MANILA, Philippines – Artifical intelligence defeated 4 top professional poker players after a 20-day marathon tournament that ended Tuesday, January 31, US time.

The competition was held at the Rivers Casino in Pennsylvania.

One of its creators, professor Tuomas Sandholm, considers it a “landmark step” for artifical intelligence because of the tricky nature of poker.

Poker deals with “imperfect information”, said the Carnegie Mellon University (CMU) professor. This required the artificial intelligence to work out strategies in spite of missing or hidden information. These could be the cards that a competing player holds or figuring out whether a person is bluffing or not.

Unlike games like chess or go, which AI had previously conquered, all the pieces are on the table. All the information pertaining to the current game situation is laid out on the board, which the AI can then analyze for the best move. In poker, there’s a certain degree of intuition required that allows one to “feel” what competing players are thinking.

In winning the tournament, this AI known as “Libratus” has apparently conquered another bastion that seemed, up to this point, an exclusively human thing.

In 2015, Libratus predecessor “Claudico” made by the same AI researchers lost in a similar tournament. Since then, the creators enhanced their AI’s computing power and approach to algorithm. 

Four of the world’s best professional poker players have been competing against CMU’s AI program, called Libratus, since Jan. 11 at @riverscasinopgh. So far, it’s been a tough fight. Tune in at 12:15 p.m. EST on Tuesday for a live update with CMU Prof. Tuomas Sandholm on the “Brains Vs. Artificial Intelligence: Upping the Ante” poker competition.

A photo posted by carnegiemellon (@carnegiemellon) on Jan 23, 2017 at 2:33pm PST

Libratus creator sent an email to Engadget highlighting this milestone in AI progress: “This is the first time that AI has been able to beat the best humans at Heads-Up No-Limit Texas Hold’em. More generally, this shows that the best AI’s ability to do strategic reasoning under imperfect information has surpassed that of the best humans.”

CMU offered some consolation to the poker players who participated in the challenge – Jason Les, Dong Kyu Kim, Daniel McAulay and Jimmy Chou – telling them that for what it’s worth, they pushed the AI to its very limits. The players also get to split a $200,000 prize pot depending on their finish. 

Professor Sandholm and Libratus co-creator PhD student Noam Brown, on the other hand, will be preparing data from the competition to be published in the future. – Rappler.com