Leveraging Data And Structure In Ontology . - Unibo.it
Leveraging Data and Structure inOntology IntegrationO. UdreaEnrico SavioliL. GetoorR.J. MillerGroup 15Andrea RealeAndrea SorbiniDEISUniversity of BolognaSearching Information in Large Spaces ConferenceMarch 11, 2009Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 20091 / 52
Outline1Brief Introduction to OntologiesWhat is an Ontology2Motivations and State of the ArtBringing data togetherExisting Solutions3ILIADSOverviewThe AlgorithmExperimental ResultsConclusions and Future Developments4DemoSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 20092 / 52
Brief Introduction to OntologiesWhat is an OntologyOntologiesa quest for meaningA very common problem in IT is data modelingIICritical in the field of Information SystemsA good data model enables good data usageTools used to describe concepts cannot express the implicitinterpretation of dataSemantic knowledge is lost after the modeling processExample musician name "F. Mercury" sings-for "Queen" / musician name "B. May" plays-for "Queen" / How can we find all of Queen’s members?Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 20093 / 52
Brief Introduction to OntologiesWhat is an OntologyOntologiesa formal modelOntologyAn ontology is a formal representation ofconcepts within a domain (Song, Performer, Composer.)relations between those concepts (plays-for, member-of, written-by.)individual instances (queen, f.mercury, b.may.)Well-defined constructs to enrich descriptions with semanticsFormal models to enable automatic reasoning (DL, FOL)(member(Y,Z) :- sings-for(Y,Z) ; plays-for(Y,Z))OWL is the W3C standard for an ontology languageSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 20094 / 52
Brief Introduction to OntologiesWhat is an OntologyOWLA simple exampleMusicsubClassOfRockMusictypetypeRide of theValkyriestypeJazzMusictypeWhole lottalovetypeThe Phantomof the OperaSmoke onthe waterRequiem inD er-ofW.A. MozartIron MaidentypetypePerformerSavioli, Reale, Sorbini (DEIS)Take the ubClassOftypetypeMusicianUGM07S. HarrissubClassOfWriterSI-LS 20095 / 52
Motivations and State of the ArtBringing data togetherMotivating ScenarioThe AAA principleAnyone can say Anything about Any topicThe same thing might be described in different waysThere is a strong need to integrate heterogeneous schemasINot only in a web scenarioExampleDB schemasXML schemasOntology schemasSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 20096 / 52
Motivations and State of the ArtBringing data togetherAn Integration ProblemMusic PieceSmoke onthe watersubClassOfsubClassOfMusicWhole ssOfsubClassOfsubClassOftypeClassicRocktypeThe Phantomof the OperaJazzMusicClassicalMusictypetypeRide of theValkyriesRequiem K626Iron MaidenplaysThe BlackSabbathperformscomposer-ofTake the "A"TrainS. er-ofIron MaidenRequiem inD minortypeSmoke onthe waterplaysThe Phantomof the ictypeQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 20097 / 52
Motivations and State of the ArtBringing data togetherIntegration TechniquesStructure Based MatchingInstance Based MatchingUses schema meta-data (e.g.informations about tables orconcepts models) to discovermapping elements amongthem, both on a structural andelement levelUses informations about datainstances (e.g. contents of atable or individuals of anontology) to discover mappingsamong entities representingthemSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 20098 / 52
Motivations and State of the ArtBringing data togetherOntology IntegrationOntologies offer further ways to uncover aligning relationsbetween entitiesExplicit theoretic model semantics can be leveraged to improveintegration qualityExampleAssume that composed-by is a functional property“Requiem K626” and “Requiem in D minor” are the sameIf we have:IIcomposed-by(‘Requiem K626’, ‘Mozart’).composed-by(‘Requiem in D minor’, ‘W.A. Mozart’).Then we might say that aligning “W.A. Mozart” and “Mozart” is agood choiceSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 20099 / 52
Motivations and State of the ArtExisting SolutionsFCA-Merge and COMA FCA-Merge (human-aided tool to merge ontologies)Collects domain related natural language documentsSearches those documents for ontologies’ concepts occurrencesDerives an alignment that has to be validated by an operatorCOMA (framework with multiple match strategies)Fragment based matchingReuse of previous matching resultsComprehensive GUI for results evaluationSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200910 / 52
Motivations and State of the ArtExisting SolutionsWhat’s missingBoth COMA and FCA-Merge use only structure level matchingMoreover none of them makes use of the semantic potentialoffered by ontologiesThere exist other solutions using reasoning support, however .Iit is used only for an a-posteriori consistency check of the resultAn interesting improvement might be to leverage it for the actualmatching processThat’s where ILIADS kicks in!Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200911 / 52
ILIADSOverviewILIADSIntegrated Learning In Alignment of Data and SchemaTakes two OWL Lite ontologies as inputCombines "traditional" schema matching approaches with alogical inference algorithmIInference results are used to influence confidence in a presumedmappingMakes use of both schema (structure) and data (individuals)Outputs a set of axioms (the alignment) that tights the inputontologies togetherSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200912 / 52
ILIADSThe AlgorithmAlgorithm OverviewINPUT: Consistent Ontologies O1 and O2OUTPUT: Alignment A 01: Initialize algorithm’s structures (O is O1 O2 )02: repeat:03:Compute similarity scores between clusters04:Heuristically select a type of clusters05:for each couple (c, c 0 ) of that type do06:Determine a candidate relationship a(c,c 0 )07:Perform incremental inference08:Update similarity score09:Select the best couple, update O and A 10: until there are clusters with similarity λt11: return A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200913 / 52
ILIADSThe AlgorithmThe AlgorithmStep by step01:02:03:04:05:06:07:08:09:10:11:Initialize algorithm’s structures (O is O1 O2 )repeat:Compute similarity scores between clustersHeuristically select a type of clustersfor each couple (c, c 0 ) of that type doDetermine a candidate relationship a(c,c 0 )Perform incremental inferenceUpdate similarity scoreSelect the best couple, update O and A until there are clusters with similarity λtreturn A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200914 / 52
ILIADSThe AlgorithmThe AlgorithmStructures initializationThe algorithms groups equivalent entities in clustersClusters are classified by the type of their entitiesIIIClusters of ClassesClusters of PropertiesClusters of IndividualsA new alignment can result inIIMerging of clustersA new subsumption relationship between clustersAt the beginning a cluster is created for each entitySavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200915 / 52
ILIADSThe AlgorithmThe AlgorithmStep by step01:02:03:04:05:06:07:08:09:10:11:Initialize algorithm’s structures (O is O1 O2 )repeat:Compute similarity scores between clustersHeuristically select a type of clustersfor each couple (c, c 0 ) of that type doDetermine a candidate relationship a(c,c 0 )Perform incremental inferenceUpdate similarity scoreSelect the best couple, update O and A until there are clusters with similarity λtreturn A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200916 / 52
ILIADSThe AlgorithmThe AlgorithmSimilarity ComputationEntities similarity scoresim(e, e0 ) λx · simlex (e, e0 ) λs · simstruct (e, e0 ) λe · simext (e, e0 )lexical: Jaro-Winkler (similar to edit distance) and thesauristructural: Jaccard for neighborhoods (Jacd (S1 , S2 ) S1 S2 ) S1 S2 extensional: Jaccard on extensionsThe set of parameters {λx , λs , λe } is different for each entity typeSimilarity between clusters is computed combining the similaritiesof their entitiesSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200917 / 52
ILIADSThe AlgorithmThe AlgorithmStep by step01:02:03:04:05:06:07:08:09:10:11:Initialize algorithm’s structures (O is O1 O2 )repeat:Compute similarity scores between clustersHeuristically select a type of clustersfor each couple (c, c 0 ) of that type doDetermine a candidate relationship a(c,c 0 )Perform incremental inferenceUpdate similarity scoreSelect the best couple, update O and A until there are clusters with similarity λtreturn A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200918 / 52
ILIADSThe AlgorithmThe AlgorithmStep by step01:02:03:04:05:06:07:08:09:10:11:Initialize algorithm’s structures (O is O1 O2 )repeat:Compute similarity scores between clustersHeuristically select a type of clustersfor each couple (c, c 0 ) of that type doDetermine a candidate relationship a(c,c 0 )Perform incremental inferenceUpdate similarity scoreSelect the best couple, update O and A until there are clusters with similarity λtreturn A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200919 / 52
ILIADSThe AlgorithmThe AlgorithmSelecting a relationshipIterates over each couple (c, c 0 ) such that sim(c, c 0 ) is above athreshold λtFor each of those a candidate relationship is chosen between:IEquivalence(Two concepts are said to be equivalent if they denote the same concept)ISubsumption(A concept subsumes another concept if it always denotes a superset of the second)The selection is done by looking at the intersection of entities’extensions:IIThe set of its instance individuals, for a classThe couples of individuals involved, for a propertySavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200920 / 52
ILIADSThe AlgorithmThe AlgorithmSelecting a relationship - Example (1)Music PieceSmoke onthe watersubClassOfsubClassOfMusicWhole ssOfsubClassOfsubClassOftypetypetypeRide of theValkyriesRequiem K626Iron MaidenplaysThe BlackSabbathperformstypecomposer-ofIron MaidenRequiem inD minorcomposer-ofTake the "A"TrainS. HarrisplaysViennaPhilharmonictypeSmoke onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypetypetypetypeThe Phantomof the ictypeQuadrant Fourcomposer-oftypeMusicArtisttypeplaysW.A. A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200921 / 52
ILIADSThe AlgorithmThe AlgorithmSelecting a relationship - Example (2)Music PieceSmoke onthe watersubClassOfsubClassOfMusicWhole ssOfsubClassOfRockMusicsubClassOfRequiem inD minorcomposer-oftypetypetypeRide of theValkyriesRequiem K626Iron MaidenTake the "A"TrainS. HarrisViennaPhilharmonicplaysThe e onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypetypetypetypeThe Phantomof the n MaidentypeQuadrant Fourcomposer-oftypeMusicArtisttypeplaysW.A. A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200922 / 52
ILIADSThe AlgorithmThe AlgorithmSelecting a relationship - Example (3)Music PieceSmoke onthe watersubClassOfsubClassOfMusicWhole petype2/3Requiem K626Iron MaidenTake the "A"TrainS. HarrisViennaPhilharmonicplaysThe e onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypeNumber of instances belongingtypetypeRideof the over theonly to "RockMusic"Valkyriestotal numberof instancesconsidered:Requiem inD minortypeThe Phantomof the n MaidentypeQuadrant Fourcomposer-oftypeMusicArtisttypeplaysW.A. A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200923 / 52
ILIADSThe AlgorithmThe AlgorithmSelecting a relationship - Example (4)Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200924 / 52
ILIADSThe AlgorithmThe AlgorithmSelecting a relationship - Example (5)Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200925 / 52
ILIADSThe AlgorithmThe AlgorithmStep by step01:02:03:04:05:06:07:08:09:10:11:Initialize algorithm’s structures (O is O1 O2 )repeat:Compute similarity scores between clustersHeuristically select a type of clustersfor each couple (c, c 0 ) of that type doDetermine a candidate relationship a(c,c 0 )Perform incremental inferenceUpdate similarity scoreSelect the best couple, update O and A until there are clusters with similarity λtreturn A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200926 / 52
ILIADSThe AlgorithmThe AlgorithmIncremental Logical InferenceThe inference step is used to:IIILook for inconsistencies of the candidate relationshipInfer logical consequences of the new axiomPossibly enforce the confidence in itThe inference is not complete (it would be E XPTIME in OWL Lite)IIOnly a small number of steps is actually performedHowever this may cause inconsistencies not to be foundSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200927 / 52
ILIADSThe AlgorithmThe AlgorithmIncremental Logical Inference - Example (1)Music PieceSmoke onthe watersubClassOfsubClassOfMusicWhole ssOfsubClassOfsubClassOftypetypetypeRide of theValkyriesRequiem K626Iron MaidenplaysThe BlackSabbathperformstypecomposer-ofIron MaidenRequiem inD minorcomposer-ofTake the "A"TrainS. HarrisplaysViennaPhilharmonictypeSmoke onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypetypetypetypeThe Phantomof the ictypecomposed-byQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200928 / 52
ILIADSThe AlgorithmThe AlgorithmIncremental Logical Inference - Example (2)Music PieceSmoke onthe watersubClassOfsubClassOfMusicWhole ssOfsubClassOfsubClassOfRequiem inD minorcomposer-oftypetypetypeRide of theValkyriesRequiem K626Iron MaidenTake the "A"TrainS. HarrisViennaPhilharmonicplaysThe BlackSabbathperformscomposer-ofplaystypeSmoke onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypetypetypetypeThe Phantomof the icsameAstypeIron Maidentypecomposed-byQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200929 / 52
ILIADSThe AlgorithmThe AlgorithmIncremental Logical Inference - Example (3)Music PieceSmoke onMusiccomposed-bythe waterinverseOf The Phantomof the OperatypeRide of theValkyriestypetypeplaysThe BlackSabbathperformsTake the "A"TrainS. HarrisViennaPhilharmonictypeSmoke onthe waterRequiem K626Iron Maidencomposer-ofplaysOrchestraltypeplaysThe Phantomof the quiem inD eWhole meAstypeIron Maidentypecomposed-byQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200930 / 52
ILIADSThe AlgorithmThe AlgorithmIncremental Logical Inference - Example (4)Music PieceSmoke onthe watersubClassOfsubClassOfMusicWhole ssOfsubClassOfsubClassOftypeRide of theValkyriesRequiem inD minorcomposed-bytypetypecomposer-ofRequiem K626Iron Maidencomposer-ofTake the "A"TrainS. HarrisViennaPhilharmonicplayssameAscomposer-ofThe BlackSabbathperformsplaystypeSmoke onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypetypetypetypeThe Phantomof the ictypeIron Maidentypecomposed-byQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200931 / 52
ILIADSThe AlgorithmThe AlgorithmIncremental Logical Inference - Example (5)Music PieceSmoke onthe watercomposed-byWhole sOfsubClassOfRockMusicsubClassOfThe Phantomof the OperatypeRide of theValkyriescomposer-oftypeParanoidtypetypeSmoke onthe waterRequiem K626Iron Maidencomposer-ofTake the "A"TrainS. omposer-ofThe ckplaysRequiem inD minorcomposed-bytypetypetypetypesubClassOfMetalThe Phantomof the peIron Maidentypecomposed-byQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200932 / 52
ILIADSThe AlgorithmThe AlgorithmIncremental Logical Inference - Example (6)Music PieceSmoke onthe watercomposed-byWhole sOfsubClassOfRockMusicsubClassOfThe Phantomof the OperatypeParanoidtypecomposer-ofRequiem K626composer-ofsameAsTake the "A"TrainS. HarrisViennaPhilharmonictypeSmoke onthe waterplayscomposer-ofThe BlackSabbathperformsplaysOrchestraltypeIron MaidentypeRide of m inD minorcomposed-bytypetypetypetypesubClassOfMetalThe Phantomof the peIron Maidentypecomposed-byQuadrant rformertypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200933 / 52
ILIADSThe AlgorithmThe AlgorithmStep by step01:02:03:04:05:06:07:08:09:10:11:Initialize algorithm’s structures (O is O1 O2 )repeat:Compute similarity scores between clustersHeuristically select a type of clustersfor each couple (c, c 0 ) of that type doDetermine a candidate relationship a(c,c 0 )Perform incremental inferenceUpdate similarity scoreSelect the best couple, update O and A until there are clusters with similarity λtreturn A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200934 / 52
ILIADSThe AlgorithmThe AlgorithmUpdate similarity scoreThis is the key point of the algorithmIA value f - the influence factor of the inference - is computedThe “f ” factorf Y(e1 ,e2 ) Qsim(e1 , e2 )1 sim(e1 , e2 )Q: the set of entity pairs that became equivalent as a consequence of inferencef is used to update the similarity score for the current coupleIsiminf (c, c 0 ) min(f · sim(c, c 0 ), 1)Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200935 / 52
ILIADSThe AlgorithmThe AlgorithmUpdate similarity score - Example (1)Music PieceSmoke onthe watersubClassOfsubClassOfMusicWhole ssOfsubClassOfsubClassOftypetypetypeRide of theValkyriesRequiem K626Iron ser-ofTake the "A"TrainS. HarrisplaysViennaPhilharmoniccomposer-ofThe BlackSabbathIron MaidenRequiem inD minortypeSmoke onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypetypetypetypeThe Phantomof the ictypecomposed-byQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200936 / 52
ILIADSThe AlgorithmThe AlgorithmUpdate similarity score - Example (2)Music PieceSimilarity :MusicSmoke onthe water0.5subClassOfsubClassOfWhole ssOfsubClassOfsubClassOftypetypetypeRide of theValkyriesRequiem K626Iron ser-ofTake the "A"TrainS. HarrisplaysViennaPhilharmoniccomposer-ofThe BlackSabbathIron MaidenRequiem inD minortypeSmoke onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypetypetypetypeThe Phantomof the ictypecomposed-byQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200937 / 52
ILIADSThe AlgorithmThe AlgorithmUpdate similarity score - Example (3)Music PieceSimilarity :MusicSmoke onthe water0.5subClassOfsubClassOfWhole ssOfsubClassOfsubClassOftypetypetypeRide of theValkyriesRequiem K626Iron ser-ofTake the "A"TrainS. ofThe BlackSabbathIron MaidenRequiem inD minortypeSmoke onthe waterplaysThe Phantomof the OperaOrchestraltypeParanoidtypetypetypetypeThe Phantomof the icPerformercomposed-byQuadrant FourtypeMusicArtisttypeplaysMozartSimilarity :typecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM070.6MusicianBilly CobhamSI-LS 200938 / 52
ILIADSThe AlgorithmThe AlgorithmUpdate similarity score - Example (4)Music PieceSimilarity :MusicSmoke onthe water0.5subClassOfsubClassOfWhole ssOfsubClassOfsubClassOftypetypeThe Phantomof the Operaf 0.6 / (1 - 0.6) 1.5Iron MaidentypeRequiem K626playstypecomposer-ofIron MaidenRequiem inD minorcomposed-bycomposer-ofS. ofThe BlackSabbathperformsTake the "A"TrainPerformertypeSmoke onthe waterplaysRide of The Phantomof the iccomposed-byQuadrant FourtypeMusicArtisttypeplaysMozartSimilarity :typecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM070.6MusicianBilly CobhamSI-LS 200939 / 52
ILIADSThe AlgorithmThe AlgorithmUpdate similarity score - Example (5)Music PieceSimilarity :MusicSmoke onthe water0.75subClassOfsubClassOfWhole ssOfsubClassOfsubClassOftypetypeThe Phantomof the Operaf 0.6 / (1 - 0.6) 1.5Iron MaidentypeRequiem K626playstypecomposer-ofIron MaidenRequiem inD minorcomposed-bycomposer-ofS. HarrisplaysViennaPhilharmoniccomposer-ofThe BlackSabbathperformsTake the "A"TraintypetypeSmoke onthe waterplaysRide of The Phantomof the iccomposed-byQuadrant rtypecomposer-ofsubClassOfW.A. MozarttypeSavioli, Reale, Sorbini (DEIS)WritersubClassOfUGM07MusicianBilly CobhamSI-LS 200940 / 52
ILIADSThe AlgorithmThe AlgorithmStep by step01:02:03:04:05:06:07:08:09:10:11:Initialize algorithm’s structures (O is O1 O2 )repeat:Compute similarity scores between clustersHeuristically select a type of clustersfor each couple (c, c 0 ) of that type doDetermine a candidate relationship a(c,c 0 )Perform incremental inferenceUpdate similarity scoreSelect the best couple, update O and A until there are clusters with similarity λtreturn A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200941 / 52
ILIADSThe AlgorithmThe AlgorithmBuilding the alignmentBy now for each candidate pair of clusters of a given typeIIIA possible relationship has been exploredA set of consequences has been inferredAn “inference-weighted“ similarity has been computedBefore restarting the loop:III The axiom a(c,c0 ) with the highest similarity score is chosenIt is added to the output alignment A 0If a(c,c0 ) is an equivalence c and c are mergedSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200942 / 52
ILIADSThe AlgorithmThe AlgorithmStep by step01:02:03:04:05:06:07:08:09:10:11:Initialize algorithm’s structures (O is O1 O2 )repeat:Compute similarity scores between clustersHeuristically select a type of clustersfor each couple (c, c 0 ) of that type doDetermine a candidate relationship a(c,c 0 )Perform incremental inferenceUpdate similarity scoreSelect the best couple, update O and A until there are clusters with similarity λtreturn A Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200943 / 52
ILIADSThe AlgorithmThe AlgorithmThe End(ing)The algorithm halts when there are no more candidate clustersIIWhen there are no clusters having similarity greater than thethreshold λtRemaining clusters are not likely to share any relationshipIntuitively ILIADS is guaranteed to terminate becauseIIPreviously used cluster pairs are not re-used unless their scorechangesThe merging process decreases the number of clustersSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200944 / 52
ILIADSExperimental ResultsComparative resultsPrecision, Recall and F-1Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200945 / 52
ILIADSExperimental ResultsTests AnalysisThe interplay between structure and instance integration delivershigher qualityIISignificant improvement of recallTests on ontologies without instance data showed comparableresults with the other systemsLambda-tuning allows ILIADS to adapt itself better to particularpairs of ontologiesInconsistent alignments due to limited inference steps were foundonly in the .5% of testsSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200946 / 52
ILIADSConclusions and Future DevelopmentsSummaryExplicit semantics provided by ontologies can improve thequality of data integrationInstance data exploitation could significantly enhance traditionalmatching techniquesILIADS leverages both these opportunities and shows promisingresultsFuture developmentsIntra-ontology differentiated λ parametersAlignment of ”distant” sections of the ontologies in parallelSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200947 / 52
DemoDemoAnd now an ultra-fancy demoSavioli, Reale, Sorbini (DEIS)UGM07SI-LS 200948 / 52
DemoThank you.Group 15A. SorbiniSavioli, Reale, Sorbini (DEIS)E. SavioliUGM07A. RealeSI-LS 200949 / 52
AppendixFor Further ReadingFor Further Reading IW3C OWL resources.http://www.w3.org/2004/OWL/.D. Aumueller, H.H. Do, S. Massmann, and E. Rahm.Schema and ontology matching with COMA .In Proceedings of the 2005 ACM SIGMOD internationalconference on Management of data, pages 906–908. ACM NewYork, NY, USA, 2005.I. Horrocks, P.F. Patel-Schneider, and F. van Harmelen.From SHIQ and RDF to OWL: the making of a Web OntologyLanguage.Web Semantics: Science, Services and Agents on the World WideWeb, 1(1):7–26, 2003.Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200950 / 52
AppendixFor Further ReadingFor Further Reading IIY. Kalfoglou and M. Schorlemmer.Ontology mapping: the state of the art.The knowledge engineering review, 18(01):1–31, 2003.E. Rahm and P.A. Bernstein.A survey of approaches to automatic schema matching.The VLDB Journal The International Journal on Very Large DataBases, 10(4):334–350, 2001.P. Shvaiko and J. Euzenat.A survey of schema-based matching approaches.Lecture Notes in Computer Science, 3730:146–171, 2005.Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200951 / 52
AppendixFor Further ReadingFor Further Reading IIIG. Stumme and A. Maedche.FCA-MERGE: Bottom-Up Merging of Ontologies.In INTERNATIONAL JOINT CONFERENCE ON ARTIFICIALINTELLIGENCE, volume 17, pages 225–234. LAWRENCEERLBAUM ASSOCIATES LTD, 2001.O. Udrea, L. Getoor, and R.J. Miller.HOMER: Ontology alignment visualization and analysis.F. Baader.The description logic handbook: theory, implementation, andapplications.Cambridge University Press, 2003.Savioli, Reale, Sorbini (DEIS)UGM07SI-LS 200952 / 52
Music Ride of the Valkyries Requiem in D minor Take the "A" Train The Phantom of the Opera type type Jazz Music Rock Music W.A. Mozart composer-of plays Performer performs composer-of Smoke on the water type Music Iron Maiden type Billy Cobham type composer-of plays plays The Phantom of the Opera Savioli, Reale, Sorbini (DEIS) UGM07 SI-LS 2009 .
Leveraging VMware Consolidated Backup for Disaster Recovery Leveraging VMware Consolidated Backup for Disaster Recovery September 10-13, 2007. Contents Contents
Our Story: Leveraging Open Source for an Enterprise CMS Solution 4 Carleton Content Management System Version 2.0 . OUR STORY: LEVERAGING OPEN SOURCE TO DEVELOP AN ENTERPRISE CMS 1. INTRODUCTION In conjunction with the Carleton University Web Services' presentation at the 2010 Ontario .
Simple and Compound Data Structures Simple Data Structure: Simple data structure can be constructed with the help of primitive data structure. A primitive data structure used to represent the standard data types of any one of the computer languages. Variables, arrays, pointers, structures, unions, etc. are examples of primitive data structures.
Discovering and Leveraging Visual Structure for Large-scale Recognition Abhinav Shrivastava CMU-RI-TR-17-63 Submitted in partial fulfillment of the requirements of the degree of Doctor of Philosophy in Robotics The Robotics Institute School of Computer Science Carnegie Mellon University August
Leveraging Big Data Analytics and Hadoop in Developing India's Healthcare Services D. Peter Augustine Assistant Professor Christ University Bangalore 560029 ABSTRACT In this paper, we analyze and reveal the benefits of Big Data Analytics and Hadoop in the applications of Healthcare where the data flow to and from is in massive volume. The
neric Data Modeling and Data Model Patterns in order to build data models for crime data which allows complete and consistent integration of crime data in Data Warehouses. Keywords-Relational Data Modeling; Data Warehouse; Generic Data Modeling; Police Data, Data Model Pattern existing data sets as well as new kinds of data I. INTRODUCTION The research about Business Intelligence and Data
Title: ER/Studio Data Architect 8.5.3 Evaluation Guide, 2nd Edition Author: Embarcadero Technologies, Inc. Keywords: CA ERwin data model software Data Modeler data modeler tools data modelers data modeling data modeling software data modeling tool data modeling tools data modeling with erwin data modelings data modeller data modelling software data modelling tool data modelling
tube in tube structure, braced frame structure, bundled tube structure, mega tube structure and outrigger frame system that can be used to enhance the lateral resisting capacity of tall buildings. 2 TYPE OF STRUCTURE 2.1 Frame tube Structure In this type of structure, the columns are placed on the periph-ery of the building with a core wall.
