Using Ontologies to Evaluate Knowledge-based Systems

摘要

Many systems have hidden assumptions about their domain that must be rendered explicit if we are to identify the principles on which they are based. Verified ontologies provide one way of making these assumptions explicit. A verified ontology consists of a specification of a class of mathematical structures together with a proof of two fundamental properties: (1) Satisfiability: every structure in the class is a model of the ontology's axioms; (2) Axiomatizability: every model of the ontology's axioms is isomorphic to some structure in the class.