Re: AI architectures in the 2000s

From: Philip Goetz (philgoetz@gmail.com)
Date: Wed Aug 09 2006 - 09:16:14 MDT


On 8/9/06, Ben Goertzel <ben@goertzel.org> wrote:
>
> Phil, do you disagree with this overall intuition regarding the nature
> of intelligence? If so, could you elaborate? What role do you think
> the "special magic trick" at the core of intelligence, if any, plays?
> Do you think it underlies both generalized and specialized pattern
> recognition ability, or do you think it's a special trick for
> generalized pattern recognition? Or would you reject this language
> for discussing things altogether?

I think the language about the "One True Unifying Principle" was a lot
more insulting than that about the kitchen sink.

I'm not big on one-principle AI architectures, at least not those I've
seen so far. Your description of the components of an AI arch is
nice. I do think there may be a fairly simple description of what
brain regions do, but I think that unifying principle may appear at a
lower level in an architecture than the unifying principles in
architectures like Soar, ACT, DTRT, the subsumption architecture, etc.
 If there is any one principle that will make AI easier, it will be a
principle of morphogenesis or self-organization, not something about
how to form chunks or resolve impasses or arbitrate resource
contentions.



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:57 MDT