[#!: full-line executeEmbeds, phraseValueList = (("fout fout)("backtrack backtrack))
[#!: path_insertIn_fHand (link d_webWork 'fin Head_one.html') fout ;
Computational neuro-genetic modelling.html
[#!: path_insertIn_fHand (link d_webWork 'fin Head_two.html') fout ;
[#!: path_executeEmbedsInsertIn_fHand (link d_webWork 'Menu.html') phraseValueList ;
[#!: path_executeEmbedsInsertIn_fHand (link d_webWork 'Menu neural nets.html') phraseValueList ;
[#!: path_insertIn_fHand (link d_webWork 'normalStatus.html' ) fout ;
Computational Neuro-Genetic Modelling (CNGM)
From my 2006 draft paper listed below:
"Abstract
– Computational Neuro-Genetic Modeling (CNGM) is discussed from
the perspective of building Artificial Neural Network architectures
starting with substantially pre-defined modules and processes
(DNAANNs). This is equivalent to assuming that DNA code in a neuron
can ultimately specify function, process and some level of data
abstraction beyond the immediate role of genes to produce proteins or
to regulate processes, and using that basis as a metaphor for
DNA-ANNs. A robust and diverse set of ensembles or modules of
DNA-ANNs is sought that is sufficient for a given problem domain, and
that generalizes well. The potential advantages that might be derived
from highly evolved, fine-grained hybrid genetic/connectionist
systems, and some of the implementation challenges that they could
present are discussed." Bill Howell, 2006
While my thrust is from the mathematics/ computer science perspective, the
concept has obvious parallels to the brain and mind, which is
definitely of interest. However, biological substantiation, if this
ever occurs, is only appearing very indirectly, so don't take that
side of it too seriously. On the other hand, work by Michael Meany's
group in Montreal, John Mattick's group at the University of
Queensland in Brisbane, and possibly Sandra Pen˜a de Ortiz in
Puerto Rico, shows that work in genetics is proceeding along many
fronts that may provide eventual substantiation for a biological
basis to CNGM. Furthermore, work by Gary Marcus at New York
University deals with modelling the growth of the brain, and this is
even more of a challenge than the CNGM described here, and indeed is
a necessary basis for it.
My concepts and thinking are explained in the following documents, only
one of which has been submitted in quality format:
William
Neil Howell "Genetic specification of recurrent neural
networks: Initial thoughts", Proceedings of WCCI 2006, World
Congress on Computational Intelligence, Vancouver, paper#2074, IEEE
Press, pp 9370-9379, 16-21 July 2006 This must be obtained
through the publisher. Because of space constaints it is more
limited in scope than the draft below, but it is also a proper,
clean, paper.
Genetic
Specification of Recurrent Neural Networks - This is my
PowerPoint presentation from the International Joint Conference on
Neural Networks in Vancouver, August 2006. The "March of the
Penguins" film fits in beautifully with one sub-class of
consequences of CNGM.
Genetic
specification of neural networks - Draft concepts and
implications of CNGM are discussed in this early version of the
paper that is in the proceedings of the IJCNN 2006 Vancouver
conference. Note that this is a very rough, intermediate draft
that I probably won't get around to cleaning up. However, this does
wander into IMPLICATIONS, from the perspective both of the
mind/brain, and of sort of a "philosophical" perspective
on the analysis of complex systems. Not that it's terribly well
described here.
Junk
DNA and Neural Networks - My conjecture on directions for, and
implications of, the direct involvement of non-protein-coding DNA,
RNA, and epigenetics for more advanced Artificial Neural Network
designs.
Unfortunately,
my electronic documents from the formative 1997-1999 period, and
earlier overview of material circa 1994-965 regarding hierarchical/
structured neural networks are no longer readable. However, there is
a huge amount of research into highly structured systems of
Artificial Neural Networks and the benefits that are being sought.
Directory of available files for this webpage:
Last updated: 05May07, original 27Jan07
[#!: full-line executeEmbeds, phraseValueList = (("fout fout)("backtrack backtrack))
[#!: path_executeEmbedsInsertIn_fHand (link d_webWork 'fin Footer.html') phraseValueList ;