Section: Software

Keywords : Information Retrieval, XML, Semantic Web, RDF, RDFS, OWL, Conceptual Graph, ontology.

CORESE

Participants : Olivier Corby [ correspondant ] , Olivier Savoie.

Description.

Corese (COnceptual REsource Search Engine) is an RDF(S)-dedicated engine based on Conceptual Graphs (CG) ( http://www.inria.fr/acacia/soft/corese ). It enables to load RDFS schemas and RDF annotations and to transform them into conceptual graph formalism. It then enables to query the base of annotations thus created, by using the projection operator offered by the conceptual graph formalism.

Corese implements RDF, RDFS, some statemEnts from OWL Lite and the query pattern part of SPARQL (Simple Protocol and RDF Query Language). Furthermore, Corese query language integrates original features such as approximate search, group, count, graph path. Approximate search consists of searching the best approximate answers to a query according to the ontology, graph path enables to search the graph structure of RDF. Corese also integrates an RDF Rule Language based on the CG Rule model. The inference rule engine works in forward chaining.

It has been applied in more than 10 applications at the INRIA and is available for download. It is a semantic web platform that enables us to design and develop semantic web applications. Corese is embedded in a Semantic Web Server (based on Tomcat) and called SEWESE

Corese was one of the first RDF engine as the development started in 1999 and a paper was published at ICCS'2000 [65] . The Acacia team has been a pioneer in applying conceptual graphs to RDF and semantic web [39] . The CG helped us for understanding and using RDF. Building a software prototype helped us for mastering the semantic web languages.

Corese benefited from an INRIA operation of software development (ODL) intended to improve quality of the implementation in order to support its diffusion.

Applications.

Corese is used as search engine:

• for the RNRT KmP project on skills management,

• for the KmP-Drire project following this RNRT project with Telecom Valley,

• for the KmP-Philips project on intra-firm skills management with Philips,

• in the Ligne de Vie project on healthcare network,

• in the MEAT project on experiment memory on transcriptome analysis,

• in our co-operation with EADS,

• in QBLS system for e-Learning.

In the past, Corese was the cornerstone of four co-operations of the Acacia team:

• the IST project, CoMMA (Corporate Memory Management through Agents) [70] , [71] ,

• the SAMOVAR project with Renault [12] , [75] , [74] ,

• the co-operative research action ESCRIRE [76] ,

• the Color action Aprobatiom with CSTB.

Diffusion.

• Corese was registered at APP.

• Corese was made available to:

  • Renault,

  • ATOS Origin,

  • T-Systems NOVA (Deutsche Telekom),

  • CSTB,

  • CSELT (Telecom Italia),

  • LIRMM,

  • Mainline team at ESSI,

  • CETU (Centre d'étude des tunnels du Ministère de l'Equipement).

  • University of Santiago Chili,

  • ENST Bretagne,

  • Tech-CICO team at Université Technologique de Troyes (UTT),

  • Facultad de Informatica, LSIIS

  • Zuhlke Engineering AG, CH

  • W3C Group on the Social Meaning of RDF Graphs, Deltek Systems, Inc. USA

  • Galaad team at INRIA Sophia Antipolis.

• In 2005, Corese was presented in demonstration to members of:

  • Semantic Systems (a Spanish SME),

  • SAP Labs,

  • New South Wales University.

• The work on Corese was published in [18] , [51] , [34] , [65] , [66] , [64] .

• There is a project of creation of a start-up, eCore, in order to market a solution for skills management based on Corese.