Monday, November 08, 2004
Center of Excellence Proposal
à
White Paper on Incident Information Orb
Architecture (IIOA) à
Adi Structural Ontology Part I
à
Cubicon language descriptive
à
Using the Readware conceptual role-up of real time harvested text
Peter Stephenson and I meet with the publisher who will publish the book on cyber attack mechanisms, authored by Peter and I. We will take between 8 – 18 months on this project, depending on when we can really start some experimental work related to what we have outlined as a notational foundation of categoricalAbstraction (cA) and eventChemistry (eC) as applied to modeling the complex world of vulnerabilities and attacks. The notational foundation is to be included in section 1 of this book, and will support the full generative methodology.
I start this note off with this news just to put into context some thoughts I had today about the opportunity to reveal Readware as a different type of technology.
I have also been revising the draft of my book, Foundations for Knowledge Science.
The issues related to what the new science is, that the old science has been on my mind, and are expressed in the revision of the preface to this book.
The generation of structural ontology allows things that cannot be done otherwise. By “structural ontology, we mean, of course, a stratified ontology that starts with the designation of a substructural ontology and precedes two levels of organization so that function and behavior finds a place as the middle ontology. The first instance of this being done is the Tom Adi structural ontology for natural language. However, we are not quite finished with all three parts. The group working under Pospelov and Finn certain was describing something that is very similar to what Tom and I have called a structural ontology.
So structural ontology is a means to conduct an investigation about something. This investigation have various levels of abstraction and in each level the investigator(s) can come to an agreement over how to characterize behavior and function. Structure is measured via some type of instrumentation.
The old science could take abstraction, mathematics and formalism related to conservation laws, and set up a laboratory and point and say "see here" this experiment goes exactly in accordance with the model which we have developed.
Complex systems, such as those we study, do not have such models.
The notion of a higher order abstraction over the event structures that make up cyber attacks is one that Peter Stephenson has brought forward from within his deeper self. His view of this helps us see something about what is in common between one application of the generative methodology and another application. The opportunity to develop a structured ontology for cyber attack mechanics is very intriguing.
Well back to this notion of a difference between the old science and the new science.
The new science can develop as a higher order abstraction, what Sandy calls the behavioral level, without there being the ability to say "see here" and point to a specific realization of that abstraction as an instance of theory. There is a reason for this. The behavior/functional level is not merely defined by one instance, but is defined as a invariance across many instances. Otherwise the function is not known to the system that needs the function to occur. (This is not really the definition of a utility function that is driving an evolution, but perhaps some will see it this way.)
These thoughts are the beginning of a lot of words that I have in my mind this evening. I will let others find this words in their own mind and in their own way.
I was deeply grateful for the opportunity to visit with Peter this afternoon.
The subject line of this note is about anticipatory e-commerce, only because there is a possible path that would allow a core team to rapidly take a step towards useful representation of real time social discourse in the form of Orb constructions useful in e-commerce.
The importance of this difference between non-stratified science and the new science is in the notion of repeatability. There is no complete fidelity in the pointing to specific instances in the world from the middle level of abstraction. Each instance of a behavior in the world is unique. This will make the descriptions of Readware cultural interests seem non-scientific. E-commerce applications will not have a problem with this imprecision, but scientific peer review has had a problem, as we have all noticed.
But the fact is that the structured ontology provides already a lot of knowledge that becomes realized in the specifics of “reading” the thematic structure at this moment or that moment in this community or that community. More can be done with the focus of applications once we have the first serious deployments.
For example, the measurement of the "stems" in English by Readware is one that can be used to parallel the Wycal patent. As Wycal’s patent describes, for text understanding, how one can carry the themes at a primary and secondary focus and pick up the social interests expressed within a community, and relate these to what Tom has called the cultural ontology.
The measurement, instrumented to look at the substructure as revealed in the stemming process, is occurring in real text. The structured ontology is supplying the higher-level abstract knowledge regarding patterns that recall cultural interests. The specifics of the details are not important, except as a means to guide the formation of this high level contextualization. Thus the measurement of results from a stemming process reveals cultural interests. These specific interests might have many ways in which they might have been expressed. One does not see this detail, only that some specific behavior or function is involved.
Then, if there is an interest, a further investigation of specifies of the words being used can be made. At this point, the dots are connected and the unique specifics of this situation are available.