4

I have been doing some research into artificial intelligence and I came across a 2007 paper titled "Checkers is Solved" on the game of checkers being solved by AI techniques after more than 16 years of trial. A solution to the game is defined by the team as "determining the final result in a game with no mistakes made by either player". The search for a solution started back in 1989 and it was finally found in 2007. Of what importance is this to the field of AI?

Thomas Owens
  • 82,739
cobie
  • 3,237
  • 2
    Could you try to be a bit more specific? In its current state your question is a good conversation starter, but not a good fit for programmers. – Simon Bergot Oct 10 '12 at 16:46
  • 2
    It might be useful if you cite the paper that contains this information. If possible, provide a link to the paper or someplace where the paper can be obtained. – Thomas Owens Oct 10 '12 at 17:07
  • Did the solution use something other than min-max? – Gort the Robot Oct 10 '12 at 17:11

1 Answers1

4

When you explore AI, you do so first with a problem space that is relatively small and bounded by specific rules, so that you can prove your techniques.

Checkers is such a space. It therefore makes an ideal playground for AI explorations. This is also true of chess, although checkers is a smaller problem space, and therefore is more suitable to simpler explorations of AI. Board games like chess and checkers are easily modelled in a computer using a two-dimensional array.

As your knowledge improves, you can begin to alter your techniques and apply AI to more complex and non-deterministic domains, like human interactions.

Steven Evers
  • 28,180
Robert Harvey
  • 199,517
  • Or the game go, which was once thought to be immune compared to relatively simple games such as chess. Go is now also being (somewhat) successfully attacked by AI techniques. – David Hammen Oct 10 '12 at 17:17
  • @DavidHammen, if humans can play it, AI can compete. – Philip Oct 10 '12 at 20:23
  • @DavidHammen I think Go is still lands in the "can't be solved" zone for the normal 19x19 boards - there's simply too many possible board positions. But I'm not completely certain. – Izkata Oct 10 '12 at 21:46
  • Current Go engines are using a Monte Carlo randomization which produces good enough moves to compete with accomplished amateurs. – Gilbert Le Blanc Oct 11 '12 at 14:34