13 August 2018: Juan Casanova

 

Title: Using abstract knowledge to automatically detect faults in ontologies.

 

Abstract:

 

Developing and using ontologies in logic-based formalisms (such as OWL) has similar issues to those well known in software engineering: errors, imprecisions in requirement capture and bad design (in general, "faults") can occur and cause serious problems. During my master's by research here at Edinburgh, I developed a proof of concept of a framework for automatically detecting faults in ontologies by formally encoding abstract knowledge, which I called "Meta-ontology fault detection". In the first year of my PhD I have tackled performance and other problems in it.

 

Meta-resolution and meta-unification are the names of the algorithms developed for this purpose, together forming an automated theorem proving algorithm using a least commitment approach for the instantiation of meta-variables to offer a more computationally attractive approach to the problem.

 

I will talk about all of this, and about the steps to follow in the remainder of my PhD.