Reasoning on data semantics
we investigate different models and algorithms for querying data (or ressources) through possibly heterogeneous and distributed ontologies.
List of participants: F. Jouanot (Associate Professor), M.-Ch. Rousset (Professor), A. Termier (Associate Professor), G. Vargas-Solar (CR1), Ch. Collet (Professor), R. Tournaire (PhD 2007- ), S. Tandabany (PhD 2007-)
The web has deeply changed the vision of modern data management systems and has forced to revisit the problem of querying data which are distributed, possibly heterogeneous and ill or semi-structured. This revolution is going to get amplified with the miniaturization of storage devices connected to the network. This opens new possibilities and raises new challenges for integrating heterogeneous decentralized and context-sensitive data. Reasoning on context and data semantics is one of the keys for attacking in a principled way those challenges. The positioning of the group is to investigate the different algorithmic issues for the scalability of querying data through possibly heterogeneous and distributed ontologies.
We plan to extend our work on data semantics in two main directions:
- Models and algorithms for reasoning on the distributed semantics of Web data. In particular, we will investigate models for handling uncertainty and trust in peer-to-peer data management systems (Dataring project), and algorithms for automatic discovery of probabilistic mappings between taxonomies of classes.
- Models and algorithms for handling semantic and contextual descriptions of devices or services. In particular, we will investigate the problem of automatic discovery and composition of services based on the semantic description of their functionalities and of the context in which the devices supporting them are deployed (CONTINUUM project).