Tutorial
"From Data to Knowledge: Creating, Validating, and Learning from Knowledge Graphs"
Abstract: Knowledge Graphs (KGs) provide a flexible way to integrate heterogeneous data with explicit semantics, but building a KG is only the first step toward reliable, actionable knowledge. KGs must be systematically created, curated, validated, and analyzed to ensure that their content satisfies expected structural and semantic requirements and supports knowledge discovery. This tutorial presents an end-to-end, hands-on pipeline for transforming heterogeneous data into validated KGs and discovering symbolic patterns from their content.
Participants learn how to model and integrate data as an RDF Knowledge Graph, identify and curate data-quality problems, and express structural and semantic constraints using the Shapes Constraint Language (SHACL). SHACL validation detects violations, inspects their causes, and guides KG curation. The tutorial then moves from knowledge validation to knowledge discovery by introducing symbolic learning techniques for extracting interpretable relational patterns and rules from the validated KG. Participants analyze the discovered patterns based on their evidence and meaning, and use them as candidates for further enrichment and validation.
Through practical exercises, participants experience an iterative KG lifecycle in which they create, curate, validate, analyze, and enrich knowledge. By the end of the tutorial, participants understand not only how to construct a KG, but also how to assess its quality and use its symbolic structure to discover explicit, interpretable knowledge.
Biography
Maria-Esther Vidal is Professor of Computer Science at Leibniz University Hannover and leads the Scientific Data Management group at TIB – Leibniz Information Centre for Science and Technology. Her research focuses on semantic data management, Knowledge Graphs, and hybrid AI, with particular emphasis on integrating symbolic knowledge and machine learning to enable transparent and interpretable AI systems.
She has extensive experience in Knowledge Graph creation, integration, curation, validation, and analytics, as well as in semantic technologies and neuro-symbolic approaches for knowledge discovery. Her research has spanned scientific and industrial domains, including medicine, environmental sciences, and research data management. She has authored more than 250 scientific publications and has coordinated and contributed to numerous national and European research projects. She has also supervised more than 30 doctoral researchers and has extensive experience in teaching and training students and researchers in databases, semantic data management, Knowledge Graphs, and AI.