Linked Data and AI Part 2: From Linked Data to Knowledge Graphs

This course details practical approaches for evolving linked data into knowledge graph systems, with specific attention to how generative AI can be strategically adapted within these infrastructures. We will examine the defining characteristics of platforms that support persistence, querying, and inference of graph data, including the Library of Congress Linked Data Service, the Share‑VDE knowledge base (BIBFRAME data), and the Wikidata knowledge graph.

Case studies will focus on two applied areas: 1) semantic retrieval from Wikidata and the Share‑VDE knowledge base with generative AI, and 2) the modeling and populating of a knowledge graph using BIBFRAME records contributed to a Wikibase via an enhanced version of Marva Quartz (marva‑vibe). This enhancement enables chat‑based linked data creation, illustrating experimental approaches such as “vibe cataloging,” which signal future directions in knowledge graph population.

Through lectures and case studies, participants will gain practical understanding of how knowledge graphs add value to metadata by providing interconnected networks that are inference capable. Students will learn how to adapt generative AI models to extend bibliographic data into knowledge graph systems, while gaining awareness of emerging experimental methods that may shape future linked data practice.