Date : July 7, 2022, 10 a.m. - Room :Salle du conseil

Efficient reasoning in heterogeneous data integration systems


The proliferation of digital data sources in many domains brings a new urgency to the need for techniques which allow to flexibly and efficiently query heterogeneous data sources (relational, JSON, key-values, graphs etc.). We follow the Ontology-Based Data Access (OBDA) paradigm where the knowledge coming from the data sources is integrated using an ontology. First, we propose a new integration architecture where the data sources are integrated into a RDF graph using an RDFS ontology with an expressive mappings language (GLAV mappings). We define several query answering strategies on this architecture that are implemented into our system called Obi-Wan. We show that our new strategy which compiles the ontological reasoning into the mappings improves the performance. Second, we study the more general setting, where the ontology is defined as a set of existential rules. We give a characterization of the rules and the mappings, for which the reasoning enabled by the rules can be compiled into the mappings. Bio: Maxime Buron is a postdoc at Inria Sophia Antopolis in the GraphIK team. He will join the University Clermont Auvergne as an associate professor starting in September 2022. He completed his PhD supervised by François Goasdoué, Ioana Monalescu and Marie-Laure Mugnier at Inria Saclay in 2020, then he worked one year with Michael Benedikt at the University of Oxford as an research associate. His research work relies on automated reasoning, databases and knowledge representation in particular in the context of Ontology-Based Data Access.