In article <[log in to unmask]>, Declan Fox
<[log in to unmask]> writes
>But do we yet have any machine "intelligent" enough to
>class as AI?
Doesn't that depend on the algorithms used? I take your point though;
unless we know how intelligence works then we cannot reproduce it. The
problem is that I don't think we even know the language yet, although
Chaos seems a promising start where the reiteration of a single, simple
algorithm such as the logistic difference equation can produce a pattern
of self-organisation.
I think as doctors we have a wonderful substrate of data where we work
intelligently with uncertainty, providing probabilities with inadequate
information and coming up with a reasonable answer. We mix information,
misunderstanding, half-truths, emotion, intuition, semi-related facts
and figures, stir it with knowledge and experience and come up with a
reasonable answer, usually. Although the average person attributes a
direct cause and effect to what we do, this is invariably a simplistic
metaphor for what actually happens.
Although I don't know much about AI, I don't think we're much beyond the
stage of genetic algorithms. If we can find how to seed the information
soup with the correct simple algorithm to produce apparently unrelated
results, except by Chaotic standards, then I think AI becomes a
realistic proposition.
Regards
George
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|