|
Knowledge Representation Foundations
New Representation
As languages have proved ineffective to adequately represent knowledge, as presented
in"The Main Language Fallacy", and since we do need to model, manage, and share knowledge and its architecture,
with the added support of advanced computing systems, we need an effective representation
approach.
Natural Structure
As language structure, grammar, and syntax are inadequate to represent knowledge,
even if they can be very practical to infer references between communicating parties
with common related knowledge contexts, the knowledge and knowledge architecture representation
approaches must be based on and derived from the natural structure of knowledge, and
not the structure of languages.
Representation Approaches
But, at the same time, representational structures are required, as well as corresponding
notation elements. Symbols, qualification, and relations, for example, need to be
defined and expressed, as they are required for languages. The main differences are
in the structure and architecture approaches.
Qualities?
Qualities are the fundamental component of knowledge, as presented in"Qualities Map Directly To Neural Patterns", but what are qualities and how can they be expressed and represented in both human
and computer accessible ways?
Everything?
How can everything that can be known be effectively expressed and represented in both
human and computer accessible ways?
Relations
How can everything that can be known be effectively related, interrelated, and correlated
in both human and computer accessible ways?
Example
Conceptually, it is in fact relatively simple, especially as it is natural and intuitive
like the natural phenomenon of knowledge is. At the core, as everything else are applications
of these components, it can be expressed in three basic statements:
- Qualities range from simple to compound, as they can be nested and combined to any
degree or level. The most fundamental qualities are, in order, identification and
relation, where classification is also a basic relation application.
- Everything we can know are identifiable quality collections, which we naturally call
resources and which encompass and map all concepts, beings, things, systems, environments,
processes, feelings, behaviors, phenomena, and relations.
- Relations are resources like any other, but they also have references to the subject
(e.g. source) and object (e.g. destination) resources that they relate and for which
they qualify each relationship.
Diagram
The following conceptual derivation diagram uses standard UML class diagram notation
to illustrate the required core knowledge architecture representation foundations,
along with the notation components at their base, as well as along with a few basic
applicative extensions, like events, processes, and adaptive cases, for example.
Right Side Up
Note that, possibly contrary to the traditional Object-Oriented approach of putting
the more general structures on top and the more specialized on bottom, this diagram
puts specialization on top and the generalized foundations at the bottom. Accordingly,
the traditional white inheritance arrows are now pointing down to the more general
"ancestor" structures. The main reason for doing so, is that it seems natural that
foundations sit below the structures that they support, and because the applications,
modeling, managing, and sharing knowledge, are the main interest and focus.
Knowledge Architecture Foundations, conceptual derivation UML diagram
Better Knowledge Representation Foundations
There is much more to knowledge and knowledge architecture representation, and much
more still to their modeling, management, and sharing, but already, we have defined
much more appropriate foundations than language structures, for the tasks at hand.
Continue to
|
