This Research-Team is a follow-up of VERSO Research-Team
Project-Team Presentation
Joint project-team with the University of Paris-Sud 11 and CNRS (LRI), Located in Orsay.
Information availale online is more and more complex, distributed, heterogeneous, replicated, and changing. Web services, such a SOAP services, should also be viewed as information to be exploited.
The goal of this project-team is twofold : first of all to study the fundamental problems that are raised by modern information anf knowledge management systems, and secondly to determine novel solutions to solve these problems. Such systems will contain rich information and must be connected to networks. GEMO's main theme is the integration of information, seen as a general concept; more precisely, to discover meaningful information or services, understand their content or goal, integrate them, and finally monitor their evolution over time.
We would like to offer environments both powerful flexible to simplify the deployment of applications that give fast access to meaningful data. The creation of data warehouses and mediators offering a wide access to multiple heterogeneous sources provides a good means of achieving these goals.
Thes new problems combine Artificial Intelligence techniques (such as classification) and Database techniques (such as indexing).
GEMO is a project-team born from the merging of INRIA-Rocquencourt project-team VERSO, with members of the IASI group of Laboratoire de Recherche en Informatique (UMR 8623 CNRS-University Paris-Sud.
Research themes
XML data mediation.
We are interested in the integration of heterogeneous data, such as XML data. Our goal is to cluster elements of a collection of XML documents that are similar to each other into classes, and provide an accurate description of every class. A global schema (ontoligy) can be constructed, and should be used as an entry point when querying this XML collection.
- Mediation for Semantic Web.
Ideally, the Semantic Web would be a Web where the semantics of each data entity (web page, service...) would be understandable by human and machine alike : Agents, search engines and information servers should also to comprehend what each data entity deals with. Ontologies will play a central role in giving semantic values to each data entity. This open whole new perspectives regarding the improvement of the quality of results provided by search engines. More precisely, GEMO deals with three major problems that hinder the scalability of such systems : mediation between ontologies, mediation between sources (peer to peer context), and mediation between the Web ans its users.
- Thematic Web-Warehouses
We would like to develop a flexible and generic approach that would enable us to specify in a declarative way, for a given warehouse, the data we are interested in, simplify the acquisition of this data from the Web, and organize the data retrieved, bearingin mind we want to query the data later on.
- Broadening the use of Web Services.
By combining approaches such as warehousing and mediation, we are also drawn to investigate how to integrate Web services (that exchange XML data). One goal would be the discovery of services useful for a given application, and the understanding of their use. To this end, we focus our research on Active XML, a model based on XML documents that incorporate web service calls.
- Data model theory
We are also interested in the theorical aspects of the data oriented vision of computer science. We use tools such as logic and complexity to characterize calculus on collections (relations) or irregular graphs (Web).
International and industrial relations
Main international cooperations and industrials partners :
PICSEL project with France Télécom R&D.
European project DbGlobe on web query evaluation
RNTL E.dot project on warehouse on food riskwith, notably,INRA.
Industry : Xyleme start-up (issued from project-team).
Scientific leader
Serge ABITEBOUL
+33 1 72 92 59 32
serge.abiteboul@inria.fr
Secretary : +33 1 74 85 42 25
Team Address
Parc Club Orsay Université
ZAC des vignes
4, rue Jacques Monod - Bâtiment G
91893 ORSAY Cedex
