240 likes | 342 Views
Semantic Hubs for Geological Projects P. DURVILLE (INRIA) , F. Husson (BRGM) & e-WOK_HUB Consortium. SeMMA 2008. e-WOK_HUB Consortium : INRIA, EADS, LISI/CRITT, BRGM, IFP, ENSMP. Outline. Introduction : Use-case/application context (geological projects)
E N D
Semantic Hubs for Geological ProjectsP. DURVILLE (INRIA), F. Husson (BRGM)& e-WOK_HUB Consortium SeMMA 2008 e-WOK_HUB Consortium : INRIA, EADS, LISI/CRITT, BRGM, IFP, ENSMP
Outline • Introduction : Use-case/application context(geological projects) • Global architecture : SOA and Semantic web technologies • Zoom on ontology management services : ECCO tool • Zoom on annotation generation services : global process • Zoom on search service : geographical search • Conclusion
Use-case / Application context • CO2 geological storage prospection projects • In which are involved people with different profiles • geologists, researchers, engineers, project leaders… • Working on a large variety of available resources • Internal or external project reports, scientific articles, databases … • Producing new knowledge that can be usefull for current and future projects improving a better reuse of knowledge. • Need the discovery, interoperability and integration of these resources
SOA and Semantic Web technologies • Semantic web technologies • Ontologies (RDFS/OWL lite) • Annotations (RDF) • Rules (CORESE rules) • Queries (SPARQL) • SOA (service oriented architecture) • ESB (Petals) • Semantic web services (SAWSDL) CORESE is used to find possible service compositions • a hub is a warehouse of semantic business resources
Service metadata creation process SAWSDL WSDL Semantic repository RDF
How to use service metadata? WSDL CORESE Orchestra BPEL activities SPARQL query Semantic repository
Ontology management service - ECCO • ECCO : ontology editor • contextual and collaborative • ontology whole life cycle • term extraction from domain texts • vocabulary development • hierarchy construction • OWL Lite representation edition • Track of modifications for annotation evolution
ECCO: Term extraction from domain texts
ECCO : Colla-borative vocabulary edition
Annotation services • Annotation generation • Annotation storage • Annotation evolution
Annotation services • Annotation generation • Annotation storage • Annotation evolution
Annotation generation services Text only Language detection + Annotation dc :language Grammatical analysis + Annotations verb, adj., noun, … Semantic annotation Statistic annotation + Annotations Of domain - Annotations verb, adj., noun, …
Annotation generation services Text only Language detection + Annotation dc :language Grammatical analysis + Annotations verb, adj., noun, … Semantic annotation Statistic annotation + Annotations Of domain - Annotations verb, adj., noun, …
“3D stratigraphic geometries of the intracratonic MesoCenozoic Paris Basin were obtained by sequence stratigraphic correlations…” Semantic annotation service Text Annotations (RDF) RDF generation Linguistic process Ontologies (OWL/RDFS) Term extractor Jaro- Winkler CORESE Pattern (SPARQL) slack <rdf:RDF xmlns:geo="http://rdf.inria.fr/geo/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:wl="http://model.core.weblab.eads.com#"> <wl:Segment rdf:about="weblab://myWS/myDocument#1_inriaForGeo_1"> <geo:geolocalized rdf:resource="http://rdf.insee.fr/geo/DEP_75"/> </wl:Segment> </rdf:RDF>
Search service • Search examples: • - Documents with content talking about Paris basin area • - Documents with content at 50 km from Reims
Many various databanks Databases ex : National Borehole Databank Reports written by BRGM and IFP • 2 different ways for the geolocalization of resources indirect geolocalization direct geolocalization Annotation : {x,y} Annotation : Paris basin Matching with ontologies Search in databases
1’ Cartography selection of an interesting zone 1’’ Textfield input fornamed entities PARIS BASIN Semantic-Geographic bridge 2 Process 3 Geolocalization Database ex : National Borehole Databank PDF report 3’’ Indirect geolocalization 3’ Direct geolocalization
2 Translation into a SPARQL query sent to and processed by a search engine PREFIX geo: <http://rdf.insee.fr/geo/> PREFIX geo2: <http://rdf.inria.fr/geo/>PREFIX wl:<http://model.core.weblab.eads.com#> SELECT distinct ?doc WHERE { ?doc rdf:type wl:Document . ?doc wl:contains ?res . { ?res geo2:geolocalized ?x . FILTER (?res = <URI1> || ?res = <URI2> || … ) } UNION { ?res geo2:geolocalized ?y . ?z geo:subdivision ?y . FILTER (?res = <URI3> || …) } } 1 Cartography selection of an interesting zone 3 Documents found are displayed to users
Conclusion • A service oriented architecture that integrates semantic web technologies • Combining different services/tools allowing semantic metadata management such as : • Ontology development in a collaborative and contextual way • Ontology evolution and impact on annotations • Annotation generation combining two different approaches • Knowledge retrieval by means of SPARQL queries on metadata generated • Considering services as resources in order to add metadata on them so as to perform retrieval and composition of services/applications
Conclusion - Perspectives • On annotation management : • Annotation generation on textsconsidering weighting • Multi view pointsmanagement of annotations • On semantic web services : • BPEL activies generation • On search aspects : • Combining semantic and specific searches: • combining queries • combining results
Gracias! Preguntas? e-WOK_HUB Consortium : INRIA, EADS, LISI/CRITT, BRGM, IFP, ENSMP