5/22/2004 9:03
AM
On the generality of the particular
Key questions on Common Upper
Ontology -> .
(new
beads are edited for a few days until the grammar is correct)
Nan Gelhard and I talked on the phone this morning about the claim that
and the discussion within the SICoP group regarding the use of government and industry resources to create a standard "Upper Ontology".
What is common between my expression of work in the form of the Tri-level (memory, awareness, anticipation) architecture, the voting procedure and formative ontology is that "all ontology is local".
Nan and I take this, "all ontology is local", as principled. I have labored long and hard to make the grounding of this principle in cognitive and quantum neuroscience as clear as possible. So we can just leave this to further discussion.
Under the hood for a tri-level architecture (for developing and using Anticipatory Webs of information) is the voting procedure. The Prueitt voting procedure was my attempt to capture the very simplest form of Finn's quasi-axiomatic theory. The issue, as Finn put it to me on several occasions (personal communication) has to do with a "route to induction". He knew, as did others in his community, that the formation of the mental event involved emergence. So here is the key to creating a computational architecture that extends the individual’s ability to know and interact with (even control) complex event structures occurring in real time in the natural world.
The voting procedure is about resolving an underconstrained computer based representation of ontology in real time as a formative (emergent) event. The tools that are needed to provide the underconstrained representation are independently important. In the simplest cases, I have used the voting procedure to route instructional materials to a user based on a model of what has been learned and what that user has not learned. I used very simple word frequencies as the profile.
John Sowa’s representation of knowledge as simple graph structure is much better than simple word frequencies. Sowa’s work is, as is the quasi-axiomatic theory, an extension of the work of C. S. Pierce. The voting procedure will work in this case also, but one has to think through the issues.
Orbs (Ontology referential bases) are in fact underconstrained knowledge representations, which when viewed by a human becomes actually knowledge. The system we have designed as part of our investment offer to In-Q-Tel would have cognitive graphs as the middle level representation of Orb sub-structure.
TAIParse (from Amnon) would be used in measuring linguistic variation with a theory of grammar type as represented in a RUG (automated rule generation) module. The nine-step actionable intelligence process model has two “first steps”
{ (instrumentation/measurement), (encoding/interpretation) }
TAIParse is one way to address the first step. The output of the first step can be then encoded as relational databases, with fixed data organization, or as Orbs, where the data organization is not fixed (underconstrained). The voting procedure (or full quasi-axiomatic theory) is used in the interpretative process.
The bead games continue to be developed:
PSP