NASA Jet Propulsion Laboratory talks how the organization built and uses an institutional knowledge graph at KMWorld Connect 2021

JPL’s IT landscape has evolved organically during its decades-long lifetime, spawning a proliferation of applications and data systems that support JPL’s missions.

These applications and systems often duplicate employee information, roles, project names, organization structure, and other institutional information, resulting in redundant, possibly inaccurate, data.

With institutional information residing in silos, concepts are often unrelated. Relationships between structured and unstructured data, as well as answers to “Who did what when?” and “Who has experience?” require time-consuming, manual investigation. Even the simple questions, such as, “Which projects are currently active?” or, “Where is the Robotics Lab?” are not readily accessible.

Ann Bernath, software systems engineer, NASA Jet Propulsion Laboratory (JPL); Bess Schrader Speaker, senior consultant, Enterprise Knowledge; and Daria Topousis, content & search services project manager, NASA Jet Propulsion Lab (JPL) discussed its vision; process for data inclusion, modeling, and stakeholder involvement; current status; success stories; and future road map during KMWorld Connect 2021.

“We’ve had many applications spring up over time holding institutional data,” Bernath said. “All that data is siloed. It makes it difficult to find information.”

Over the past several years, JPL has invested in semantic technologies in the form of a taxonomy-driven community-of-practice application, semantic proof-of-concepts, and deployment of an enterprise semantic platform (ESP).

“Our desired state is ‘one-stop shopping’ where data is in one place,” Bernath said.

Leveraging the ESP’s semantic tool suite, its small team delivered an institutional knowledge graph (IKG), a centrally maintained, ever-evolving knowledge graph that identifies and describes JPL’s common concepts and the relationships between them, providing a single source of discovery for users and applications.

The IKG is based on standards, models, and conventions established by a small action team of IT and JPL Library representatives and is reviewed by an advisory committee made up of a broader list of enterprise-wide stakeholders.

According to Topousis, JPL needed project team leaders focused on delivering and meeting a schedule. Balancing the vision with the day to day execution is crucial.

“We wanted to bring people who are IT people and information management experts together,” Topousis said.

The IKG is built on top of taxonomy and ontology tool and a data orchestration piplene tool, and RDF Triple Store to deliver results and answers, Schrader said. Making the IKG included:

• Prioritizing data needs
• Identifying institutional gold sources
• Develop and ontology
• Transform data
• Validate data
• Provide access

The approach is continuous improvement via data inclusion and outreach, Topousis said. Users can utilize a visual search interface, Schrader noted.

“We’re getting some traction and reaching our desired state little by little,” Bernath said.

In the future JPL would like to incorporate semantics into enterprise-wide search, add IKG knowledge panels, and linking employees or specialists with their expertise and skills, Schrader said.

KMWorld Connect 2021 is going on this week, November 15-18, with workshops on Friday, November 19. On-demand replays of sessions will be available for a limited time to registered attendees and many presenters are also making their slide decks available through the conference portal. For more information, go to www.kmworld.com/conference/2021.