Knowledge Base Revision Through Exception-Driven Discovery and Learning

Abstract

We are currently witnessing a trend toward an architectural separation of a knowledge base (KB) into an ontology and a set of rules. The ontology is a description of the concepts and relationships from the application domain; the rules are problem solving procedures expressed with the terms from the ontology. Moreover, terminological standardization taking place in more and more domains has led to the development of domain ontologies. These two developments raise the prospect of reusing existing ontologies when building a new knowledge based system. For instance, the Disciple approach for building a knowledge based agent relies on importing ontologies from existing repositories of knowledge, and on teaching the agent how to perform various tasks, in a way that resembles how an expert would teach a human apprentice