Title

Topics in OWL Reasoning for BFO-based Ontologies

Workshop type

Tutorial

Organizer

Alan Ruttenberg, University at Buffalo, NY

Reasoning can bring a number of benefits to ontology application and development, among them

• Increased clarity of the scope of a term,
• A level of automated quality control,
• Enhanced expressiveness of queries, 
• Discovery and characterization of ontology content
• Integration across ontology structured data sources
• Approaches to debugging ontologies

This tutorial teaches a number of approaches for axiomatizing terms, using real-world cases from biology and medicine for the most part. During the first section we review the basics of how to see the effect of axioms, giving examples that exploit OWL reasoning, SPARQL query, and Triple Store use. 

In the second section we expose a variety of different kinds of reasoning problems that can be solved by exploiting the capabilities of OWL reasoners to find inconsistency, and determine subsumption relations.  We examine cases where reasoning is already effectively used, as well as examples where it is less so. As a practical example to show the benefits of reasoning we take one or two under-axiomatized ontologies, add axioms, and explore the consequences.
 
In the third section we proceed to a selection of advanced topics, exploring less commonly used OWL constructs that can be beneficial in ontology construction, for example how to encode local reflexivity, how to reason with individuals, use of rules, how to determine effective domain and range of properties, approaches to modeling temporal phenomena, and use of ontology transformation as an aid for debugging.  Participants who have special topics they wish covered may request that they be addressed with sufficient notice before the tutorial.

This is an intermediate- to advanced-level course. Participants will ideally have a working knowledge of OWL, for instance having using Protégé and attempted to axiomatize terms, and will have run a reasoner and thought about the results. Experience with a scripting language will be beneficial as some examples involve programmatic development of axioms and queries.

Case studies will be taken from existing BFO-based and OBO Library ontologies and materials from the tutorials will be available. The instructor will be available for a period after the tutorial for attendees who wish to discuss topics in more detail or about how to apply them to their own ontology work.

Few, if any, courses that teach the application of more advanced OWL to biomedical ontology development and application. As a consequence, too few people are in a position to bring the full power of logical reasoning to their ontology work, or even have a working knowledge of what the potential benefits are. When it comes time to defend our work (for instance at grant renewal time) they have a harder time than necessary explaining what the benefit of doing formal ontology work is.

Courses taught by computer scientists often have uninteresting or misinformed use cases, and courses for biomedical specialists often struggle with introducing basic OWL, never mind more complicated constructs. We need to remedy this by developing material that teaches this area and engaging ontologists in the learning and use of it.

The materials presented in this tutorial will be taken from a number of years of personal experience being at the leading edge of development of OWL, of BFO, and of OBO Foundry ontologies. I have familiarity with and have collaborated on many ontologies, bringing most of the methods I will describe to that work.

This is a good time to offer such a tutorial, I think. I have observed that quite a few ontologies have mastered the basic use of OWL and some have gone further. Simpler reasoning-based tools such as Term Genie have more adoption and might bring in participants who have been intrigued by it’s utility. OWL tools and toolkits are reasonably mature – they were not when the first ICBO conference took place.

I think this tutorial needs a good 4 hours to present, however if the schedule permits it would be good to have a hands-on section without having to worry about the pace. This could be offered by making this a 6 hour tutorial, with the 3rd slot of 2 hours being a "clinic”, or by offering to meet with participants in a room other than where workshops and tutorials are taking place.  I leave this to the discretion of the the conference organizers.