Program & Location

Detailed Program

Wednesday, July 18

Keynote Session, 09:00 - 10:00 - Room 1

• 09:00 - 11:00: Magdalena Ortiz, Is ontology-mediated graph querying finally within reach?

Coffee Break, 10:00 - 11:00

Session 1, Temporal and Ontological Foundations for Semantic Data Management

• 11:00 - 11:30: Kajsa Pedersen, Greta Adamo, Roman Kontchakov, Davide Lanti, Andrey Rivkin, Towards Monitoring of Patients with Bipolar Disorder

• 11:30 - 12:00: Julien Corman, Ognjen Savkovic, Temporal Path Queries: Challenges and Open Questions

Lunch Break, 12:00-14:00

Session 2: Semantic Technologies for Data-centric AI and Knowledge Graph Construction

• 14:00 - 14:30: João P. Lima, Renata Wassermann, FORGE: Framework for Ontology-grounded Retrieval and Graph Extraction 

• 14:30 - 15:00: Jose Parente de Oliveira, Antonio Silveira Dantas, Using Knowledge Graphs to Restrict the Context of LLM Analysis: A GraphRAG Approach for Personnel Management

• 15:00 - 15:30: Kamal Hamaz, Marco Di Panfilo, Davide Lanti, Diego Calvanese, AgenticVKG : Towards Memory-Driven Agentic AI for VKG Mapping Construction

Invited Talk

Prof. Magdalena ortiz

Is ontology-mediated graph querying finally within reach?

Ontology-based data access (OBDA) has been around for almost two decades as a successful paradigm for querying graph-structured data. Yet, we still lack ontology-mediated graph queries with their distinctive path-navigation capabilities. Although the complexity of ontology-mediated navigational queries has long been understood, no implementations have been attempted, and existing systems remain restricted to conjunctive (relational) queries that are rewritten into SQL or SQL-like fragments of SPARQL. In this talk, we argue that, thanks to the recent standardisation of GQL and SQL/PGQ, long-standing barriers to ontology-mediated graph queries have been lifted, and we have new paths for addressing other challenges, such as extending pure rewriting techniques beyond DL-Lite, a topic that has received surprisingly little attention.  We present recent advances towards identifying expressive GQL fragments that subsume conjunctive and regular path queries while admitting pure rewritings.  On the ontology side, we extend pure rewritability to Horn logics like ELHI, suitably restricted to preserve NL data complexity. 

Location