The MATHESIS meta-knowledge engineering framework: Ontology-driven development of intelligent tutoring systems

Abstract

The effect of the knowledge acquisition bottleneck is still limiting the widespread use of knowledge-based systems (KBS), especially in the area of model-tracing tutors, as they demand the development of deep domain expertise, tutoring and student models. The MATHESIS meta-knowledge engineering framework for model-tracing tutors, presented in this article, aims at maximizing knowledge reuse. This is achieved through ontological representation of both the declarative and procedural knowledge of a KBS (model-tracing tutor), as well as of the declarative and procedural authoring knowledge of the process to develop a KBS. Declarative knowledge is represented in Ontology Web Language (OWL). Procedural knowledge is represented using the concepts of atomic and composite processes of OWL-S web services description ontology. The framework provides knowledge engineering tools, integrated into the Protégé OWL ontology editor, for the development and management of the KBS's ontological representation. It also provides meta-knowledge engineering tools for the ontological representation of the knowledge engineering expertise as a set of composite knowledge engineering processes and atomic knowledge engineering statements. The latter constitute a language, ONTOMATH, for building executable knowledge engineering models that, when executed by the tools, guide non-expert knowledge engineers like domain experts to the creation of new knowledge-based systems (model-tracing tutors). The framework, being in an experimental stage, was used for the development of a monomial multiplication and division tutor. However, the overall design and implementation aimed at constituting the framework as a proof-of-concept system that can be used for the meta-knowledge engineering of more complex model-tracing tutors.