资源描述：
《ZaiQing_Nie20130708》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、Web-ScaleEntityMiningandCrowdsourcingZaiqingNieMicrosoftResearchAsia“BigDataPerspectives”WorkshopRenminUniversityofChina,Beijing,July8,2013HeterogeneousEntityKnowledgeinBigDataApplicationsSemiScientificPapers-StructuredelementswithinaknowledgeblockEntityItemsProductWebpagesBorninLansing,Mic
2、h.,HugheswasanadcopywriterandcreativedirectoratLeoBurnettinChicagobeforebreakingthroughwiththescreenplaystothe1983hits"Mr.Mom"and"NationalLampoon'sVacation."ThenworkingintheIllinoisFilmOffice,AlCohnrememberstheyoungHughescominginandwantingtolaunchhisdirectingcareerintheChicagoarea.NewsBioVi
3、rtualCharactersNaturallanguagetextwithinaknowledgeblock2BuildaBigEntityGraph•EntityKnowledgeinBigDataApplications–WebPages•StructuredHTMLelements•NaturalLanguageTextParagraphs–StructuredDatabases•Wikipedia,IMDB,Freebase,ACMDL,etc–Documents•PDFs,WordDocuments,PPT,eBook,etc•Question:Canweextr
4、actandintegrateallentityknowledgeandconstructaBigEntityGraph?High-QualityStartingwithPeople-CentricEntitiesWhatisEntityGraph•Nodes–EntityPageswithProfileInformation•Links–Relationships(entity-entity,entity-web,entity-social)WhyBuildingEntityGraph•ANewKnowledgeGraph–ComplementarytoWebGraph
5、andSocialGraph•AValuableKnowledgeAssetforBigDataAnalytics–LinkingtheHeterogeneousDataforBetterSearchandRetrieval•EntityQueries•FactQueries–EnablingEntitySearchandKnowledgeBrowsinginBigDataScenariosDeployedEntityMiningandSearchEngines•FullyAutomated:Web-ScaleEntityExtractionandIntegration–Li
6、braAcademicSearch(2004)•AcademicEntityGraphfromResearchPapers•NowMicrosoftAcademicSearch–Renlifang(2008)•People-CentricEntityGraphfrom1BChineseWebpages–EntityCube(2009)•People-CentricEntityGraphfrom3BEnglishWebpages•Crowdsourcing:InteractiveKnowledgeMining–AuthorFeedbackinMicrosoftAcademicS
7、earch(2010)–Renlifang2.0:GuanxiWiki(2011)–20QuestionsGame(2011)SystemOverviewWebCosmos&CrawlerDataFeeds(UnstructuredData)ACM,IEEE,IMDBClassification(StructuredData)EntityExtractionDataFeedParserEntityIntegration&NameDisambiguationWebEntityStoreStaticRank
