Learning Ontologies from RDF Annotations

19
Learning Ontologies from RDF Annotations Alexandre Delteil, Catherine Faron-Zucker, Rose Dieng ACACIA project, INRIA, 2004 Sophia Antipolis, France

description

Learning Ontologies from RDF Annotations. Alexandre Delteil, Catherine Faron-Zucker, Rose Dieng ACACIA project, INRIA, 2004 Sophia Antipolis, France. TOC. Introduction RDF & RDFS Background Ontology Example Approach to Ontology Learning Conclusion Future Work. Introduction. - PowerPoint PPT Presentation

Transcript of Learning Ontologies from RDF Annotations

Page 1: Learning Ontologies from RDF Annotations

Learning Ontologies from RDF Annotations

Alexandre Delteil, Catherine Faron-Zucker, Rose Dieng

ACACIA project, INRIA, 2004Sophia Antipolis, France

Page 2: Learning Ontologies from RDF Annotations

TOC

IntroductionRDF & RDFS BackgroundOntology ExampleApproach to Ontology LearningConclusionFuture Work

Page 3: Learning Ontologies from RDF Annotations

Introduction

Build ontologies from information extracted from RDF annotations

“We have … a method to learn ontologies from RDF annotations by systematically generating the most specific generalization of all the possible sets of resources.”

Page 4: Learning Ontologies from RDF Annotations

RDF Annotation

Triplet statement (resource, property, value), (Njal, type, Cat)Easily represented as a graphXML syntax provided

Resource or literalProperty

Page 5: Learning Ontologies from RDF Annotations

XML Serialization of RDF Annotation

<rdf:Description about=‘#Njal’><rdf:type resource=‘#Cat’ /><livesIn>

<rdf:Description><rdf:type resource=‘#House’ /><ownedBy rdf:resouce=‘Catherine’ />

</rdf:Description></livesIn>

</rdf:Description>

Anonyms Resource

Page 6: Learning Ontologies from RDF Annotations

RDF Schema (RDFS)

RDFS -> schema specification languageSpecifies ontological knowledge used in RDF statementsConsists of a set of declarations of classes and propertiesDefines class and property hierarchiesMultiple inheritance

Page 7: Learning Ontologies from RDF Annotations

RDFS Metamodel and RDFS Ontology

Page 8: Learning Ontologies from RDF Annotations

Pieces of Knowledge and Descriptions

Piece of Knowledge -> set of nodes directly connected with the resource…Descriptionn -> largest set of nodes connected with the resource and having a path length <= nComplete Description -> the set of nodes connected to the resource through all possible properties

Page 9: Learning Ontologies from RDF Annotations

Piece of Knowledge Relative to Njal

Page 10: Learning Ontologies from RDF Annotations

D1(Njal): Description

Page 11: Learning Ontologies from RDF Annotations

D2(Njal): Description

Page 12: Learning Ontologies from RDF Annotations

Ontology Learning

Systematically consider all concepts covering a set of resource nodes RDF graph resource extraction

techniques preliminary first step

Group concepts and resources based on intensions and extensionsIncrementally build generalization hierarchy

Page 13: Learning Ontologies from RDF Annotations

Building of S1

Page 14: Learning Ontologies from RDF Annotations

D1(Njal): Concept Hierarchy

Page 15: Learning Ontologies from RDF Annotations

Hierarch Based on Descriptions of Length N

Construct triples of intensions and related extensionsIteratively join triple L1 with triple in pathJoin all possible triples and pathsConstruct intensions of length n Build sets Sn from inclusion relations between node extensions

Page 16: Learning Ontologies from RDF Annotations

Building of S2 from S1

Page 17: Learning Ontologies from RDF Annotations

D2(Njal): Concept Hierarchy

Page 18: Learning Ontologies from RDF Annotations

Conclusion

Lacks clarityGaps in logic in explanation, S1 -> OntologyRelies on RDF annotations previously generatedResult complexity can increase exponentiallyRequires no training dataLittle or no user inputImplemented and tested inside European IST Comma Project

Page 19: Learning Ontologies from RDF Annotations

Future Work

Inclusion of heuristicsInsertion of domain specific criteriaGraphical UIBounding methods to reduce complexityRDF annotation generator