ORB Visualization
(soon)
Graphical representation of human
knowledge thread -> .
4/18/2004 7:21 AM
The Orbs are theoretical systems having correspondences to physical, biological, psychological or social systems (called collectively “natural systems”). The direct access to order triples in the form < a, r, b > allows for very quick changes in the physical system so that alignments with the theoretical system can be made by changes in what is anticipated, what is in the memory structure and therefore what is in the set of causes of a present engineered physical system. To achieve this control of a physical system one needs a means to act in the physical world based on abstraction information that is developed in the theoretical system.
Natural language understanding is one key application. The theoretic system is needed to develop some representation of the concepts being expressed with natural language, as well as reading between the lines and capturing the motivation, intentions and uncertainties that can be desired or interpreted. How does one control the physical system? An example is given in the next figure:
Figure 1: The use of Ontology referential bases to evoke mental state transitions
In Figure 1, we give a screen shot of a local topological neighborhood over the graph theoretic representation of an abstract Orb where the meaning of the nodes is not known to the reader. However, in the next figure, meaning about the nodes are known and the local topological neighborhoods will evoke in the human a whole series of meanings that depend on shared experiences and on individual tacit knowledge and anticipations.
Figure 2: Two Orb subject indicators from the FCC public ruling (1997 – present)
In the case of taxonomy generation (back of the book index generation) for unstructured text, this physical system is a human or social system. If the Orb subject matter indicators in Figure 2 where generated through an examination of diplomatic discourse about alternatives to beliefs held in Iraq regarding the development of an written National Iraqi Constitution, one might be able to navigate away from the “belief” subject-matter indicator center to the “suggestion” center to see that there are, (I am making this up), four alternatives to a dangerous belief. This navigation in the Orb index could lead to some communication that a specific belief might be perfectly acceptable as an alternative to a belief that violence was necessary to rid Iraq of the Western influence before the natural Iraqi constitution could be reified as a written document. The belief might have been expressed by a group of diplomats who suggested in an obscure journal, or in a chat room, that the Western powers had to accept an insult as a way of saving face. The consequence of allowing this insult affords the positive natural Iraqi constitution to be expressed in the formation of a friendly and democratic nation.
The rationality of this “suggestion” might not be found until the linkage with concepts related to consensus building and disputes are put together in the mind of the President of the United States. Orbs will do these types of things easily, but it is unclear that a relational database would be so agile.
In diplomatic discourse, human use of rhetoric and oratory has long been the answer, and this answer captures a lot of meanings. Clearly, therefore we need a system that generates evocative signs that induce in the human, social system, biological system some internal change in state related to recognition. Something similar is said about the use of Orbs to control complex manufacturing processes, such as pharmaceutical manufacturing process or a process that is producing silicon or nano-technology products. Some type of “induction” whether electromagnetic, chemical or behavioral mechanism is needed that changes the internal state of the physical process due to an “experience of” the control mechanism.
Evocative sign systems (such as gesturing systems) can be used to provide a degree of precise control over human behavior. Of course, we do this all the time and every day.
The role of the theoretical system, the abstractions, is thus to bring the human to a state of understanding using sign systems recognized by humans, biological systems, and complex physical systems. We think about physical manufacturing processes as an example where Orb-based control can be more predictable and precise than in the case of natural language understanding. However, the Orb for a complex manufacturing system will still have points of ambiguity where causal “tipping points” exist and need to be explored in such a way that the tacit knowledge and anticipations of a human in the loop is instrumental to discovering the relationships that point to the cause of manufacturing failures. We believe that this can be demonstrated in a short period of time, say three months; given the technology we have and the understanding and associations between BCNGroup scholars that exists. Additional application to the Orb control of plant growth processes is suggested in a study of the application of general Framework theory (gF) to the measurement and control of fishponds.
So why are relational database not able to generate precise control over human interpretation of evocative sign systems (such as a set of Orb subject matter indicators using an Upper Controlled Taxonomy)? The answer is that the relational database has organizational structure that over-constrains the relationships between sub-structural elements (in this case words and patterns of co-occurrence of words). XML losses this pre-specified organizational structure to a certain extent, but XML does not give very fast access to the data. The concepts of machine representation of environmental affordance (from which our notion of computational abduction is derived) and of the set of invariances in data (from our notion of computational mutual induction) cannot be placed together into very fast action-perception cycles unless the data set is very small and the kinds of questions the human has are restricted.
The SLIP (Shallow Link analysis, Iterated scatter gather and Parcelation) technology, which Ontologystream Inc invented in 2002, fully demonstrates the quickness and the transparency of data encoding and interpretation into and from Orbs. We have developed many tutorials and the software is currently free and easy to use. Orb technology and SLIP technology can be integrated within 90 days and made available for weblog index creation that allows an Human-centric Information Production capability for the representation of real time social discourse. However, OntologyStream Inc is stuck without the minimal investment or a one-year project needed to take the next step. We are waiting for this economic problem to resolve and have proposed several technology collaborations to rapidly advance US intelligence production capabilities. (4/18/2004 8:09 AM). Following from this investment and new contracts, the BCNGroup has proposed a National Center of Excellence to be built in Northern Virginia.
One should see the future application domains as being similar to what one might have imagined at the beginning of the database era. When relational databases where being developed, in the 1980s, there was an expectation that the database would become like human experience of memory – except having perfect recall.
Relational databases require a prespecification of structure to the entire set of data. This is the so-called database schema. The Federal Enterprise Architecture (FEA) addresses some of the issues with respect to database schema interoperability, but the use of the relational data itself creates interoperable issues at root, and thus one needs to have web services such as reconciliation processes over ambiguation.
So, for example, one might wish the encode
< Jones, people living in a house, 13519 Leith >
along with many other facts about the Jones, and about other families. The relationship between people and “living in houses” is abstracted into a data model where a column of a relational database table is labeled address and the row of the database table is associated with people. “Jones” then becomes the entry of the cell of the column that is also the cell of a row. The relationship can also be called
< people, name, address >
The limitation on relational databases comes because the definition of rows and columns has to be done up front. This limitation can impose certain awkwardness when trying to get at the data fast. If the relational data model and the way in which the data needed to be address are aligned then things work well.
However, in exploratory work the human mind often will want to examine the memory of the past based on real time anticipation rather than on a structure developed based on a different set of anticipations. This restriction leads to phenomenon like “pre-mature closure” where an investigation comes to a certain conclusion not based on the full spectrum of anticipation by the investigator, but based on what can be manufactured easily.
Conjecture on System Reason
for Intelligence Failures
Comments to portal@ontologystream.com.