From: Ben Goertzel (firstname.lastname@example.org)
Date: Tue May 28 2002 - 22:47:58 MDT
I finally got around to reading some of Arthur T. Murray's stuff on the
Mentifex AI approach.
I didn't look at the code (yet), I just read the docs on the site.
If anyone is interested, these two are fairly clear and simple (esp. the
The "Nolarbeit Theory Journal" documents on the site, I found rather hard to
follow, although interesting in a literary and psychological sort of way.
I like the simplified brain theory underlying the Mentifex system. It is
really pretty similar to the brain theory I put forth in From Complexity to
Creativity (and earlier, in an article published in Complexity magazine):
Had I known about the Mentifex work then, I would have cited it there....
Both theories use the hierarchical structure of the cortex, and connect it
to the dual-network-like structure of mind. This stuff provides background
for my more recent work on hebbian logic -- hebbian logic is about reasoning
among neural clusters, and this is about how neural clusters work more
However, my own AI work is not based on my theories about how the brain
works except in a very loose way. Murray proposes an AI architecture based
more closely on his theory of the brain.
He says a lot about how brain structure affects linguistic processing. Here
I agree with most of his general conclusions about how linguistic structure
probably relates to brain structure and dynamics. However, I doubt that
these general conclusions can be implemented simply in software. I think
that the linguistic behaviors he discusses do come out of the brain
structures he mentions, but that there's a lot of detailed brain structure
and dynamics intervening between the two levels, which his software and
theory don't embody. I could go into more details but don't have time right
In my view, although Arthur sometimes *presents* his ideas in a somewhat
kooky way (by the standards of the mainstream scientific community, and even
by the standards of this list), the ideas themselves are significantly
better than most of what passes for cognitive science and AI. There is some
deep thinking there. If anyone else but me is trying to survey all serious
thinking on AGI, Murray's two papers I cited above should be looked at for
-- Ben G
This archive was generated by hypermail 2.1.5 : Sun May 19 2013 - 04:00:41 MDT