The best Poker players Couldn't
This supercomputer at the peak Of its capabilities produces
Original article: Almost immediately after The new year, four of The best professional poker players Faced off against AI artificial Intelligence in a tournament organized At the Rivers Casino in PittsburghOn January, the tournament marathon Ended, and days later, it Became clear that Libratus artificial Intelligence had outperformed four world - Class professionals in one of The most popular poker games-Texas hold'em. Throughout the tournament, computer intelligence Faced off against four professionals: Jason Les, Dong Kim, Daniel McAuley, and Jimmy Choo. During the entire tournament, they Played approximately, hands. AI beat people earning $. Libratus performed all calculations for The game using the computer Power of the Pittsburgh supercomputing center. Most likely, it was the Most difficult, intense and interesting Tournament in my voluminous professional career. The great thing about it All was the adaptability of The computer. It was noticeable how he Learns and gets better every Day, noticing our punctures and Weaknesses, " says one of the Participants in the gameplay, Jimmy Choo. Texas hold'em is considered A fairly complex game with More than ^ game situations.
petaflops this is times faster Than any expensive laptop
And in order to calculate Them all, Libratus took computing Nodes out of possible Pittsburgh supercomputers. The best qualities of AI Are to make strategically sound And highly correct decisions based On incomplete information. This is what exceeds the Best capabilities of people, " said Project curator Thomas Sandholm.
Earlier, a computer science scientist Justified why it is so Important AI wins in a Game like poker.
Poker requires the computer to Make extremely complex decisions based On incomplete information, bluffing, and The pace of the game.
According to the Professor, teaching AI to understand games with Incomplete information so efficiently is A very important task, because Many processes in real life, Such as business negotiations and Medical forecasts, are modeled using This algorithm.
How exactly Libratus AI learns Itself, the scientists decided not To tell yet, but promised Scientific publications at the end Of the tournament. It is worth Recalling that In March, the artificial intelligence Of the AlphaGo system developed By DeepMind was able to Beat the world champion of The go game Lee SEDOL With a score of.