Re: Logics at multiple levels of abstraction

From: Michael Wilson (mwdestinystar@yahoo.co.uk)
Date: Tue Sep 13 2005 - 16:31:38 MDT


Phil Goetz wrote:
> if it did appear by chance, it would suggest that you
> hadn't separated levels in your design properly

Personally I'm not nearly as keen on the 'clean separation
of layers' principle as I used to be, and indeed the SIAI
has adopted a more sophisticated view of this post-LOGI.
At present I would see the key components of AGI more as
a fairly complicated directed (dependency) graph, from the
simplest stand-alone mechanisms up to the most complex
composite ones. There is no obvious single metric by which
you could generate a neat 'application stack' diagram.
Furthermore this is all cognitive structure, as opposed to
say Minsky-style agent dependencies are a description of
cognitive /content/. I agree that a simple layering would
be nice, but it doesn't appear to exist and attempting to
shoehorn one in (as with oversimplification in general) will
just break the design. On the plus side, at least in AGI we
can have a reasonable clear and useful (though still artificial)
distinction between structure and content, whereas this is
much harder when attempting to reverse engineer the brain.

>> What makes this a rotten design?
>
> A design in which the different levels of abstraction are
> tightly coupled is harder to understand and debug.

I'm happy that you agree that the human brain is a rotten
design.

 * Michael Wilson

                
___________________________________________________________
To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:52 MDT