If you see yourself as a top poker player you might want to think again, as artificial intelligence is slowly starting to conquer the field. A poker-playing software called Cepheus is said to constantly win in front of people in a variant of Texas Hold ’em. The new algorithm could one day be a tool for taking out military opponents or even negotiate investment deals.

Researchers have designed a ‘basically unbeatable ‘ software algorithm that can play an essentially flawless session of poker — bluffing included.The poker software is impossible to beat by any adversary in a fair match, as indicated by its maker, researcher Michael Bowling and his partners at the University of Alberta who worked with Finnish programming engineer Oskari Tammelin. In a test, the algorithm managed to solve a specific poker game, dubbed heads-up limit hold’em (HULHE).

According to Bowling, the game of poker has been a test issue for manmade artificial intelligence   for about 40 years, and until today’s innovation, heads-up limit Texas hold’em poker was unsolved. Under the poker umbrella falls a series of games that display flawed information, where players cannot fully know past events. The most well-known variation of poker today is Texas hold’em. The heads-up limit hold’em is designed for only two opponents who play with fixed wager-sizes and number of raises (limit).

While littler than checkers, the defective data nature of heads-up limit hold’em makes it a significantly more difficult game for machines to play or tackle. Bowling noted that his algorithmic achievement turn game-theoretic thinking into big-scale models of  any type in a more controllable manner.

Bowling and his associates developed their calculation in a way it should learn from past games.  The algorithm got its champion-level abilities only after playing over 1,500 sessions of poker.During the first games, settled on its choices haphazardly and afterwards upgraded itself by adding a ‘regret’ value to every choice, contingent upon how inadequately it performed. The computer’s development’s strategy also included the display of some bluffing during the games. Even if the ability to bluff seems to be rooted in human behavior and a psychological feature of the  game, the scientists say this action is a part of game theory in general and computer poker in particular.

Bowling also pointed out that the methodology may be helpful real-life circumstances when one needs to settle on choices with insufficient data — for instance, for dealing with an investments’ portfolio .The group is presently concentrating on applying their methodology to in the medical field, as doctors are frequently in a passion requiring them to take a quick decision without knowing all the facts.  The team is working now with diabetes experts.

Image Source: Science News