AI system proves mental superiority for the first time, clinches Poker battle against human experts in Texas

Scientists have managed to create an Artificial Intelligence (AI) system that has successfully defeated human experts at a game of Texas hold ‘em poker for the first time ever.

The artificial intelligence is not just confined to the Poker table and has implications far beyond the game. The artificial intelligence can be implied from helping make more robust medical treatment recommendations to developing better strategic defence planning.

Researchers at the University of Alberta in Canada created DeepStack that bridges the gap between the approaches used for games of perfect information like Chess and Go where players can see everything on the board - with those used for imperfect information games by reasoning while it plays, using “intuition” honed through deep learning to reassess its strategy with each decision.

Michael Bowling, professor at the University of Alberta said that Poker is one game which has been a challenge in artificial intelligence.
“Poker has been a long-standing challenge problem in artificial intelligence,” he said.

“It is the quintessential game of imperfect information in the sense that the players don’t have the same information or share the same perspective while they’re playing,” said Bowling.

In order to test their theories, AI researchers have been using parlour games because such games are mathematical models describing how decision-makers interact.

DeepStack has the ability to think about each situation during the play to imperfect information games using a technique called continual re-solving.

DeepStack determines the correct strategy for particular situation in Poker as it uses its “intuition” to evaluate how the game might play out in the near future and it doesn’t think about the entire game.

“We train our system to learn the value of situations. Each situation itself is a mini poker game,” said Bowling.

“Instead of solving one big poker game, it solves millions of these little poker games, each one helping the system to refine its intuition of how the game of poker works,” he said.

“This intuition is the fuel behind how DeepStack plays the full game,” he added.

For the complex problems like heads-up no-limit hold’em, thinking about each situation as it arises is important. It has vastly more unique situations than there are atoms in the universe, largely because of the ability of the players to wager different amounts including the dramatic “all-in.”

DeepStack takes action at human speed despite the complexity of the game with an average of only three seconds of “thinking” time - and runs on a simple gaming laptop.

Last December, DeepStack played against a pool of professional poker players in order to test the approach. These experts were recruited by the International Federation of Poker.

No less than 33 players from 17 countries played a 3,000-hand match over a period of four weeks. Each of the 11 players who finished their match, were beaten by DeepStack with only one outside the margin of statistical significance.