4 poker pros lost $1.8 million to an AI program

When it comes to poker, humans have traditionally had the upper hand on computers.

But this week, it was announced that four of the world's best poker players lost nearly $1.8 million (£1.4 million) to an artificial intelligence (AI) program developed by scientists from Carnegie Mellon University (CMU).

The professional players - Dong Kim, Jimmy Chou, Daniel McAulay, and Jason Les - took on the "Libratus" AI agent at a version of poker called no-limit heads-up Texas hold 'em.

The marathon match, held at Rivers Casino in Pittsburgh, Pennsylvania, lasted for 30 days but in the end the AI won $1,776,250 (£1,408,743) over 120,000 hands.

It involved the human players staring at a computer screen for 10 hours a day and being repeatedly trounced by Libratus, according to The Register.

The pros will split a $200,000 (£159,000) prize purse based on their respective performances during the event.

The victory is being hailed as a major breakthrough by those that developed the AI. Tuomas Sandholm, cocreator of Libratus and a machine learning professor at CMU, hailed the event as a landmark moment.

The researchers said that the victory was only possible thanks to a supercomputer, which the AI used to compute its strategy before and during the event.

In a statement on the university's website, Sandholm described how Libratus improved as the match went on.

"After play ended each day, a meta-algorithm analysed what holes the pros had identified and exploited in Libratus' strategy," Sandholm said. "It then prioritised the holes and algorithmically patched the top three using the supercomputer each night. This is very different than how learning has been used in the past in poker. Typically, researchers develop algorithms that try to exploit the opponent's weaknesses. In contrast, here the daily improvement is about algorithmically fixing holes in our own strategy."

Andrew Ng, chief scientist of Chinese tech giant Baidu, compared the victory to when DeepMind's AlphaGo agent beat Lee Se-dol at Go and IBM's Deep Blue, which became the first chess playing programme to beat a human world champion.