The Ultimate Poker Showdown: Professionals Versus Artificial Intelligence

20 years ago, in 1997, IBM made headlines for building Deep Blue, the super-computer that defeated world chess champion, Garry Kasparov, in a six-game match.

14 years later, IBM came back with the new-and-improved Watson, and took artificial intelligence to the next level in 2011. They paired Watson against two Jeopardy champions, and the world watched as ten racks of ten Power 750 servers defeated took home the victory.

And just last year, the world of artificial intelligence was brought again to the forefront when it defeated a human champion at the game Go.

Now? According to The Verge, a team from Carnegie Mellon University is striving for a new milestone. They want to know whether or not artificial intelligence can defeat some of the best human players at Heads-Up No-Limit Texas Hold'em poker.

What makes this challenge slightly different than the chess, Jeopardy, or Go scenarios, is that Texas Hold'em is what's called an "imperfect-information game."

Professor Tuomas Sandholm, who built the Libratus systema along with PhD student Noam Brown, told The Verge, that, "in a complete information game you can solve a subtree of the game tree. With incomplete-information games, it's not like that at all. You can't know what cards the other player has been dealt. That means you don't know exactly what subgame you're in. Also, you don't know which cards chance will produce next from the deck."

If you've ever watched a match of Texas Hold'em on ESPN, you are aware of how much the game depends on human interaction--no matter how subtle. It is a game of bluffing, strategically timed bets, a close watching of nervousness, and subconscious giveaways such as ticks, tense finger tapping, and more. A human game, to say the least.

To provide an added perspective to what this sort of artificial intelligence experiment could mean for the world of poker in general, I reached out to Dan Fleyshman--the serial entrepreneur behind Victory Poker, which at the time was one of the largest online poker brands in the market.

"I have always thought how crazy it is that a huge percentage of poker players are proud that they've never read a strategy book or utilized technology," he said. "They believe because they've been playing poker for years that the experience makes them elite. What they don't realize is they've been practicing certain wrong habits or decision patterns for thousands of hours."

This is fascinating commentary alongside the development of artificial intelligence, and the level to which we believe we can calculate our way through "imperfect-information" exchanges: which include things like language, for example. At what point will the scales tip, and will the pure act of hours and practice for a human not be able to compete with the algorithmic advantages of a computer?

One of the four poker pros participating in this head-to-head showdown with artificial intelligence system named Libratus, is Jimmy Chou, who told The Verge, "I do not believe that poker is different enough from chess and Go, and ultimately think that computers will dominate the game."

The matches begin on January 11th in Pittsburgh, Pennsylvania, at the Rivers Casino. At stake is a prize of $200,000. In order to come home with a victory, Libratus will have to duke it out against four pro poker players over the course of 20 days--during which a whopping 120,000 hands will be played. 

This article passed through the Full-Text RSS service - if this is your content and you're reading it on someone else's site, please read the FAQ at http://ift.tt/jcXqJW.
Recommended article: The Guardian's Summary of Julian Assange's Interview Went Viral and Was Completely False.

from Poker