Inferenceon the Semantic WebMyungjin Lee
Artificial Intelligence
FussySystemthe intelligence of machinesmethodologyMachineLearningNeuralNetworkgoalmethodologyhasApproachArtificialIntelligenceGeneticAlgorithmmethodologyKnowledgeBaseApproachLogichasApproachbasedonApproaches of AI
What is Semantic Web?WebtargetSemanticWebArtificialIntelligencegoalthe intelligence of machinespurposea vision of informationthat is understandable by computers,so that they can performmore of the tedious work involvedin finding, sharing, and combining informationon the web.dc:description
approachApproach of Semantic WebSemanticWebKnowledgeBaseApproachLogicSentencebasedonbasedonuserepresentationrepresentationOntologyrepresentationPropositionalLogicPredicateLogicFist OrderLogicDescriptionLogicpartOfpartOf
Ontology on the Semantic WebOWLOntologycomponentSCOTcomponentRDFRDFSvocabularycomponentSKOSdc:descriptionXMLcomponentvocabularySIOCURIAn ontology is a formalexplicit specification ofa conceptualization.FOAF
Merits of OntologyDatabaseOntologyowl:sameAs¬owl:differentFromimageimagedifferencesdifferencesrdf:Bagrdf:lirdf:lirdf:lia powerofrepresen-tationInferenceSemantics
Task of InferenceInferencebeing able to derive new datafrom data that you already knowdc:descriptiontasktaskdc:descriptionRuleInferenceTBoxInferenceOntologyInferencestatementsthat describe a systemin terms of controlledvocabulariestaskdc:descriptiontaskdc:descriptionABoxInferenceTBox-compliant statementsabout that vocabularyto produce valid statementswithin systembased on rule
Ontology InferenceOntologyInferenceto derive additional facts to be inferredfrom instance data and class descriptionsbased on own semanticsdc:descriptionRDF SemanticsPerson<x, y> is in IEXT(I(rdfs:subClassOf))if and only if x and y are in ICand ICEXT(x) is a subset of ICEXT(y)ManMyungjin( Man		rdfs:subClassOf	Person )( Myungjinrdf:type		Man )( Myungjinrdf:type	Person )
TBox InferenceTBoxInferenceOntologyInferencestatementsthat describe a systemin terms of controlledvocabulariesdc:descriptiontask<rdfs:Classrdf:about="http://xmlns.com/foaf/0.1/Document" rdfs:label="Document”>	<rdfs:subClassOfrdf:resource="http://xmlns.com/wordnet/1.6/Document"/></rdfs:Class><rdfs:Classrdf:about="http://xmlns.com/foaf/0.1/PersonalProfileDocument”>	<rdfs:subClassOfrdf:resource="http://xmlns.com/foaf/0.1/Document"/></rdfs:Class>http://xmlns.com/foaf/0.1/PersonalProfileDocumentrdfs:subClassOf					http://xmlns.com/wordnet/1.6/Document
ABox InferenceOntologyInferenceABoxInferenceTBox-compliant statementsabout that vocabularydc:descriptiontask<rdf:Propertyrdf:about="http://xmlns.com/foaf/0.1/homepage” rdfs:label="homepage“ >	<rdfs:subPropertyOfrdf:resource="http://xmlns.com/foaf/0.1/page"/></rdf:Property><foaf:Personrdf:about="#me" xmlns:foaf="http://xmlns.com/foaf/0.1/">	<foaf:name>Dan Brickley</foaf:name>	<foaf:homepagerdf:resource="http://danbri.org/" /></foaf:Person>http://xmlns.com/foaf/0.1 /#me		foaf:page		http://danbri.org/
Rule InferenceRuleInferenceto produce valid statementswithin systembased on ruledc:descriptionif	hasParent(?x, ?y)hasParent(?x, ?z)	Man(?y)	Woman(?z)then	hasWife(?y, ?z)hasWifehasParenthasParent
SWRL (Semantic Web Rule Language)SWRLHorn-likeRuleMemberSubmissionrepresentationstatuseditorformsubLanguageSWRLTabRuleMLBodyrdf:Seqrdf:lirdf:liHeadplugInProtégéscreenshot
Inference Engine for Semantic WebBossama forward chaining rule engine supports SWRLdc:descriptionrdf:typePelletan open-source Java OWL DL reasoner has SWRL-supportdc:descriptionInferenceEnginerdf:typeKAON2an infrastructure for managing OWL-DL, SWRL, and F-Logic ontologiesdc:descriptionrdf:typeRacerPrordf:typeprocessing of rules in a SWRL-based syntax by translating them into nRQL rulesdc:descriptionrdf:typeJenato derive additional RDF assertions, the axioms and rules associated with the reasonerdc:description
SMART SystemIntelligenceInformationSystemLabYonseiUniversityjava framework for semantic web applicationlocatedIncreateddc:descriptionSMARTfunctionsupportrdf:Bagrdf:Bagrdf:lirdf:lirdf:lirdf:lirdf:liOntologyProcessSPARQLProcessrdf:liSPARQLRDFRDFSRuleInferenceOntologyInferenceSWRLOWL
Example DemoSWRL Ruleif	sioc:Post(?x)sioc:Post(?y)sioc:topic(?x, ?a)sioc:topic(?y, ?b)rdf:type(?a, ?z)rdf:type(?b, ?z)then	sioc:related_to(?x, ?y)sioc:Postsioc:Postrdf:typerdf:typesioc:related_toclouds-with-siocsample-postsioc:topicsioc:topicsemantic-websws.geonames.orgSPARQL QueryPREFIX sioc: <http://rdfs.org/sioc/ns#>SELECT ?u ?v WHERE{	?u	sioc:related_to	?v .}rdf:typerdf:typesemanticweb
Issue of InferenceRDF 상에서 어디에 추론을 쓰지?새로운 관계 발견을 통한 네트워크 분석또 다른 RDF Vocabularies 혹은 도메인 온톨로지와의 관계 규칙 정의 및 추론고민할 문제추론을 위한 표현력과 복잡도많은 룰에 의한 충돌세상사를 반영한 규칙의 생성철저한 준비?표현력의 한계?
? !

Inference on the Semantic Web

  • 1.
  • 2.
  • 3.
    FussySystemthe intelligence ofmachinesmethodologyMachineLearningNeuralNetworkgoalmethodologyhasApproachArtificialIntelligenceGeneticAlgorithmmethodologyKnowledgeBaseApproachLogichasApproachbasedonApproaches of AI
  • 4.
    What is SemanticWeb?WebtargetSemanticWebArtificialIntelligencegoalthe intelligence of machinespurposea vision of informationthat is understandable by computers,so that they can performmore of the tedious work involvedin finding, sharing, and combining informationon the web.dc:description
  • 5.
    approachApproach of SemanticWebSemanticWebKnowledgeBaseApproachLogicSentencebasedonbasedonuserepresentationrepresentationOntologyrepresentationPropositionalLogicPredicateLogicFist OrderLogicDescriptionLogicpartOfpartOf
  • 6.
    Ontology on theSemantic WebOWLOntologycomponentSCOTcomponentRDFRDFSvocabularycomponentSKOSdc:descriptionXMLcomponentvocabularySIOCURIAn ontology is a formalexplicit specification ofa conceptualization.FOAF
  • 7.
  • 8.
    Task of InferenceInferencebeingable to derive new datafrom data that you already knowdc:descriptiontasktaskdc:descriptionRuleInferenceTBoxInferenceOntologyInferencestatementsthat describe a systemin terms of controlledvocabulariestaskdc:descriptiontaskdc:descriptionABoxInferenceTBox-compliant statementsabout that vocabularyto produce valid statementswithin systembased on rule
  • 9.
    Ontology InferenceOntologyInferenceto deriveadditional facts to be inferredfrom instance data and class descriptionsbased on own semanticsdc:descriptionRDF SemanticsPerson<x, y> is in IEXT(I(rdfs:subClassOf))if and only if x and y are in ICand ICEXT(x) is a subset of ICEXT(y)ManMyungjin( Man rdfs:subClassOf Person )( Myungjinrdf:type Man )( Myungjinrdf:type Person )
  • 10.
    TBox InferenceTBoxInferenceOntologyInferencestatementsthat describea systemin terms of controlledvocabulariesdc:descriptiontask<rdfs:Classrdf:about="http://xmlns.com/foaf/0.1/Document" rdfs:label="Document”> <rdfs:subClassOfrdf:resource="http://xmlns.com/wordnet/1.6/Document"/></rdfs:Class><rdfs:Classrdf:about="http://xmlns.com/foaf/0.1/PersonalProfileDocument”> <rdfs:subClassOfrdf:resource="http://xmlns.com/foaf/0.1/Document"/></rdfs:Class>http://xmlns.com/foaf/0.1/PersonalProfileDocumentrdfs:subClassOf http://xmlns.com/wordnet/1.6/Document
  • 11.
    ABox InferenceOntologyInferenceABoxInferenceTBox-compliant statementsaboutthat vocabularydc:descriptiontask<rdf:Propertyrdf:about="http://xmlns.com/foaf/0.1/homepage” rdfs:label="homepage“ > <rdfs:subPropertyOfrdf:resource="http://xmlns.com/foaf/0.1/page"/></rdf:Property><foaf:Personrdf:about="#me" xmlns:foaf="http://xmlns.com/foaf/0.1/"> <foaf:name>Dan Brickley</foaf:name> <foaf:homepagerdf:resource="http://danbri.org/" /></foaf:Person>http://xmlns.com/foaf/0.1 /#me foaf:page http://danbri.org/
  • 12.
    Rule InferenceRuleInferenceto producevalid statementswithin systembased on ruledc:descriptionif hasParent(?x, ?y)hasParent(?x, ?z) Man(?y) Woman(?z)then hasWife(?y, ?z)hasWifehasParenthasParent
  • 13.
    SWRL (Semantic WebRule Language)SWRLHorn-likeRuleMemberSubmissionrepresentationstatuseditorformsubLanguageSWRLTabRuleMLBodyrdf:Seqrdf:lirdf:liHeadplugInProtégéscreenshot
  • 14.
    Inference Engine forSemantic WebBossama forward chaining rule engine supports SWRLdc:descriptionrdf:typePelletan open-source Java OWL DL reasoner has SWRL-supportdc:descriptionInferenceEnginerdf:typeKAON2an infrastructure for managing OWL-DL, SWRL, and F-Logic ontologiesdc:descriptionrdf:typeRacerPrordf:typeprocessing of rules in a SWRL-based syntax by translating them into nRQL rulesdc:descriptionrdf:typeJenato derive additional RDF assertions, the axioms and rules associated with the reasonerdc:description
  • 15.
    SMART SystemIntelligenceInformationSystemLabYonseiUniversityjava frameworkfor semantic web applicationlocatedIncreateddc:descriptionSMARTfunctionsupportrdf:Bagrdf:Bagrdf:lirdf:lirdf:lirdf:lirdf:liOntologyProcessSPARQLProcessrdf:liSPARQLRDFRDFSRuleInferenceOntologyInferenceSWRLOWL
  • 16.
    Example DemoSWRL Ruleif sioc:Post(?x)sioc:Post(?y)sioc:topic(?x,?a)sioc:topic(?y, ?b)rdf:type(?a, ?z)rdf:type(?b, ?z)then sioc:related_to(?x, ?y)sioc:Postsioc:Postrdf:typerdf:typesioc:related_toclouds-with-siocsample-postsioc:topicsioc:topicsemantic-websws.geonames.orgSPARQL QueryPREFIX sioc: <http://rdfs.org/sioc/ns#>SELECT ?u ?v WHERE{ ?u sioc:related_to ?v .}rdf:typerdf:typesemanticweb
  • 17.
    Issue of InferenceRDF상에서 어디에 추론을 쓰지?새로운 관계 발견을 통한 네트워크 분석또 다른 RDF Vocabularies 혹은 도메인 온톨로지와의 관계 규칙 정의 및 추론고민할 문제추론을 위한 표현력과 복잡도많은 룰에 의한 충돌세상사를 반영한 규칙의 생성철저한 준비?표현력의 한계?
  • 18.