Seminar: From Ontologies to Knowledge Graphs: The Evolution of the Semantic Web Towards Explainable AI and Retrieval-Augmented Generation (RAG)

The seminar offers a brief overview of the evolution of semantic technologies, from the Semantic Web and ontologies as structured knowledge representation models, to Knowledge Graphs (KGs), which today serve as a bridge between semantics and artificial intelligence. We will explore how Knowledge Graphs, as a pragmatic evolution of ontologies, not only enable meaningful organization and linking of information, but also enhance generative AI applications through approaches such as Retrieval-Augmented Generation (RAG). By integrating structured sources (such as KGs) with language models, the quality, traceability, and explainability of AI-generated responses are significantly improved, leading to more reliable and auditable systems. The second part will present an ongoing case study: the digitization, semantic enrichment, and intelligent access to historical documents from the archives of Acción Católica Española. The project combines natural language processing techniques, semantic knowledge generation, and hybrid retrieval using knowledge graphs and vector databases. Additionally, as a future line of work, the transformation of extracted knowledge into interoperable ontological formats such as CIDOC-CRM is considered, to facilitate its publication and reuse within infrastructures like the European Cultural Heritage Cloud.

This seminar is aimed at students and professionals interested in the evolution of the Semantic Web, the current role of Knowledge Graphs in AI, and their real-world application, for instance, in projects focused on the preservation and digital transformation of documentary heritage.