[Somewhat Off Topic] AI Contest

Peter Hansen peter at engcorp.com
Mon Aug 6 02:00:54 CEST 2001


Roman Suzi wrote:
> 
> On 5 Aug 2001, Doug Fort wrote:
> 
> >I'm not convinced that logic programming == AI. I'm more of a bottom-up
> >genetic algorithm fan myself.
> 
> Well, "I" could be different. I always thought having logic is necessary
> to be intelligent. Evolution is not intelligent, but it solves problem.
> Probably, it could be called Artificial Intuition.

I'm with Doug on this.  Intelligence does not necessarily imply logic,
at least not in most of the formal senses.  Intelligence does, however, 
imply the acquisition of knowledge (learning) and its application.
(At least using the first definition from dictionary.com :-)

I consider genetic algorithms a good current example of what we
used to call artificial intelligence.  For certain kinds of problems,
especially some with which even we humans have difficulty, they are
very effective learners.  

The knowledge a GA acquires is encoded in the genomes of the 
evolved population.  Once a GA has found an acceptable solution, 
it is possible to analyze the best genomes and "reverse engineer" 
the encoded information, revealing things which would not have been
readily apparent to a human.  I call that intelligence of a sort,
and it sure beats my cat for problem solving.  (Disclaimer: my
cat is not the brightest cat, and some cats may exhibit greater
intelligence than current genetic algorithsm; no offense to
cats in general was intended by this posting.)

And I've been waiting for a chance to use Python (Stackless, of
course!) for some GA work.

-- 
----------------------
Peter Hansen, P.Eng.
peter at engcorp.com



More information about the Python-list mailing list