Friday, October 22, 2004
[1] ß Discussion between Peter
Stephenson and Paul Prueitt
[3] ß Discussion on the Adi framework for knowledge processors
White Paper on Incident Information Orb
Architecture (IIOA) à
Adi Structural Ontology Part I à
Tom
A set of comments on Part II can be made over the phone, I have hand written notes on a printed copy of Part II, and we can discuss a few of the points and then perhaps you might wish to make a revision based on these discussions.
Part I seems correct except for the first two paragraphs of section 2:
"Adi observed that Arabic vowels and consonants indicated different
realizations (concrete applications) of certain abstract aspects of
"things." In a refinement
process, he chose a label for each abstract aspect he encountered. In the examples below, the abstract aspects
are in bold type and they are simply referred to as aspects.
"This is a big collection of examples (and it will grow as time allows) and the reader may not wish to read them all. The formal presentation of the ontology follows. More advanced examples will be discussed later on."
I would change these paragraphs and replace the two of them with the following:
"Adi observed that Arabic vowels and consonants indicated
different realizations (concrete applications) of certain abstract aspects of
"things." In a refinement
process, he chose a label for each abstract aspect he encountered. In the examples below, the abstract aspects
are in bold type and they are simply referred to as aspects. Abstract aspects delineate the contents of
frames within three layers of abstraction, (1) substructural, (2) superconcept,
and (3) topics+themes. Superconcepts
and topics+themes are to be discussed in Part II.
"To
illustrate and motivate substructural ontology, a collection of descriptions
are organized in a grouped correspondence to a banding of an 8x4 matrix holding
the set of 32 letters in Arabic. The
banding treats the first row as the top "band". The second, third and fourth rows are
treated as a grouped second band of the matrix. The fifth, sixth and seventh rows are treated as the third
band. The eight, and last, row of the
matrix is treated as the last band of the matrix. The banding following a decomposition of the number 8 (rows) into
1 + 3 + 3 + 1 (rows), which are the numbers of subsets of size 0, 1, 2, or 3 of
a universal set of size 3. The bands
partition the cells of the 8X4 matrix into groups having 4, 12, 12, 4 elements
in each group. Each group has a
categorical difference of some importance.
The formal presentation of the substructural ontology follows, where we
will establish a one-to-one correspondence between this 8x4 matrix and a 8X4
matrix representing what we conjecture to be a reasonable way to organize
categories of abstractions over the set of natural kind addressed in human
communication."
****
I will be watching foot ball and exercising to get the soreness to go away. Please call anytime today to talk about Part II.
We areincluding this note as a bead game communication for specific reasons. Ben Goertzel and Paul Werbos may be able to make a helpful observation about the structured ontologies that Adi discovered and that Prueitt and Adi are presenting in a formal system. Brain Welter may see in the three layers of abstraction the internal working of his very successfully deployed NdCore system for conceptual role-up.
Ballard or Sowa may choice to make a comment at this point.
Comments should be address to portal@ontologystream.com , and be considered a communication into the BCNGroup bead games.