Dean Allemang, data.world, principal solution architect

Knowledge Graphs are integral components in enterprise data management and foundational elements of reliable AI-based systems. But the question remains - how do we construct a knowledge base and populate it with our information?

For several years, an emerging approach involves initiating the process with documentation encompassing descriptions of the organization's knowledge and proceeding with "entity extraction". This strategy involves identifying the entities mentioned in the documents and discerning their interconnections. An important nuance to entity extraction is the role of metadata; how do we correspond our entities with an established structure (ontology) within a specific domain?

Recent developments in AI are enhancing this method for creating Knowledge Graphs by offering an effective method to assess large volumes of text.

This tutorial invites attendees on an exploration of using an ontology to facilitate entity extraction. Participants will then assess the derived knowledge graph, adapt the ontology for an even more profound extraction process. Attendees will be able to witness the evolution of their knowledge graph through a data catalogue, an interactive platform that allows for tracking, managing, and integrating with other data assets.

By the conclusion of the session, each participant or group will have created a unique knowledge graph tailored to their interpretation of the domain.