An Introduction to Knowledge Computing

Authors: Vagan, Terziyan; Oleksandr, Shevchenko; Mariia, Golovianko
Year: 2014
Venue: Eastern-European Journal of Enterprise Technologies, Vol. 1, No. 2 (67), pp. 27-40.
Product of the Action: Yes

Keystone Members Authors:

This paper deals with the challenges related to self-management and evolution of massive knowledge collections. We can assume that a self-managed knowledge graph needs a kind of a hybrid of: an explicit declarative self-knowledge (as knowledge about own properties and capabilities) and an explicit procedural self-knowledge (as knowledge on how to utilize own properties and the capabilities for the self-management). We offer an extension to a traditional RDF model of describing knowledge graphs according to the Semantic Web standards so that it will also allow to a knowledge entity to autonomously perform or query from remote services different computational executions needed. We also introduce the concepts of executable knowledge and knowledge computing on the basis of adding an executable property to traditionally used (datatype and object) properties within the RDF model. The knowledge represented with such an extended model we call as an executable knowledge, or the one which contains explicit (executable) instructions on how to manage itself. The appropriate process of the executable knowledge (self-) management we call as a Knowledge Computing. Unlike the knowledge answering machines, where computations over knowledge are used just for addressing a user query, the knowledge computing in addition provides computations for various self-management purposes. The paper also presents some pilot (proof-of-concept) implementation of the executable knowledge as a plug-in to Protégé ontology development environment. P.S. Among other objectives the paper aims at content self-management where RDF model for content structuring is extended with the executable properties; and, e.g., search in such systems would mean content execution (i.e., search outcomes are rather being computed than being found).