Knowledge graphs are a valuable tool that organizations can use to manage the vast amounts of data they collect, store, and analyze.
At Data Summit 2022, Joseph Hilger, COO, Enterprise Knowledge LLC and Sara Nash, senior consultant, data and information management, Enterprise Knowledge, LLC presented an “Introduction to Knowledge Graphs” during their workshop session.
The annual Data Summit conference returned in-person to Boston, May 17-18, 2022, with pre-conference workshops on May 16.
Enterprise knowledge graphs’ representation of an organization’s content and data creates a model that integrates structured and unstructured data. Knowledge graphs have semantic and intelligent qualities to make them “smart.”
“We have information overload,” Hilger said. “We’re beyond the days where we know everything we’re working on.”
Knowledge graphs are the foundation for AI tools such as Apple’s assistant Siri, Amazon’s Alexa, Netflix’s user interface, and more, Hilger explained.
Knowledge graphs allow for explainable AI. These graphs store information in a machine readable way that humans can still understand. Knowledge graphs can aggregate information from disparate solutions and provide context for data that lies within these multiple systems, he said.
Taxonomy is being able to associate synonyms and form hierarchies of information that is easily searchable. This is one of the main pillars of a mature knowledge graph solution. Ontologies can connect the formed relationships between these taxonomies, Nash said.
“Taxonomy is the backbone of search,” Nash said. “Data should be considered something that is consistent regardless of the way it’s represented.”
A knowledge graph is comprised of a graph database to store the information, then there’s a data orchestration system, and then taxonomy/ontology management, Hilger explained.
Knowledge graphs can identify similar things based on an ontology model, Nash noted. Google has been a main driver of this type of knowledge graph.
Preferred use cases for a graph model include highly interrelated data, sparse data, or when a flexible schema is required, Nash said.
“Graph says, ‘Bring it on, this is exactly what I want to do,’” Hilger said. “To get it to really work, you need to have consistency.”
They presented a few examples of use cases that Enterprise Knowledge had worked on where Hilger and Nash constructed a knowledge graph solution, including companies that had inconsistent data throughout the organization and organizations that wanted to create personalized content offerings.
Many Data Summit 2022 presentations are available for review at https://www.dbta.com/DataSummit/2022/Presentations.aspx.