From: droleary.usenet@1q2008.subsume.com
In article <479a8a8a$1@news.unimelb.edu.au>, jonesrob@emporia.edu
wrote:
> On Jan 15, 8:21 pm, egs wrote:
> > Can someone help me to find truly simple AI problems?
> >
> > Enrico
> >
>
> I would think that something like:
>
> man(socrates).
> mortal(X):- man(X).
> ?-mortal(socrates).
>
> is the sort of thing that constitutes
> a "Hello World" for AI. I often try that
> sort of thing every time I try out a new
> computer language.
I say that fails the test of intelligence because Socrates is not a man;
he is Socrates. Many a "new computer language" still suffers from doing
class instantiation instead of instance classification. Put another
way, simply saying "birds fly" and "a robin is a bird" doesn't say
anything about any *individual* robin's ability to fly. Nor does it
account for a future observation that "an ostrich is a bird". The
intelligence is not in the data.
I'll join the camp that says tic-tac-toe is a nice beginner AI problem.
It is more approachable to students as a game, and it involves an
element of competition between two players. It is also a solved problem
with a small search space. That means brute force can be used, but you
can also raise higher-level abstractions (e.g., mirrored/rotated
positions) in building a more sophisticated program. From that you can
discuss the distilled rules that lead to a draw and how *that* result is
the intelligence that we strive to get from a computer.
--
My personal UDP list: 127.0.0.1, 4ax.com, buzzardnews.com, googlegroups.com,
heapnode.com, localhost, ntli.net, teranews.com, vif.com, x-privat.org
[ comp.ai is moderated ... your article may take a while to appear. ]
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
