From: Cliff Stabbert (cps46@earthlink.net)
Date: Sat Jul 27 2002 - 08:50:18 MDT
In rereading some of Douglas Hofstadter's articles on Copycat, I was
struck by a point he makes in the post script of "Analogies and
Roles in Human and Machine Thinking", published in _Metamagical
Themas_.
(For a brief background on Copycat, see
http://www.csu.edu.au/ci/vol03/paper4/node3.html
The Copycat project is discussed at greater length in Hofstadter's
_Fluid Concepts and Creative Analogies_, and lisp source is readily
available)
Specifically, in arguing that some analogies "are better" than others,
he argues for a "survivability value" -- i.e., some analogies give you
a better chance of survival, or some ways of analogizing give you a
better chance of survival than others.
With that in mind, I'm curious -- and this may not be the right place
to ask, but I'm hoping someone here has an idea -- whether any work
has been done to combine Copycat-like projects with neural networks,
genetic algorithms or something along those lines, i.e., evolving
Copycat's underlying rules and structures based on some type of
virtualized environment. I mostly agree with Hofstadter's (implied)
assessment that the human ability to form analogies both lies close to
the center of intelligence and has a high survival value.
In briefly following up, I find Marshall and Hofstadter on
"self-watching":
http://www.cs.indiana.edu/event/maics96/Proceedings/Marshall/marshall.html
...which is interesting, but disappointing to me in that the
survivability thread appears to have been dropped.
I'm curious whether others on this list have been looking into
emergent analogy-forming Copycat-like systems and searching for
automated or semi-automated (e.g., via virtual environments) ways of
evaluating and/or evolving their utility. I realize that the focus of
this list is not quite so narrow but am having trouble locating a
narrow list wide enough.
-- Cliff
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:40 MDT