After obtaining a PhD in computer science in 2001, I was recruited as a full time INRIA researcher in the EXMO team of INRIA Grenoble Rhône-Alpes. I moved to Montpellier in 2007 to prepare the proposal of the INRIA GraphIK project team, created in january 2010.
During my PhD, I began working on a knowledge representation formalism known as Conceptual Graphs. Syntactically, these objects are graphs (a mathematical structure), semantically, they correspond to first-order logic formulae.
I consider now graphs as a pivotal structure to study the properties of different knowledge representation formalisms, be it logics, databases, or semantic web languages such as RDFS or OWL. An important problem in that direction is Ontological Conjunctive Query Answering. Generally, this problem can be stated as follows: given a knowledge base K that contains a database F (facts) and an ontology O, does K contain an answer to the conjunctive query Q? In its first-order logic version, we can consider both F and Q as the existential closure of conjunctions of positive atoms, and O as a set of first-order logic formulas that can be expressed in a given ontology language. The problem is then restated as: is Q logically deducible from F and O?
My research has focused on using rules to express the ontology. The KR language obtained (it can be seen as Datalog+ in databases) is then so expressive that the deduction problem becomes undecidable. Finding efficient algorithms for decidable subclasses of the problem becomes now a crucial task, more so now that the applications we aim at will involve large and possibly distributed knowledge bases.
Address: LIRMM, 161 rue ADA, F34392 Montpellier Cedex 5, Montpellier, France
Phone number: (33)(0) 67 41 85 31
Fax number: (33)(0) 67 41 85 00