Thursday, June 24, 2004
Theory of Anticipatory
Information
Thursday, June 24, 2004
Communication from: BCNGroup and colleagues
Rick, Leo , everyone... please forward
Here is a recent paper on the use of Semantic Web concepts, as applied to the notion of giving a complete semantics to the NASA Earth Observation Data.
http://esto.nasa.gov/conferences/estc2004/papers/a5p1.pdf
Clearly the implied completeness of a solution to the issue of semantics follows the clarity of a statement that many problems are solved if one has an OWL type representation of human knowledge. This is a style issue, as almost no one actually thinks the an fixed upper ontology will solve the common sense problem, as promised by Doug Lenat at Cyc Corp now almost two decades ago. Lenat's micro-theories are less then complete.
The paper, above, is well presented and does indicate issues that do need to be address with reconciliation of terminology, but the criticality of reconciliation and non-stability on meaning of structure is stressed less that it could be.
In paragraph starting second column:
"Semantic understanding of text by automated tools is enabled through the combined use of i) ontologies and ii) software tools that can interpret the ontologies. An ontology is a formal representation of technical concepts and their interrelations in a form that supports domain knowledge. Generally an ontology is hierarchical with child concepts having explicit properties to specialize their parent concept(s)."
Our problem is in the generality of the solution, and the sense given that this is a rather completely satisfactory solution. Upper ontology for NASA is what is being proposed here, and there is value in any upper ontology that is reasonably developed. But....
The Orbs are capable of producing hierarchical ontology, but this is not what is most natural for Orbs. What is most natural is to not impose an inheritance theory simply because hierarchical structures allow the common types of OIL (Ontology Inference Language). Other types of inference awaits in the wings, as indicated below.
Orbs allow a more complex, as in underdetermined, structure to be defined. The structure is a simple graph that is underconstrained in it's set form, and when aggregated into a specific graph requires some type of convolution. Given a different convolution the limiting distribution, or the "retrieval", is quite likely to be different in major and minor ways. Formative and differential ontology follows...
as discussed in:
http://www.bcngroup.org/beadgames/InOrb/theoryOfInformation.htm
The agility of the knowledge representational encoding, ie the schema, is but one issue that Orb address different than does OWL and OIL. The other issue is the complete separation of structural analysis from imposition of meaning. So a set of qualitative structure activity relationship (Q-SAR) techniques are possible as indicated in
http://www.bcngroup.org/area3/pprueitt/kmbook/Chapter6.htm
Comments..
Dr. Paul Prueitt
703-981-2676