AI Program Beats Humans In The Poker Game
A poker-playing AI has beaten four human players in the marathon match which has lasted 20 days.
An artificial intelligence program that was developed at Carnegie Mellon University named Libratus was trained to play the variant of the game known as no-limit heads-up Texas hold ’em.
The humans won in a similar tournament in 2015.
The victory has been hailed as the significant milestone for the AI, by the team responsible for building it.
The AI has won more than $1.5m (£1.2m) worth of chips from the humans. The matches were held at the Rivers Casino in Pittsburgh, were the live-streamed over gaming site Twitch.
Click here for more Technology News
The professor of computer science at the Carnegie Mellon, Tuomas Sandholm, said that the event as “historic.”
He has said that “heads-up no limit Texas hold ’em was in a way the last frontier of all the games.”
“Chess, Othello, Go, Jeopardy has all been conquered, but this remained elusive which was a landmark in AI gameplay.”
Prof Sandholm mentioned that the algorithm could be transferred to a range of other uses.
“It was not just about the poker,” he said.
“The algorithms could take the information and output a strategy in a range of scenarios, including negotiations, finance, medical treatment, and cyber security.”
He also stated that they had proven the ability of AI to do strategy and reasoning. There were many potential applications in future.
Jimmy Chou, one of the professional poker players, had admitted at the halfway point in which the AI was proving a tougher opponent.
“The robot gets better and better each day and looks like a more terrible version of us,” he stated.
Mr.Chou said that”in the first couple of days, they had had high hopes.”
“But every time they had found a weakness, it learns from them, and the fault disappears the next day.”
Click here for Latest Hot News
He also added that the professionals had been sharing their notes and tips to find the weaknesses in the AI’s gameplay.
But they were not only the ones who does the homework. Once after the play ends each night, the Pittsburgh Supercomputing Centre has added computations to sharpen the AI’s strategy.
All the four human players shared of about the $200,000 (£159,000) prize fund and ranked in the order of how well they have played the poker against the AI.
Jason Les, who ranked as fourth, summed up the feelings of all the four players and he added that the match had been “incredibly challenging.”
He had said that he was impressed with the quality of poker Libratus has played.
He further mentioned that “all of us had tried everything that we could do, but it was just too difficult. It became very demoralizing.”