AI Still Unable To Beat The Game of Poker

Published Thursday, January 28, 2016 -
AI Still Unable To Beat The Game of Poker

Artificial intelligence is an issue for online poker players who worry they are being taken advantage of by deceptive practices. A recent article in Discover magazine has revealed that artificial intelligence still  has not mastered the game of poker. The article explains that although the computer gurus have developed algorithms that have mastered the ancient 2,500 years old Chinese game of Go. Go is generally considered the most difficult strategy game to master. For artificial intelligence researchers, building an algorithm that could conquer Go’s world champion is the holy grail of artificial intelligence goals.

Researchers David Silver and Demis Hassabis involved in the Google DeepMind project designed an algorithm, called AlphaGo, which in October 2015 defeated Europe’s Go champion Fan Hui five games to zero. Although the researcher Hassabis claims this breakthrough is “It’s fair to say that this is five to ten years ahead of what people were expecting, even experts in the field,” there is a game that has not been beaten by a computer just yet and it is poker.

Ten-player, no-limit poker still qualifies as a struggle for researchers of artificial intelligence who are working on solving the big picture issues but AlphaGo is a step in the right direction. Games such as chess, checkers and Go are played with a framework of well-defined rules. Players have all the information in front of them that is clear. Computer algorithms take advantage of this environment but a game such as no-limit poker players are working with incomplete information. Teaching an algorithm to go beyond defined rules to make assessments about its environment is the obstacle in the way of advancement in artificial intelligence. Professor Bart Selman from the Cornell University computer science lab says the ultimate goal for computers is “common sense understanding,” computers that see the world in the same way humans do.

Toby Walsh, a professor of AI at Australia’s University of New South Wales said, “Apart from the other variants of poker, uncertainty and randomness, there’s a third feature; psychology.”



Related news

Return to Latest News