際際滷

際際滷Share a Scribd company logo
A Linked Data hub

       Julien Cojan

                        Wimmics


 COST workshop  April 5th 2013
Head (and INRIA contact): Fabien Gandon

Vice Head (and I3S contact): Catherine Faron-Zucker
                                                      PhD students:
Researchers:                                          1. Pavel Arapov, 1st year (EDSTIC-I3S)
 Michel Buffa, MdC (UNS)                             2. Franck Berthelon, 3nd year (UNS-EDSTIC)
 Olivier Corby, CR1 (INRIA)                          3. Ahlem Bouchahda, 3rd year (UNS-SupCom Tunis)
 Alain Giboin, CR1 (INRIA)                           4. Khalil Riad Bouzidi, 3rd year (UNS-CSTB)
 Nhan Le Thanh, Pr. (UNS)                            5. Luca Costabello, 2nd year (INRIA-CORDI)
 Isabelle Mirbel, MdC, HDR (UNS)                     6. Papa Fary Diallo, 1st year (AUF-UGB-INRIA)
 Peter Sander, Pr. (UNS)                             7. Corentin Follenfant, 2nd year (SAP)
 Andrea G. B. Tettamanzi, Pr. (UNS)                  8. Maxime Lefran巽ois, 2nd year (EDSTIC-INRIA)
 Serena Villata, RP (INRIA)                          9. Nguyen Thi Hoa Hue, 1st year (Vietnam-CROUS)
                                                      10. Nicolas Marie, 2nd year (Bell-ALU, INRIA)
Post-doc:                                             11. Zide Meng, 1st year (INRIA, ANR OCKTOPUS)
 Zeina Azmeh (I3S)                                   12. Rakebul Hasan, 1st year (INRIA ANR-Kolflow)
 Elena Cabrio (CORDIS)                               13. Oumy Seye, 2nd year, (INRIA Rose Dieng allocation)
                                                      14. Imen Tayari, 3rd year (UNS-Sfax Tunisie)
Research engineers:
 Julien Cojan (INRIA, Ministry of Culture)           Assistants:
 Christophe Desclaux (Boost your code)                Christine Foggia (INRIA)
 Amosse Edouard (I3S)                                 Sarah Choulet (I3S)
research problem
socio-semantic networks: combining formal
semantics and social semantics on the web
research fields
socio-semantic networks: combining formal
semantics and social semantics on the web
 web-supported epistemic communities
 model and support actors, actions & interactions
 graph-based representation & reasoning
web landscape and graphs
(meta)data of the relations and the resources of the web


         =              +              +              +               +              +
web         sites          social       of data   of services        semantics




         =              +              +              +               +              +
typed          web          networks   linked data        workflows       schemas
graphs       (graphs)       (graphs)     (graphs)          (graphs)       (graphs)
projects
            isicil.inria.fr (ANR)
               enterprise social networking
               business intelligence, watching, monitoring
               communities of interest, of practice, of experts

            datalift.org (ANR)
DATALIFT




               from raw public data to interlinked data and schemas
               a platform and documentation to assist the process
               validation on real datasets


            kolflow.univ-nantes.fr (ANR)
               reduce the overhead of communities building knowledge
               federated semantic: distributed blackboard for man-machine coop.
projects
 dbpedia.fr (Ministry of Culture)
    Extract and publish data and facts from French version of wikipedia
    Publication on the Linked Data
    Member of the DBpedia internationalization committee


 QAKIS
    Question answering system in natural language
    Relate DBpedia data with expression in Wikipedia text
    Queries on Linked Data


 DiscoveryHub
    Recommendation system
    Interest propagation over DBpedia graph
    Cross-domain suggestions
wimmics and DBpedia FR
Semantization of French speaking resources of Wikimedia


Projects :
        DBpedia en fran巽ais
        Semantization of the Wiktionnary
        Interconnexion with other cultural ressources
        Complementary tools (Lookup, Spotlight)
http://semanticpedia.org
wikipedia
                                            encyclopedic pages




際際滷 adapted from F. Gandon presentation
wikipedia


                         dbpedia
                                            extraction framework


                                            encyclopedic data




際際滷 adapted from F. Gandon presentation
Data extracted from Wikipedia
Data extracted from Wikipedia


     48.861073       2.335784
                                           Cat辿gorie:
                                       Mus辿e_辿gyptologique
            latitude
 Infobox                     longitude
  Mus辿e                                        cat辿gorie
               mod竪le
                             Mus辿e du Louvre
  10/08/1793

                    adresse                 ville
99, rue de Rivoli
  75001 Paris
                             pays          Paris
                                                     cp
                                    pays
                                                       75056
                    France                 r辿gion

                                pays
                                           le-de-France
RDF representation

                   Mus辿e du Louvre           ville            Paris


        dbFr:Mus辿e_du_Louvre          dbFr-prop:ville    dbpedia-fr:Paris

                  (Subject)                (predicate)      (object)



Mus辿e du Louvre
                    Subject of Wikipedia articles :
                      http://fr.dbpedia.org/resource/[Page_Name]
                                 dbFr:
     ville           Properties extracted textually :
                      http://fr.dbpedia.org/property/[property]
                              dbFr-prop:
social
                    maintenance
                         dbpedia            wiki for vocabulary and
                                            extraction patterns




際際滷 adapted from F. Gandon presentation
Normalized vocabulary

Issues raised by the textual extraction
   Ambiguities
            Mus辿e du Louvre   pays        France


              Raymond         pays        France
              Poulidor


 Terms Variations

             Emmanuel Kant    naissance     22/04/1724


                              date de
              Le Corbusier                  06/10/1887
                              naissance
Normalized vocabulary

Vocabulary collaboratively defined on
      http://mappings.dbpedia.org

Ontology :
 Class hierarchy

        http://dbpedia.org/ontology/Museum
 Properties

        http://dbpedia.org/ontology/city   dbo:
                          Infobox                                dbo :
              infobox                                rdf:type
Mappings :                 Mus辿e                                Museum
       Mus辿e du                                Mus辿e du
        Louvre                                  Louvre

                  ville    Paris                     dbo:city    Paris
Interlinking
                     Outgoing links to :
                          Freebase
                          CIA factbook
                          Geonames
                          Project
                           Guttenberg
                          DrugBank
                     ...

               http://wiki.dbpedia.org/Interlinking
Multilingual Linked Open Data

                    Member of DBpedia
                     internationalization comittee
                     [1] (13 chapters)
                    Interconnexion
                          Equivalence links
                           between instances
                          Common vocabulary
                           (DBpedia Ontology)


                  [1] http://dbpedia.org/Internationalization
DBpedia En vs DBpedia Fr

      Number of common values                                   343 583
      Number of differentes values                              175 227
      Number of values only in DBpedia Fr                       460 586

      Number of values only in DBpedia En                      2 022 557



Causes for the differences :
   Error/vandalism
   Disagreement (China superficie)
   Granularity Level
    DBpedia En, Seine spring location : Burgundy (Bougogne),
    DBpedia Fr, Seine spring location : plateau de Langres.
Interconnexions
DBpedia En                                           DBpedia Fr

                                                                          Paris

                                                                                      Troyes
                                                             dbo:city
                                                                           dbo:city

                               Seine                         Seine

             dbo:river_mouth                             dbo:river_mouth

                        dbo:river_mouth
      English                                               Manche
      Channel                                                (mer)
                         Baie de
                         la Seine


         Property                      English label              French label
         dbo:river_mouth               river mouth                embouchure
         dbo:city                      city                       ville
Interconnexions
DBpedia En                                           DBpedia Fr

                                                                          Paris

                                                                                       Troyes
                                                              dbo:city
                                                                           dbo:city

                               Seine                          Seine

             dbo:river_mouth                             dbo:river_mouth

                        dbo:river_mouth                                       dbo:river_mouth
      English                                                Manche
      Channel                                                 (mer)
                         Baie de                                                       Baie de
                         la Seine                                                      la Seine


         Property                      English label              French label
         dbo:river_mouth               river mouth                embouchure
         dbo:city                      city                       ville
wiktionary



                         dbpedia
                                            extraction framework


                                            lexicographic data




際際滷 adapted from F. Gandon presentation
Wiki source



                         dbpedia
                                            extraction framework


                                            Literature data




際際滷 adapted from F. Gandon presentation
https://www.wikidata.org


   Knowledge Base collaboratively edited
    (Wikimedia commons for data)

   Overlap with DBpedia
        Phase 1 : interlanguage links
        Phase 2 : infoboxes

   Future
       Interlinking (compatibility)
       Migrate tools
http://dbpedia.inria.fr/qakis
wimmics and DBpedia FR
wimmics and DBpedia FR
wimmics and DBpedia FR
wimmics and DBpedia FR
wimmics and DBpedia FR
http://semreco.inria.fr
Zone 47   http://zone47.com/db/
HdA Lab   http://hdalab.iri-research.org/hdalab/
wimmics and DBpedia FR
web@Inria
graphes, arbre et r辿seaux           visualisation et interaction
 Asap                               Aviz
 Graphik                            Gravite
 Maestro                            In-Situ
 Mostrare                           Imedia 2
                                     Texmex

gestion des bases de donn辿es         structure, s辿mantique
 Dahu                                Exmo
 Leo                                 Orpailleur
 Zenith                              Score
                                      Wam
programmation, r辿seaux, distribution  Wimmics
 Indes
 Oasis
 Planete
 Smis
WEB
science
          37
demain, celui qui contr担lera les m辿tadonn辿es,
           contr担lera le web.




                                http://fabien.info
                                @fabien_gandon

More Related Content

wimmics and DBpedia FR

  • 1. A Linked Data hub Julien Cojan Wimmics COST workshop April 5th 2013
  • 2. Head (and INRIA contact): Fabien Gandon Vice Head (and I3S contact): Catherine Faron-Zucker PhD students: Researchers: 1. Pavel Arapov, 1st year (EDSTIC-I3S) Michel Buffa, MdC (UNS) 2. Franck Berthelon, 3nd year (UNS-EDSTIC) Olivier Corby, CR1 (INRIA) 3. Ahlem Bouchahda, 3rd year (UNS-SupCom Tunis) Alain Giboin, CR1 (INRIA) 4. Khalil Riad Bouzidi, 3rd year (UNS-CSTB) Nhan Le Thanh, Pr. (UNS) 5. Luca Costabello, 2nd year (INRIA-CORDI) Isabelle Mirbel, MdC, HDR (UNS) 6. Papa Fary Diallo, 1st year (AUF-UGB-INRIA) Peter Sander, Pr. (UNS) 7. Corentin Follenfant, 2nd year (SAP) Andrea G. B. Tettamanzi, Pr. (UNS) 8. Maxime Lefran巽ois, 2nd year (EDSTIC-INRIA) Serena Villata, RP (INRIA) 9. Nguyen Thi Hoa Hue, 1st year (Vietnam-CROUS) 10. Nicolas Marie, 2nd year (Bell-ALU, INRIA) Post-doc: 11. Zide Meng, 1st year (INRIA, ANR OCKTOPUS) Zeina Azmeh (I3S) 12. Rakebul Hasan, 1st year (INRIA ANR-Kolflow) Elena Cabrio (CORDIS) 13. Oumy Seye, 2nd year, (INRIA Rose Dieng allocation) 14. Imen Tayari, 3rd year (UNS-Sfax Tunisie) Research engineers: Julien Cojan (INRIA, Ministry of Culture) Assistants: Christophe Desclaux (Boost your code) Christine Foggia (INRIA) Amosse Edouard (I3S) Sarah Choulet (I3S)
  • 3. research problem socio-semantic networks: combining formal semantics and social semantics on the web
  • 4. research fields socio-semantic networks: combining formal semantics and social semantics on the web web-supported epistemic communities model and support actors, actions & interactions graph-based representation & reasoning
  • 5. web landscape and graphs (meta)data of the relations and the resources of the web = + + + + + web sites social of data of services semantics = + + + + + typed web networks linked data workflows schemas graphs (graphs) (graphs) (graphs) (graphs) (graphs)
  • 6. projects isicil.inria.fr (ANR) enterprise social networking business intelligence, watching, monitoring communities of interest, of practice, of experts datalift.org (ANR) DATALIFT from raw public data to interlinked data and schemas a platform and documentation to assist the process validation on real datasets kolflow.univ-nantes.fr (ANR) reduce the overhead of communities building knowledge federated semantic: distributed blackboard for man-machine coop.
  • 7. projects dbpedia.fr (Ministry of Culture) Extract and publish data and facts from French version of wikipedia Publication on the Linked Data Member of the DBpedia internationalization committee QAKIS Question answering system in natural language Relate DBpedia data with expression in Wikipedia text Queries on Linked Data DiscoveryHub Recommendation system Interest propagation over DBpedia graph Cross-domain suggestions
  • 9. Semantization of French speaking resources of Wikimedia Projects : DBpedia en fran巽ais Semantization of the Wiktionnary Interconnexion with other cultural ressources Complementary tools (Lookup, Spotlight) http://semanticpedia.org
  • 10. wikipedia encyclopedic pages 際際滷 adapted from F. Gandon presentation
  • 11. wikipedia dbpedia extraction framework encyclopedic data 際際滷 adapted from F. Gandon presentation
  • 12. Data extracted from Wikipedia
  • 13. Data extracted from Wikipedia 48.861073 2.335784 Cat辿gorie: Mus辿e_辿gyptologique latitude Infobox longitude Mus辿e cat辿gorie mod竪le Mus辿e du Louvre 10/08/1793 adresse ville 99, rue de Rivoli 75001 Paris pays Paris cp pays 75056 France r辿gion pays le-de-France
  • 14. RDF representation Mus辿e du Louvre ville Paris dbFr:Mus辿e_du_Louvre dbFr-prop:ville dbpedia-fr:Paris (Subject) (predicate) (object) Mus辿e du Louvre Subject of Wikipedia articles : http://fr.dbpedia.org/resource/[Page_Name] dbFr: ville Properties extracted textually : http://fr.dbpedia.org/property/[property] dbFr-prop:
  • 15. social maintenance dbpedia wiki for vocabulary and extraction patterns 際際滷 adapted from F. Gandon presentation
  • 16. Normalized vocabulary Issues raised by the textual extraction Ambiguities Mus辿e du Louvre pays France Raymond pays France Poulidor Terms Variations Emmanuel Kant naissance 22/04/1724 date de Le Corbusier 06/10/1887 naissance
  • 17. Normalized vocabulary Vocabulary collaboratively defined on http://mappings.dbpedia.org Ontology : Class hierarchy http://dbpedia.org/ontology/Museum Properties http://dbpedia.org/ontology/city dbo: Infobox dbo : infobox rdf:type Mappings : Mus辿e Museum Mus辿e du Mus辿e du Louvre Louvre ville Paris dbo:city Paris
  • 18. Interlinking Outgoing links to : Freebase CIA factbook Geonames Project Guttenberg DrugBank ... http://wiki.dbpedia.org/Interlinking
  • 19. Multilingual Linked Open Data Member of DBpedia internationalization comittee [1] (13 chapters) Interconnexion Equivalence links between instances Common vocabulary (DBpedia Ontology) [1] http://dbpedia.org/Internationalization
  • 20. DBpedia En vs DBpedia Fr Number of common values 343 583 Number of differentes values 175 227 Number of values only in DBpedia Fr 460 586 Number of values only in DBpedia En 2 022 557 Causes for the differences : Error/vandalism Disagreement (China superficie) Granularity Level DBpedia En, Seine spring location : Burgundy (Bougogne), DBpedia Fr, Seine spring location : plateau de Langres.
  • 21. Interconnexions DBpedia En DBpedia Fr Paris Troyes dbo:city dbo:city Seine Seine dbo:river_mouth dbo:river_mouth dbo:river_mouth English Manche Channel (mer) Baie de la Seine Property English label French label dbo:river_mouth river mouth embouchure dbo:city city ville
  • 22. Interconnexions DBpedia En DBpedia Fr Paris Troyes dbo:city dbo:city Seine Seine dbo:river_mouth dbo:river_mouth dbo:river_mouth dbo:river_mouth English Manche Channel (mer) Baie de Baie de la Seine la Seine Property English label French label dbo:river_mouth river mouth embouchure dbo:city city ville
  • 23. wiktionary dbpedia extraction framework lexicographic data 際際滷 adapted from F. Gandon presentation
  • 24. Wiki source dbpedia extraction framework Literature data 際際滷 adapted from F. Gandon presentation
  • 25. https://www.wikidata.org Knowledge Base collaboratively edited (Wikimedia commons for data) Overlap with DBpedia Phase 1 : interlanguage links Phase 2 : infoboxes Future Interlinking (compatibility) Migrate tools
  • 33. Zone 47 http://zone47.com/db/
  • 34. HdA Lab http://hdalab.iri-research.org/hdalab/
  • 36. web@Inria graphes, arbre et r辿seaux visualisation et interaction Asap Aviz Graphik Gravite Maestro In-Situ Mostrare Imedia 2 Texmex gestion des bases de donn辿es structure, s辿mantique Dahu Exmo Leo Orpailleur Zenith Score Wam programmation, r辿seaux, distribution Wimmics Indes Oasis Planete Smis
  • 38. demain, celui qui contr担lera les m辿tadonn辿es, contr担lera le web. http://fabien.info @fabien_gandon