Building a Knowledge Graph Framework with PoolParty for KGC

Matthias S.; Glenn Clatworthy; Bob L.; Larry S.; Dean A. This is an initial one-page draft of our conversation today answering the questions "What are we doing?" and "Why do we need PoolParty?" What are we doing? Summary (Glenn): We are building a framework for tracking the content of books on knowledge graphs covered in the KGC Book Club. We are trying to find a structure that would help the newcomer navigate relevant subjects (from RDF to OWL) to acquire practical skills relevant to their role. A significant barrier to the growth of the knowledge graph industry is the learning curve associated with mastering the semantic stack. Helping newcomers acquire skills relevant to their role would help scale adoption of semantic technologies. KGC has built a large community around knowledge graphs and needs an indexing system and other tools to help our community navigate an increasing number of resources. People (Matthias): We are a small, collaborative, geographically dispersed group (6-8 people) that recognizes we can’t build an ontology alone—we need a tool that enables collaboration. We have the distinct advantage of Dean A.’s expert advice. Our team is literate in semantic web domains and most of us have taken Semantic Web courses that introduce PoolParty functionality. Content (Bob):  We are starting with the first book featured in the KGC Book Club, Dean Allemang’s Semantic Web for the Working Ontologist. We recognize that using the intellectual property of the books is a potential barrier; but, the product we are building includes terminology, taxonomies, and an ontology that describes the structure of the books and point to the content of the book. We plan to include other books featured in the KGC Book Club including Michael Uschold’s Demystifying OWL for the Enterprise and Elisa Kendall and Deborah McGuinness’s book Ontology Engineering. With permission, we could also include The Knowledge Graph Cookbook. Why do we need Pool Party? Process (Matthias): We want to start small and grow. To do so, we need to get in touch with processes associated with ontology development, including automated methods for text mining, developing terminology (like card sorting), constructing consistent URIs, managing taxonomies, and defining patterns in ontologies. Technology (Glenn): PoolParty offers the most complete ensemble of tools. In a broad sense, this is a text analysis project: we are analyzing books, extracting terminology, building taxonomies, and evolving them into ontologies. PoolParty enables distributed development—many people working from different locations—and offers support for text analysis, taxonomies and ontologies: Text Analysis: NLP functionality would help us extract entities from the corpus recognition and build terminology on the front end. PoolParty offers support to automating the process of developing terminology. Taxonomy: Once entities are identified we can organize them hierarchically and build associated taxonomies that could support recommender systems to make search more efficient. Pool Party could support the taxonomy development process with card sorting functionality. Ontologies: We would evolve taxonomies into a single knowledge graph for the ontology associated with several different Knowledge Graph books and resources. We appreciate that PoolParty can output the ontology in RDF so used by ontology editors using different syntax. Goals (Larry). Our intent is to learn about knowledge graphs and to share meta-knowledge about knowledge graphs. This "knowledge graph about knowledge graphs" would have a variety of use cases--we would explore which would be helpful to the KGC community and the most marketable.