Posted on 05/10/2015 6:15:05 AM PDT by WhiskeyX
Strategy games such as chess have long been considered important ways to measure artificial intelligence. But A.I. researchers at Carnegie Mellon University chose a different, and in some ways, more challenging game: poker. Susan Koeppen of CBS Pittsburgh station KDKA reports on what happens when the chips are down.
Doug Polk, 26, is considered the best heads up, or one on one, no limit Texas hold 'em player in the world. He's defeated countless opponents and won millions of dollars.
Polk bet his reputation that he could beat Claudico, Carnegie Mellon's artificial intelligence super computer.
(Excerpt) Read more at cbsnews.com ...
An article from yesterday with a few comments:
http://www.freerepublic.com/focus/f-chat/3287982/posts
CMU computer beaten by humans [Texas Hold ‘Em] ^
A computer is only as smart or skillful as it’s programmers.
A meaningless endeavor
Poker played perfectly without the human element of bluffing and reading players becomes a pure game of chance.
If a human player plays mathematically perfectly, of course he can beat a computer player being randomly dealt slightly weaker hands during THAT game.
For infinite games, the computer would eventually win because humans make mistakes. Or it would be perfectly 50/50 IF the human player never makes a mistake
I think your analysis is simplistic. The human can make a machine-learning computer think the human is a “bluffer” by repeatedly losing hands with small stakes. Then the human sinks his teeth in when he has a good hand. This is very complex and I would hate to call it simply a game of chance.
I imagine the poem excerpted below has probably been removed from any modern day curriculum, but I remember it from the late 60's.
John Henry by Anonymous.
When John Henry was a little tiny baby
Sitting on his mama's knee,
He picked up a hammer and a little piece of steel
Saying, "Hammer's going to be the death of me, Lord, Lord,
Hammer's going to be the death of me."
.
.
.
John Henry hammered on the right-hand side.
Steam drill kept driving on the left.
John Henry beat that steam drill down.
But he hammered his poor heart to death, Lord, Lord,
He hammered his poor heart to death.
Well, they carried John Henry down the tunnel
And they laid his body in the sand.
Now every woman riding on a C and O train
Says, "There lies my steel-driving man, Lord, Lord,
There lies my steel-driving man."
The answer is no. I can bluff that machine until it’s power supply shorts out.
But the point of an AI poker playing program is precisely to include bluffing and reading players (on the basis solely of their play pattern) not just a perfect calculation of probabilities. And yes, that too can be mathematically modeled so that an algorithm can make good decisions on when to bluff and judge the likelihood that the opponent is bluffing — a lot of game theory deals with situations that are exactly like the chance-less aspect of poker.
One could even equip a computer with a computer vision system and adaptive algorithms that read the players facial expressions and gestures, but since the computer has no tells outside of any found in its play pattern, that would seem to be an unfair advantage to the machine (like the perfect rock-paper-scissors playing machine has).
I think it depends on your point of view.
In actual numbers, the humans beat the computer, by combining all of their winnings and coming up with a net of 732,000.
But, that was after 80,000 hands, over two weeks. (750 hands at a time, twice a day, for 13 days, with each of 4 different poker players). In total, $170 million was bet.
So, a $732K edge was effectively a statistical tie.
More info here: https://www.cs.cmu.edu/brains-vs-ai
How is this poker if the money was not real and the computer does not have a spouse at home with whom he/she has to discuss why they no longer own their home?!!!
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.