchers who released the results of the game in a paper few days back.
The poker games prove to be challenging for AI which will help to train them further. It is considered a great achievement whenever AI is able to beat human players in a game, especially when the game is a complicated one that involves strategies and bluffing techniques.
Last year there was a similar round of games held between AI and humans and the latter were able to win, though with a small margin. DeepStack was able to win with a rate of 450 million big blinds per game which was average for every win. The wins were calculated over 44000 hands of the game that were played.
This is certainly a trend that is being seen in all research areas involving development of AI. For instance, Mark Zuckerberg had announced that his team had been working on AI and testing it against human players in a game of Go. This was a successful stance and an AI team had won against the human players. Poker, on the other hand, has always proved to be more of a challenge due to complex strategies that are involved in many variants of the game. This particular game event is being used in AI research by Carnegie Mellon University to test their AI development projects. It seems that this time their advances are in the right direction.