Effective communication requires a common vocabulary. A knowledge graph is a combination of such a vocabulary and facts stated using that vocabulary. The vocabulary, or ontology, provides the terminology, concepts and relationships that are the foundation for knowledge graphs used in smart search, decision management, data integration, and a variety of other applications. Well-designed ontologies provide a declarative encoding of the meaning of vocabulary terms that are critical to enabling communication, among people and between machines. They are the starting point for automated extraction of content from documents using natural language processing (NLP) techniques, for seeding learning algorithms, for question answering and recommendation systems, and for understanding and leveraging data gathered from independently-developed systems and repositories.

More than 10 years ago, when we first developed this tutorial, there were few in our audience who had ever heard “the o word”. Today is a very different story, though. Because of Google’s Knowledge Graph and schema.org efforts, emerging applications in commercial and industrial IoT, sophisticated pharmacogenomics and related bioinformatics work, and increasing usage of semantics to address issues in finance such as “know your customer” (KYC), counterparty risk, and data governance more generally, our audiences are far more knowledgeable.

This tutorial provides an overview of the knowledge representation landscape and attempts to de-mystify some of the ‘black art’ of ontology development. We will outline basic methodology steps developed over time from a combination of:

• Business requirements analysis derived from best practices in business architecture capability and value stream analysis
• Domain analysis using business requirements adapted from software engineering and use case analysis more generally
• Conceptual modeling and terminology work
• Best practices from Semantic Web colleagues as well as our own experience in building large, operational systems

Examples ranging from question answering to supporting docents in a historic garden, to recommendation systems for targeted advertising and marketing, to automated support for better detection of offside offenses in soccer matches, to applications in finance and healthcare will be covered, with a focus on the Web Ontology Language (OWL). We will touch on representation patterns, including extensions for business rules systems that dramatically improve rule set quality, reduce error, and increase manageability / understanding of the rules. We will also briefly cover appropriate use of OWL and more expressive languages such as RuleML, to help potential users understand both the power and limitations the languages impose on applications. Finally, we will provide an update on current trends in standardization of ontologies and the related infrastructure to use them in enterprise systems.

This tutorial provides a great introduction for those who are just beginning to “get their feet wet” in the field and can be helpful in setting the stage for the rest of the conference.