minsky at media.mit.edu (Marvin Minsky) wrote:
>marko_amnell at hotmail.com (Marko Amnell) wrote:
> > HOW BRAINS THINK: THE EVOLUTION OF INTELLIGENCE
> > by William Calvin is a good exposition of the
> > "neural Darwinism" thesis by the man who coined
> > the term "Darwin machines". Occasional RAB contributor
> > Marvin "The Society of Mind" Minsky gets a mention in
> > a footnote for his 1994 Scientific American article
> > "Will robots inherit the earth?".
>> I've put that on my home page at
>http://www.media.mit.edu/people/minsky/.> Actually it's more about the prospects for human immortality.
Thanks. _Consciousness: How Matter Becomes Imagination_ by
Gerald Edelman and Giulio Tonori is another good book on
Neural Darwinism. It's more comprehensive than William Calvin's
_How Brains Think_ (and even contains a few mathematical
equations). I believe Edelman actually was the first to propose
the Neural Darwinism hypothesis in the late 1970s. What is it?
Well, just listing a few principles doesn't explain much, but
this is how Edelman and Tononi introduce Neural Darwinism
(pp. 79-85):
"In considering the origin of species, Charles Darwin made a
great contribution that centered on population thinking: the
idea that variation or diversity among individuals in a population
provides a basis for competition during natural selection. Natural
selection is reflected in the differential reproduction of fitter
individuals in a species. In principle, selective events require
the continual generation of diversity in repertoires of individual
variants, the polling by environmental signals of these diverse
repertoires, and the differential amplification or reproduction
of those repertoire elements or individuals that match such
signals better than their competition. Could it be that the
brain follows such principles? We believe it does, and in this
chapter we briefly review some aspects of neuronal group selection,
or Neural Darwinism. This theory embraces these selective principles
and applies them to the functioning brain. Its main tenets are
(1) the formation during brain development of a primary repertoire
of highly variant neuronal groups that contribute to neuroanatomy
(developmental selection),
(2) the formation during experience of a secondary repertoire of
facilitated neural circuits as a result of changes in the strength
of connections or synapses (experiental selection), and
(3) a process of reentrant signaling along reciprocal connections
between and among distributed neuronal groups to assure the
spatiotemporal correlation of selected neural events.
Together, the three tenets of this global brain theory provide a
powerful means for understanding the key neural interactions that
contribute to consciousness. ...
This theory of neuronal group selection (TNGS), or Neural Darwinism,
has three main tenets ...
1. Developmental selection. During the early development of
individuals in a species, formation of the initial anatomy of
the brain is certainly constrained by genes and inheritance.
But from early embryonic stages onward, the connectivity at the
level of synapses is established, to a large extent, by somatic
selection during each individual's ongoing development. For
example, during development, neurons extend myriads of branching
processes in many directions. This branching generates extensive
variability in the connection patterns of that individual and
creates an immense and diverse repertoires of neural circuits.
Then, neurons strengthen and weake their connections according
to their individual patterns of electrical activity: Neurons
that fire together, wire together. As a result, neurons in a
group are more closely connected to each other than to neurons
in other groups.
2. Experiental selection. Overlapping this early period and
extending throughout life, a process of synaptic selection
occurs within the repertoire of neuronal groups as a result
of behavioral experience. It is known, for example, that
maps of the brain corresponding to tactile inputs from the
fingers can change their boundaries, depending on how much
different fingers are used. These changes occur because
certain synapses within and between groups of locally
coupled neurons are strengthened and others are weakened
without changes in the anatomy. This selectional process
is constrained by brain signals that arise as a result of the
activity of diffusely projecting value systems, a constraint
that is continually modified by successful output.
3. Reentry. The correlation of selective events across the
various maps of the brain occurs as a result of the dynamic
process of reentry. Reentry allows an animal with a variable
and uniquely individual nervous system to partition an
unlabeled world into objects and events in the absence of a
homunculus or computer program. As we have already discussed,
reentry leads to the synchronization of the activity of
temporally coherent output. Reentry is thus the central
mechanism by which the spatiotemporal coordination of
diverse sensory and motor events takes place.
The first two tenets, developmental and experiental selection,
provide the bases for the great variability and differentiation
of distributed neural states that accompany consciousness. The
third tenet, reentry, allows for the integration of those states. ...
It is important to emphasize that reentry is not feedback. Feedback
occurs along a *single* fixed loop made of reciprocal connections
using previous *instructionally* derived information for control
and correction, such as an error signal. In contrast, reentry
occurs in selectional systems across *multiple* parallel paths
where information is not prespecified. Life feedback, however,
reentry can be local (within a map) or global (among maps and
whole regions).
Reentry carries out several major functions. For example, it can
account for our ability to discern a shape in a display of moving
dots, based on interactions between brain areas for visual movement
and shape. Thus, reentry can lead to the construction of new
response properties. It can also mediate the synthesis of brain
functions by connecting one submodality, such as color, to another,
such as motion. It can also resolve conflicts among competing
neural signals. Reentry also ensures that changes in the efficacy
of synapses in one area are affected by the activation patterns
of distant areas, thereby making local synaptic changes context-
dependent. Finally, by assuring the spatiotemporal correlation
of neuronal firing, reentry is the main mechanism of neural integration.
Since the formulation of the TNGS, considerable evidence to support
the theory has accumulated."
Unfortunately, Edelman and Tononi are not very explicit about
the nature of this evidence. Does anyone know what is the best
current evidence in support of the hypothesis of Neural Darwinism?
Or is there some crucial evidence against the theory? Can anyone
recommend other good books on Neural Darwinism, for or against?
Are there good articles or papers on the subject that are
available on the world wide web? Thanks in advance.
