Bryan Moss wrote:
>
> Yes, but doesn't your position suggest a more homogeneous
> brain, which seems at odds with pre-wired instincts? This
> is the problem I see: there's a trade off between
> specificity and constraint on the one hand and generality
> and adaptability on the other hand. That is, I can either
> have a "fragile" brain where I can make specific changes, or
> a flexible brain where I can only make very general changes.
There is *no* tradeoff between specificity and adaptability.
Adaptability is the *result* of specificity. They discuss this *exact
topic* in _The Adapted Mind_. "In the passage from Standard Model to
post-Standard Model psychology it seems fair to say that the greatest
reversal lay in how content-independence and domain-generality came to
be regarded..." Domain-specificity is not a constraint that prevents
the system from behaving correctly, it is a bonus that enables the
system to exhibit more complex adaptive behaviors.
Bryan, you have to read this book!
> (I tend to think of Cyc and "pattern catchers" as being
> opposite extremes in this spectrum.)
I tend to think of classical and connectionist AI as two sides of the
same coin. You might call it Single-Layer AI.
--
sentience@pobox.com Eliezer S. Yudkowsky
http://singinst.org/beyond.html
This archive was generated by hypermail 2b29 : Mon Oct 02 2000 - 17:34:22 MDT