Debugging and Repair Support for Ontology Developers

Ontologies are formal frameworks to organise information and are widely used to provide structured representation of large knowledge bases in fields such as artificial intelligence, semantic web, biomedical informatics, systems engineering and many more. Ontologies are used to define the meaning of concepts and describe properties and interrelationships between concepts. They provide compact representations of the knowledge about a domain, from which information can be inferred using reasoning. Given the nature of ontologies, ontology development is an exceedingly difficult and error-prone task.

The aim of this project is to support ontology engineers and users with new ways of debugging and analysing often very large ontologies. Concretely, the aim will be to develop abduction methods for expressive description logics in order to debug and repair ontologies. Abduction is the task of, given a set of observations and a knowledge base of background knowledge, finding a rational explanation for these observations. The project is expected to build on work conducted in the group for computing uniform interpolants of ontologies. Uniform interpolation and abduction are related problems.

Depending on the students interests the emphasis of the project could be on foundations and theory, it could be more user and application-focussed involving the exploration of particular use cases and applications, the focus could also be more programming intensive involving the development of an ontology debugging and repair toolkit for integration with existing APIs and ontology editors.

This project would suit a student interested in ontology-based knowledge processing and AI.