DeepStack AI superbrain beats poker pros at no-limit Texas Hold’em

  • AI beats professional poker players at no-limit Texas Hold’em
  • Ability to play poker represents a revolutionary step in the development of AI

An artificial intelligence (AI) system has beaten a selection of professional poker players at a game of no-limit Texas Hold’em. The victory marks a huge step in the development of AI.

Poker Hand
DeepStack has proven machine tops man when it comes to Texas Hold’em poker. Picture: Thinkstock.

Named ‘DeepStack’, the new poker playing machine was developed by scientists from the University of Alberta in Canada, the Czech Technical University and Charles University in Prague. In a gruelling, weeks-long contest 33 players from 17 countries played a 3,000 hand match. DeepStack beat each of the 11 players who finished their match, with only one of those victories falling outside the margin of statistical significance.

2017 is looking set to be a big year for the development of poker playing AI. In January, an AI called Libratus beat four of the world’s best poker players in an intense, twenty-day no-limit Texas Hold’em tournament.

How does poker playing AI work?

Although AI algorithms for games like chess and Go have been around a while, developing a system to play poker has long been a challenge. That’s because, unlike chess, poker is an imperfect information game where players are unaware of the actions chosen by their opponents.

DeepStack’s developers got around this by creating a system they refer to as ‘continual resolving’. Put simply, rather than attempt to predict how an entire game will play out, DeepStack focuses on creating a strategy for individual situations to evaluate how the game will develop in the short term. This, the scientists say, allows DeepStack to refine its own ‘intuition’ for poker.

Poker AI’s other applications

Games such as heads-up no-limit Hold’em provide complex problems for AI, with players’ ability to wager different amounts meaning there are countless situations which can occur in the game. Despite this, DeepStack takes an average of only three seconds thinking time per move, and runs on a simple laptop.

The strategic decision-making systems DeepStack and Libratus employ to play poker have a host of other potential applications, from negotiations to military strategy or planning medical treatments.

Approved Casinos

Want to try an online casino? Approved Casinos

Choose an approved casino from our carefully selected list. VIEW CASINOS