AI's have beaten the best human players in chess, go, and now poker.
In a landmark achievement for artificial intelligence, a poker bot developed by researchers in Canada and the Czech Republic has defeated several professional players in one-on-one games of no-limit Texas hold'em poker.
Perhaps most interestingly, the academics behind the work say their program overcame its human opponents by using an approximation approach that they compare to "gut feeling."
"If correct, this is indeed a significant advance in game-playing AI," says Michael Wellman, a professor at the University of Michigan who specializes in game theory and AI. "First, it achieves a major milestone (beating poker professionals) in a game of prominent interest. Second, it brings together several novel ideas, which together support an exciting approach for imperfect-information games."
Source: Poker Is the Latest Game to Fold Against Artificial Intelligence
Is there anything at which AI's won't soon be able to beat humans?
(Score: 2) by FatPhil on Friday January 13 2017, @03:45PM
What makes you say that? From TFP itself (available on Arxiv) "... using deep learning". This is almost certainly a completely generic program and hardware, it's simply been trained to map poker game states as inputs into poker moves as outputs. It could almost certainly just as easily be trained to take youtube URLs as inputs and return the number of cats in the video as output.
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 2) by FatPhil on Friday January 13 2017, @03:53PM
"DeepStack is a general-purpose algorithm for a large class of sequential imperfect information games."
"depth limited lookahead where subtree values are computed using a trained deep neural network"
"Instead of solving subtrees to get the counterfactual values, DeepStack uses a learned value function intended to return an approximation of the values that would have been returned by solving."
"general-purpose", "trained", and "learned" say "AI" rather than expert system to me.
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 1, Informative) by Anonymous Coward on Friday January 13 2017, @05:46PM
The problem you're facing is unfamiliarity with the terminology of the field.
General-purpose in this context means that it isn't tuned to a specific game. It's not an architecture for arbitrary implementation of machine cognition, it's an "algorithm for a large class of sequential imperfect information games." "trained" means that they used a set of data to tweak the neurally computed function. "learned" as opposed to hard-coded.
I see how you got there, but the problem is semantic overload in the context of domain-specific jargon.
(Score: 0) by Anonymous Coward on Friday January 13 2017, @11:41PM
thats certainly moer useful and appealing.