Tutorial: OWL Distilled

When: October 11th/12th, 2015
Where: Bethlehem, Pennsylvania, US - Co-located with the 14th International Semantic Web Conference (ISWC 2015)

Overview | Objectives | Motivation | What's unique? | Program | Audience | Presenters

Overview

If you think that getting your (or someone else’s) head around OWL is challenging, you have found the right tutorial. There are just a few foundational elements that everything else is built from:

That’s pretty much it. In this tutorial we will describe the many ways that these things can be combined and used. Most importantly, there are triples that assert relationships between things -- e.g. JaneDoe worksFor Microsoft. There is inference to generate new triples and a few more key things, but not as much as you think. The topics we will cover include: 

If you are getting started in OWL, this tutorial has everything you need and nothing you don't.

Objectives

After this tutorial, students will be familiar with the 30% of OWL that gets used 90% of the time. More specifically, they will: 

Motivation

Building, understanding and using ontologies is growing in importance. Being the only well supported standard, OWL is the usual choice for companies and government organizations. However, the OWL language is notoriously difficult to learn, especially for those with a limited technical background. Confusion arises from a number of places. The language is very complex with many syntaxes, languages, sublanguages and constructs. Much of this hardly ever gets used and is expressed using unfamiliar and often non-intuitive technical terms. 

Some of the commonly used constructs are also non-intuitive. The best example of this is owl:Restriction, which is arguably the most important way to define meaning using OWL. To encode a simple rule that means: “every patient visit has at least one care provider” requires you to 1) create a class for patient visit and then to 2) create a ‘Restriction’ that means: “anything that has a care provider” and 3) to say that the first class is a subclass of the second class. The second class (the restriction) often does not have a name, which leads to blank nodes and blank stares from those aspiring to master OWL. 

What makes this tutorial unique?

Gentle introduction to OWL: with minimal technical experience required. We start out using common everyday language to introduce the central modeling constructs, and later introduce the formal terminology.

Focus on what matters most: We focus on the 30% of OWL that gets used 90% of the time.

Relevance First: We present material in an order that makes sense to the participant who wants to build ontologies in subject matter that they care about. Every construct is motivated by a real world usage before it is introduced, ensuring that the material is relevant and memorable.

Real World Examples: Every example comes from one or more commercial ontologies that we have built in a variety of industries. There are no toy problems. 

Socratic Style: There is a great deal of audience participation, which deepens learning. You will have a chance to start modeling a domain or your choosing.

Program Outline

Part 1:  Getting Started: what do we need to say?

We briefly introduce the idea of an ontology as a model of some subject matter that you care about, and that OWL is the language for expressing the model.  Through plentiful class participation and real-world examples, we identity the main things that one has to say when describing subject matter as an ontology. This includes:

 Part 2:  How do we say it in OWL?

We transition from informal everyday terminology to the more precise technical language of OWL.  For example, a ‘kind’ is an owl:Class; ‘a kind of’ is rdfs:subClassOf; a way two things can be related to each other is an owl:ObjectProperty. To “say something” is to assert an axiom, or a triple. Drawing conclusions is called ‘inference’.

Specific topics covered are listed below, approximately in order or presentation. The actual order is dictated by a natural progression of building a simple ontology in healthcare.

Part 3:  Patterns and Pitfalls

We go over some common patterns for building ontologies in OWL as well as highlight a few common pitfalls. We introduce some naming conventions and common patterns for creating definitions from more primitive concepts. We describe a pattern for catching many errors.

We discuss how easy it is to read too much into  text labels and comments. We explain how domain and range are often misunderstood and misused. Finally we mention the open world, arguable the most challenging part of OWL, but do not discuss at any length.

Part 4: Summary and Conclusion

This is OWL distilled. We introduced 41 of the 142 constructs in the OWL 2 Reference Guide that comprises the 29% of OWL that you use 90% of the time.  We summarize an additional 11% of OWL that is commonly used and give a brief overview of the 60% of OWL that are advanced topics or for other reasons, infrequently used.

Audience & Prerequisites

This tutorial is suitable for complete beginners, there are no prerequisites.  The intended audience includes anyone who is interested in a gentle introduction to OWL, or who wants to be able to better explain OWL to others.  This might include:

Presenters

Michael Uschold and Dave McComb have extensive experience building enterprise ontologies in OWL and teaching how to do so.  In 2015 alone, customized versions the material in this tutorial has been presented in-house to Disney, Wells Fargo and Sentara Healthcare.

Michael Uschold, Semantic Arts

Michael has over two decades experience in developing and transitioning semantic technology from academia to industry. He pioneered the field of ontology engineering, co-authoring the first paper and giving the first tutorial on the topic in 1995 (in London). This leveraged the work he did in creating the influential "Enterprise Ontology". From October 2010, he has been working as a senior ontology consultant at Semantic Arts, training and guiding clients to better understand and leverage semantic technology.  He has built commercial enterprise ontologies in finance, healthcare, legal research, electrical power, consumer products, manufacturing and corporation registration for state government.  He has given numerous tutorials and training classes. He received his Ph.D. in AI from Edinburgh University in 1991 and an MSc. from Rutgers University in Computer Science in 1982.

Dave McComb, Semantic Arts

Dave is a hands-on practitioner and thought leader in the area of applying Semantic Technology to Enterprise Architecture and Applications. For fourteen years as co-founder and President of Semantic Arts he has managed major Semantic Technology projects with over a dozen large enterprises, including Goldman Sachs, Broadridge Financial Systems, Procter & Gamble, Lexis Nexis, Sentara Healthcare, Sallie Mae, and seven different Agencies in the States of Colorado, Texas and Washington.  

Dave is the author of the book, Semantics in Business Systems, and was the co-founder of the Semantic Technology Conference, the go to place for companies looking to commercialize semantics.  He is a frequent speaker and writer on the topic and has inspired many to enter the field. In the early 90’s Dave pioneered an approach to ontology development based on facilitated brainstorming in focused Semantic Modeling sessions.  In the intervening 20 years he has lead over 200 of these sessions and built almost as many ontologies. 

Prior to founding Semantic Arts, Dave spent 13 years with Andersen Consulting (the part that became Accenture) designing and building Enterprise Applications for large firms including: Boise Cascade, Georgia Pacific, Wildish Construction, Norton Abrasives, the US Geological Survey, Bougainville Copper, US West and Martin Marietta (now Lockheed Martin). He founded First Principles and co-founded Velocity Healthcare.

Are you interested in an occasional email about semantics, enterprise ontology and enterprise information systems?
Sign up for the Semantic Arts newsletter.