Ort: Wilhelmshoeher Allee 73, R: 0315,

Beginn:16:15

Abstract: Ontology-Based Data Access (OBDA) is a paradigm for accessing and managing data that has recently become very popular for dealing with data sets that are incomplete, unstructured, and possibly stem from multiple sources (such as web data). The central idea of OBDA is to enrich such data sets with an ontology that provides a uniform and well-defined query vocabulary and links this vocabulary to the actual representation of the data. In this talk, I start with taking a look at the data complexity of OBDA query answering, which delineates feasible cases from infeasible ones and is closely related to practical techniques for implementing OBDA based on classical relational database systems. I then advocate a very fine-grained,non-uniform study of data-complexity and show that this approach gives rise to a previously unknown, but very tight connection between OBDA and constraint satisfaction problems (CSP). It turns out that this connection can also be used to derive interesting decidability and complexity results for relevant OBDA problems, and to clarify the descriptive complexity and expressive power of OBDA query languages.