From: Ben Goertzel (ben@goertzel.org)
Date: Sat Sep 28 2002 - 15:54:50 MDT
Hi,
I know there are some folks on this list who know the contemporary cognitive
neuroscience literature better than I do.
For you guys, I have a question.
I am interested in the variety of ways in which the brain may store learned
procedural knowledge (knowledge of how to do things -- these could be motor
actions, or they could be cognitive actions, etc.).
How well-demonstrated is it that the brain uses several different types of
representation for the same procedure?
Are there known cases in which the brain uses more than one representation
for the same procedure, because the different representations suit different
purposes?
Is it the case that some common procedure-representation-schemes provide
faster procedure execution than others, and that some provide richer
integration with other parts of the brain than others?
Are those procedure-representation-schemes providing richer integration,
also slower?
I know there is good evidence for a neural distinction between procedural,
episodic and declarative memory. But I haven't seen much of anything on
distinctions between different kinds of procedural memory.
I know that knowledge is "routinized", eg. once you get good at serving a
tennis ball or walking or whatever, your body handles the process in a
different way than before when you were learning. But is it known whether
the routinized knowledge, as represented in the brain, has fewer connections
back to the rest of the brain than the pre-routinized knowledge?
I am asking this stuff because I'm rewriting the section of my in-process
book on Novamente about procedures. Novamente has two different
representations for procedures, one of which provides rich integration with
other parts of the system, the other of which provides fast execution. Some
procedures may be represented the one way, some the other way, and some
both.
I'd like to be able to say something, on a general level, about the relation
between this kind of "dual representation" for procedures, and how
procedural knowledge is dealt with in the brain.
If this kind of question is just beyond the scope of contemporary
neuroscience, that's OK. That's basically what I've said in the current
rough draft of the book. I just want to be a little more certain I'm not
wrong.
I'm going to run the whole book past a friend who's a prominent
neuroscientist, before publication, to be sure I haven't made any brain
science goofs. But he happens not to be available at the moment.... [And I
hasten to add, for those who aren't familiar with Novamente as described on
www.realai.net, that Novamente is NOT based on detailed emulation of the
human brain, although the human brain has certainly been inspirational for
the design in many ways.]
-- Ben G
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:40 MDT