/
HowNAGAUncoils:SearchingwithEntitiesandRelationsGjergjiKasneciMax-Plan HowNAGAUncoils:SearchingwithEntitiesandRelationsGjergjiKasneciMax-Plan

HowNAGAUncoils:SearchingwithEntitiesandRelationsGjergjiKasneciMax-Plan - PDF document

conchita-marotz
conchita-marotz . @conchita-marotz
Follow
364 views
Uploaded On 2015-11-23

HowNAGAUncoils:SearchingwithEntitiesandRelationsGjergjiKasneciMax-Plan - PPT Presentation

1InthemythologiesofHinduismandBuddhismNAGAisahugesnakeHereitstandsforthesizeanddiversityoftheWeb extractedbyLEILA5Asampleofthisknowledgebaseisshownbelow 3QUERYANDANSWERMODELInthespiritofthee ID: 203021

1InthemythologiesofHinduismandBuddhism NAGAisahugesnake.Here itstandsforthesizeanddiversityoftheWeb. extractedbyLEILA[5].Asampleofthisknowledge-baseisshownbelow: 3.QUERYANDANSWERMODELInthespiritofthee

Share:

Link:

Embed:

Download Presentation from below link

Download Pdf The PPT/PDF document "HowNAGAUncoils:SearchingwithEntitiesandR..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.


Presentation Transcript

HowNAGAUncoils:SearchingwithEntitiesandRelationsGjergjiKasneciMax-Planck-InstitutSaarbruecken/Germanykasneci@mpii.mpg.deFabianM.SuchanekMax-Planck-InstitutSaarbruecken/Germanysuchanek@mpii.mpg.deMayaRamanathMax-Planck-InstitutSaarbruecken/Germanyramanath@mpii.mpg.deGerhardWeikumMax-Planck-InstitutSaarbruecken/Germanyweikum@mpii.mpg.de 1InthemythologiesofHinduismandBuddhism,NAGAisahugesnake.Here,itstandsforthesizeanddiversityoftheWeb. extractedbyLEILA[5].Asampleofthisknowledge-baseisshownbelow: 3.QUERYANDANSWERMODELInthespiritoftheexamplequeryintheintroductionwepresentataxonomyofqueriessupportedbyNAGA.Basi-cally,aNAGAqueryisadirectedgraphG(V;E)whereVisasetofpossiblylabelednodesandEasetofpossiblylabelededges.EVIDENCEQUERIES(EQ)Anevidencequeryisaconnecteddirectedgraphthenodesandlabelsofwhicharelabeled. Inthiscase,theuserwantstoknowwhetherthereisanyevidenceforacertainhypothesis.DISCOVERYQUERIES(DQ)Adiscoveryqueryisaconnecteddirectedgraphthenodesandedgesofwhichmaybeunlabeled. Here,theuserwantstodiscoverpiecesofmissinginforma-tionsuchasentitiesorrelations,representedbyunlabelednodesoredgesrespectively.Theabovequeryisthequeryexamplefromtheintroduction.RELATEDNESSQUERIES(RQ)Arelatednessqueryisaconnecteddirectedgraphthenodesandedgesofwhichmaybeunlabeledandatleastoneoftheedgesislabeledwitharegularexpressionoverrelationshiplabels. Inthissetting,theuserisinterestedinthebroadrelationbetweenpiecesofinformation.Obviously,itholdsEQDQEQ.TheanswertoaqueryisasubgraphAoftheknowledgegraphthatmatchesthequery.Forexample,NAGAcouldanswertheabovediscoveryqueryasfollows: Thenumbersontheedgesofananswergraphrepresentthecertaintiesofthefacts.Inordertocomputetheover-allcertaintyofananswer,thecertaintiesoftheedgeshavetobeaccumulated.Wethinkofthecertaintyvalueastheprobabilitythatafactiscorrect.Sincethefactshavebeenextractedindependently,theprobabilitythatthecompleteanswerAiscorrectisjusttheproductofthecertaintiesof theedges:c(A)=YeedgeinAc(e)Incasetherearemultipleanswerstoaquery,wereturnthetop-kanswersrankedbytheircertainty.4.PRELIMINARYRESULTSNAGAisimplementedinJavaontopofanOracledatabase.ThequeryenginetranslatestheuserqueriesintoSQLandperformsgraphsearchestosolverelatednessqueries.WeconductedseveralpreliminaryexperimentstoassessthequalityofNAGA'sanswers.Below,wepresentsomesam-plequeriestoillustratethedi erencebetweenNAGAandGoogle: Query Google NAGA WhenwasBBCNewsestablished? Answerfoundonfollowinglink. 1922. WhichotherphysicistswereborninthesameyearasAlbertEinstein? Fails.GiveslotsoflinkstoEinstein'sbiography. vonLaue,Pfund,Burton,andseveralothers. WhatistheconnectionbetweenEinsteinandBohr? LinkstothedebatebetweenBohrandEinsteinonquantumtheory. Botharescientists.ThereareMooncratersandas-teroidbeltsnamedafterthem.TomCruisecon-nectsthembybeingaveg-etarian(asEinstein)andbybeingbornin1962(whenBohrdied). WhatistheconnectionbetweenIndiraGandhiandMargaretThatcher? Linksofeventsinvolvingthetwowomen Botharefemaleheadsofgovernment,bothat-tendedOxford,bothwerepoliticiansinEnglish-speakingcountries(IndiaandUK,respectively) 5.CONCLUSIONANDOUTLOOKWepresentedNAGA,anewsemanticsearchenginewithaquerylanguagethatcanexpresscomplexqueriesbymeansofgraphstructuresandregularexpressionsoverrelations.Forfuturework,weplantoexploitlogicalinferencesintheknowledgegraphandtopredictnotonlythecorrectness,butalsotheinterestingnessofanswers.NAGAcanbetriedoutonlineat:http://www.mpii.mpg.de/~suchanek/naga. 6.BLAH 7.REFERENCES [1] M.Cafarella,C.Re,D.Suciu,andO.Etzioni.Structuredqueryingofwebtextdata:Atechnicalchallenge.InCIDR,2007. [2] SaraCohen,JonathanMamou,YaronKanza,andYehoshuaSagiv.Xsearch:Asemanticsearchengineforxml.InVLDB,pages45{56,2003. [3] J.Graupmann,R.Schenkel,andG.Weikum.Thespheresearchengineforuni edrankedretrievalofheterogeneousXMLandwebdocuments.InVLDB,2005. [4] V.Kacholia,S.Pandit,S.Chakrabarti,S.Sudarshan,R.Desai,andH.Karambelkar.Bidirectionalexpansionforkeywordsearchongraphdatabases.InVLDB,2005. [5] FabianM.Suchanek,GeorgianaIfrim,andGerhardWeikum.Combininglinguisticandstatisticalanalysistoextractrelationsfromwebdocuments.InKDD,2006. [6] FabianM.Suchanek,GjergjiKasneci,andGerhardWeikum.Yago:ACoreofSemanticKnowledge.In16thinternationalWorldWideWebconference(WWW2007),NewYork,NY,USA,2007.ACMPress. [7] MaxVolkel,MarkusKrotzsch,DennyVrandecic,HeikoHaller,andRudiStuder.Semanticwikipedia.InWWW,2006.