From: Ben Goertzel (ben@goertzel.org)
Date: Wed Mar 06 2002 - 19:02:24 MST
Hi,
In the document that I posted (a link to) here recently, I defined three
kinds of intelligent self-modification: noetic, dynamical and substructural.
Here are some more explicit comments on how these relate to my Novamente
work.
In the beginning of that paper, I say:
****
Suppose we have a system S, which is attempting to achieve goal G relative
to environment E. Let S(t) denote the state of S at time t.
There are three different senses in which we can say that S is
self-modifying. These may be quantified as three different quantities
associated with a system S at a time t:
· Intelligent noetic self-modification: the degree to which changes in the
system’s state correlate with increases in the system’s intelligence
· Intelligent dynamical self-modification: the degree to which changes in
the system’s dynamics correlate with increases in the system’s intelligence
· Intelligent substructural self-modification: the degree to which changes
in the system’s “pragmatically meaningful state space” correlate with
increases in the system’s intelligence
****
The paper also gives a formal definition of intelligence, based on the
notion of intelligence as "achievement of complex goals in complex
environments," and the quantification of complexity in terms of algorithmic
pattern.
Now, any AI system worthy of the name will display intelligent noetic
self-modification. All of Novamente’s cognitive dynamics are specifically
oriented toward this kind of self-modification: changing the state of the
system to make the system achieve its goals more effectively.
Intelligent dynamical self-modification is to some extent implicit in
intelligent noetic self-modification, but, in Novamente, it becomes much
more significant when schema learning becomes a major part of the system’s
dynamics. Schema learning is the mechanism by which Novamente explicitly
modifies procedures by which it does things. When the system begins
modifying its own basic cognitive dynamics, then intelligent dynamical
self-modification becomes truly prevalent.
Finally, intelligent substructural self-modification is happening all the
time in Novamente as well – but it gets more and more significant as schema
learning and basic-cognitive-dynamic learning rise. And it becomes truly
overpowering when the system becomes able to rewrite its own sourcecode and
transform itself into something entirely different than it once was – in
this case it is modifying its own substructure in a direct and obvious way.
-- Ben
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT