Wednesday, January 04, 2006
New
discussion about signal pathways
and complex ontology
This is part of a discussion that will be moved to a Wiki page soon.
Joanne,
I developed my PhD thesis
modeling Pribram's cognitive neuroscience about frontal lobe function, and
immune theory reaction networks, Jerne, Richter, Burnet.
This set the stage of three years
(1990- 1993) of operating a Neural Network Research Facility at Georgetown
Univ.
Since then I have been working in
the IT consulting industries trying to figure out how to position the
phenomenon of complexity (true natural complexity) ... with some success.
A more recent communication,
this morning, sets up more of the research agenda that I am working on.
I post some part of my discussions at a weblog (called
the Glass Bead Game) but am interested in learning about Wiki pages.
I know Steven Grossberg and Lofti
Zadeh and Paul Werbos fairly well.... and the histories associated with these
individuals.
I am very encouraged by the
BioPax ontology, which I am looking at with Protege - with OWL Plugin.
I might develop a specific
position regarding a step like process for producing
1) "data ontology" such as the BioPax level1
2) data mining programs that act from an external point of view using OWL "retrieval" is the simplest possible form to produce co-occurrence within category data aggregation for human (biological scientist) inspection.
3) controlled vocabulary knowledge management infrastructure (using SchemaLogic's schema server or similar)
4) a
formative and situational biological research tool primarily involved in
realizing QSAR procedures (with Mill's logic)
I developed this position for US
Customs (world wide commodity transactions create "signals"
also :) ) but this work was ignored by the government, but
picked up by www.datawarehouse.com
see link to data warehouse’s serialization
of my work from [118].
If things go right, perhaps we
can look at ontological modeling were degeneracy is part of the architectural
design?
http://www.bcngroup.org/beadgames/generativeMethodology/117.htm
I have an emerging e-commerce
system www.virtualTaos.net
that supports my private research and some industry consulting on Service
Oriented Architecture and OASIS standards for business ontology models.
But my sense is that most in the business
community will not be able to overcome the current IT orientation towards
strong simplifications and anti-complexity. However, in this BioPax
community I see that natural complexity is seen as an open problem - and
unrelated to the computer science "definition" of complexity. So the biologist can more easily say to the
IT person that the system they offer is not sufficient for the full
solution. The BioPAX OWL ontology will
be very important, however, in my estimation.
So the challenge is to develop,
in some way, a functional architecture for scientific work on systems with emergence
and the observed phenomenon of variability in response and selectionism..
Then perhaps we can look at
ontological modeling where degeneracy is part of the architectural design? This means a new generation of business
interaction tools, and tools for e-commerce.
The degeneracy, found ubiquitously
in nature, cannot be formalized (completely) using predetermined categorization
... for two reasons
1) a new category may in fact be emerging in response to evolutionary pressures.
2) the
entailment is both categorically driven, ie can be reduced to classes or at
least concept representations, and structurally entailed.
The mechanisms involved in the formation of natural categories are likely a set of replicator mechanisms, like genes… but unlike the very shallow concept of meme.
This structural entailment was
almost the complete core of the life's work of bio-mathematician Robert Rosen.
(I am be traveling in a couple of
days so I wanted to lay out the tentative position here.... )