Debugging and Repair of Owl Ontologies Debugging and Repair of Owl Ontologies

Title of dissertation: DEBUGGING AND REPAIR OF OWL ONTOLOGIES Aditya Kalyanpur, Doctor of Philosophy, 2006 Dissertation directed by: Professor James Hendler Department of Computer Science With the advent of Semantic Web languages such as OWL (Web Ontology Language), the expressive Description Logic SHOIN is exposed to a wider audience of ontology users and developers. As an increasingly large number of OWL ontologies become available on the Semantic Web and the descriptions in the ontologies become more complicated, finding the cause of errors becomes an extremely hard task even for experts. The problem is worse for newcomers to OWL who have little or no experience with DL-based knowledge representation. Existing ontology development environments, in conjunction with a reasoner, provide some limited debugging support, however this is restricted to merely reporting errors in the ontology, whereas bug diagnosis and resolution is usually left to the user. In this thesis, I present a complete end-to-end framework for explaining, pinpointing and repairing semantic defects in OWL-DL ontologies (or in other words, a SHOIN knowledge base). Semantic defects are logical contradictions that manifest as either inconsistent ontologies or unsatisfiable concepts. Where possible, I show extensions to handle related defects such as unsatisfiable roles, unintended entailments and nonentailments, or defects in OWL ontologies that fall outside the DL scope (OWL-Full). The main contributions of the thesis include: • Definition of three novel OWL-DL debugging/repair services: Axiom Pinpointing, Root Error Pinpointing and Ontology Repair. This includes formalizing the notion of precise justifications for arbitrary OWL entailments (used to identify the cause of the error), root/derived unsatisfiable concepts (used to prune the error space) and semantic/syntactic relevance of axioms (used to rank erroneous axioms). • Design and Analysis of decision procedures (both glass-box or reasoner dependent, and black-box or reasoner independent) for implementing the services • Performance and Usability evaluation of the services on realistic OWL-DL ontologies, which demonstrate it’s practical use and significance for OWL ontology modelers and users DEBUGGING AND REPAIR