(AMS Bumper Sticker)

Web math-frolic.blogspot.com

Thursday, January 26, 2017

Update on Poker-playing Robot

A couple of weeks ago I mentioned that AI programs were now taking on poker, and a nice followups to that storyline appear this week:
In an ongoing Texas hold’em tournament a poker-playing robot named “Libratus” is so far up by almost $800,000 against its human competitors.
From the the first article: 
“Poker requires reasoning and intelligence that has proven difficult for machines to imitate. It is fundamentally different from checkers, chess, or Go, because an opponent’s hand remains hidden from view during play. In games of ‘imperfect information,’ it is enormously complicated to figure out the ideal strategy given every possible approach your opponent may be taking.”Further it is noted that “...an AI player has to randomize its actions so as to make opponents uncertain when it is bluffing.”
And from the 2nd article:
"One of the things Libratus does well is bluff..."Mastering the art of the bluff requires AI that can calculate risk and reward in real time without having perfect information about what its opponent can do in return. It implies the system does more than simply play a perfectly safe game where it only grinds out wins when it has the stronger hand."
Libratus is specifically programmed (as I understand it) to be skilled at just one specific poker game, so even if it wins this tournament (and it looks like it will, as it seems to be getting stronger over time), it doesn’t mean that bots are on the verge of taking over all professional poker… or, at least not yet. Of course the real congratulations go, not to the bot, but to the clever humans (in this case from Carnegie-Mellon) programming it.

No comments: