知识图谱梳理.pptx

《知识图谱梳理.pptx》由会员分享，可在线阅读，更多相关《知识图谱梳理.pptx（28页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、知识图谱架构知识图谱一般架构:来源自百度百科复旦大学知识图谱架构:早期知识图谱架构第1页/共28页知识图谱一般架构:来源自百度百科第2页/共28页第3页/共28页架构讨论数据检索预处理构建关系矩阵网络图谱参数调整可视化数据规范化处理结果导读早期知识图谱架构第4页/共28页知识抽取实体概念抽取实体概念映射关系抽取质量评估第5页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014A sampler of research problemsGrowth:

2、knowledge graphs are incomplete!Link prediction:add relationsOntology matching:connect graphsKnowledge extraction:extract new entities and relations from web/textValidation:knowledge graphs are not always correct!Entity resolution:merge duplicate entities,split wrongly merged onesError detection:rem

3、ove false assertionsInterface:how to make it easier to access knowledge?Semantic parsing:interpret the meaning of queriesQuestion answering:compute answers using the knowledge graphIntelligence:can AI emerge from knowledge graphs?Automatic reasoning and planningGeneralization and abstraction9第6页/共28

4、页关系抽取定义:常见手段:语义模式匹配频繁模式抽取,基于密度聚类,基于语义相似性层次主题模型弱监督第7页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014Methods and techniquesSupervised modelsSemi-supervised modelsDistant supervision2.Entity resolutionSingle entity methodsRelational methods3.Link pre

5、dictionRule-based methodsProbabilistic modelsFactorization methodsEmbedding models80Not in this tutorial:Entity classification Group/expert detection Ontology alignment Object ranking1.Relation extraction:第8页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August

6、 24,2014 Extracting semantic relations between sets of grounded entitiesNumerous variants:Undefined vs pre-determined set of relationsBinary vs n-ary relations,facet discoveryExtracting temporal informationSupervision:fully,un,semi,distant-supervisionCues used:only lexical vs full linguistic feature

7、s82Relation ExtractionKobeBryantLA LakersplayForthe franchise player ofonce again savedman of the match forthe Lakers”his team”Los Angeles”“Kobe Bryant,“Kobe“Kobe Bryant?第9页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014Supervised relation extract

8、ionSentence-level labels of relation mentionsApple CEO Steve Jobs said.=(SteveJobs,CEO,Apple)Steve Jobs said that Apple will.=NILTraditional relation extraction datasetsACE 2004MUC-7Biomedical datasets(e.g BioNLP clallenges)Learn classifiers from+/-examplesTypical features:context words+POS,dependen

9、cy path betweenentities,named entity tags,token/parse-path/entity distance83第10页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014Semi-supervised relation extractionGeneric algorithm(遗传算法)1.2.3.4.5.Start with seed triples/golden seed patternsExtract

10、patterns that match seed triples/patternsTake the top-k extracted patterns/triplesAdd to seed patterns/triplesGo to 2Many published approaches in this category:Dual Iterative Pattern Relation Extractor Brin,98Snowball Agichtein&Gravano,00TextRunner Banko et al.,07 almost unsupervisedDiffer in patter

11、n definition and selection86第11页/共28页founderOfKDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014Distantly-supervised relation extraction88Existing knowledge base+unlabeled text generate examplesLocate pairs of related entities in textHypothesizes that the

12、 relation is expressedGoogle CEO Larry Page announced that.Steve Jobs has been Apple for a while.Pixar lost its co-founder Steve Jobs.I went to Paris,France for the summer.GoogleCEOcapitalOfLarryPageFranceAppleCEOPixarSteveJobs第12页/共28页Distant supervision:modeling hypothesesTypical architecture:1.Co

13、llect many pairs of entities co-occurring in sentences from text corpus2.If 2 entities participate in a relation,several hypotheses:1.All sentences mentioning them express it Mintz et al.,09“Barack Obama is the 44th and current President of the US.”(BO,employedBy,USA)89KDD 2014 Tutorial on Construct

14、ing and Mining Web-scale Knowledge Graphs,New York,August 24,2014第13页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014Sentence-level featuresLexical:words in between and around mentions and their parts-of-speech tags(conjunctive form)Syntactic:depen

15、dency parse path between mentions along withside nodesNamed Entity Tags:for the mentionsConjunctions of the above featuresDistant supervision is used on to lots of data sparsity of conjunctiveforms not an issue92第14页/共28页Distant supervision:modeling hypothesesTypical architecture:1.Collect many pair

16、s of entities co-occurring in sentences from text corpus2.If 2 entities participate in a relation,several hypotheses:1.2.All sentences mentioning them express it Mintz et al.,09At least one sentence mentioning them express it Riedel et al.,10“Barack Obama is the 44th and current President of the US.

17、”(BO,employedBy,USA)“Obama flew back to the US on Wednesday.”(BO,employedBy,USA)95KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014第15页/共28页Distant supervision:modeling hypothesesTypical architecture:1.Collect many pairs of entities co-occurring in sente

18、nces from text corpus2.If 2 entities participate in a relation,several hypotheses:1.2.3.All sentences mentioning them express it Mintz et al.,09At least one sentence mentioning them express it Riedel et al.,10At least one sentence mentioning them express it and 2 entities can expressmultiple relatio

19、ns Hoffmann et al.,11 Surdeanu et al.,12“Barack Obama is the 44th and current President of the US.”(BO,employedBy,USA)“Obama flew back tothe US justWednesday.”said.”employedBy,USA)98KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014was born in on he alway

20、s(BO,(BO,bornIn,第16页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014Distant supervisionProsCan scale to the web,as no supervision requiredGeneralizes to text from different domainsGenerates a lot more supervision in one iterationConsNeeds high qual

21、ity entity-matchingRelation-expression hypothesis can be wrongCan be compensated by the extraction model,redundancy,language modelDoes not generate negative examplesPartially tackled by matching unrelated entities101第17页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New

22、 York,August 24,2014104KobeBryantGasolteammatebornInplayInLeagueBlackMambaEntity resolutionLA LakersplayForplayForPau35ageKobe B.BryantVanessaL.BryantmarriedTo1978Single entityresolutionRelational entityresolution第18页/共28页DEF:DEF:We consider the entity resolution(ER)problem(also We consider the enti

23、ty resolution(ER)problem(also known as deduplication,or mergepurge),in which known as deduplication,or mergepurge),in which records determined to represent the same real-world records determined to represent the same real-world entity are successively located and entity are successively located and

24、mergedmerged the problem of extracting,the problem of extracting,matchingmatching and and resolving resolving entity mentions in structured and entity mentions in structured and unstructuredunstructured datadataMethodsMethodsEntity resolution/deduplication Multiple mentions of the same entity is wro

25、ng and confusing.第19页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014Single-entity entity resolutionEntity resolution without using the relational context of entitiesMany distances/similarities for single-entity entity resolution:Edit distance(Leve

26、nshtein,etc.)Set similarity(TF-IDF,etc.)Alignment-basedNumeric distance between valuesPhonetic SimilarityEquality on a boolean predicateTranslation-basedDomain-specific105第20页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014Relational entity resolut

27、ion Simple strategies Enrich model with relational features richer context for matchingRelational features:Value of edge or neighboring attributeSet similarity measuresOverlap/JaccardAverage similarity between set membersAdamic/Adar:two entities are more similar if they share more items that areover

28、all less frequentSimRank:two entities are similar if they are related to similar objectsKatz score:two entities are similar if they are connected by shorter paths114KobeBryant1978teammatebornInplayForplayInLeagueBlackMambaLA LakersplayFor35agePauGasol第21页/共28页KDD 2014 Tutorial on Constructing and Mi

29、ning Web-scale Knowledge Graphs,New York,August 24,2014KobeBryant1978teammatebornInplayForplayInLeagueBlackMambaLA LakersplayFor35agePauGasolRelational entity resolution Advanced strategiesDependency graph approaches Dong et al.,05Relational clustering Bhattacharya&Getoor,07Probabilistic Relational

30、Models Pasula et al.,03Markov Logic Networks Singla&Domingos,06Probabilistic Soft Logic Broecheler&Getoor,10115第22页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014LINK PREDICTION116第23页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Kno

31、wledge Graphs,New York,August 24,2014117KobeBryantLink predictionNY KnicksPauGasolteammateplayInLeagueteamInLeagueopponentplayForLA LakersplayFor Add knowledge from existing graph No external source Reasoning within the graph1.Rule-based methods2.Probabilistic models3.Factorization models4.Embedding

32、 models第24页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014First Order Inductive Learner FOIL learns function-free Horn clauses:118Gasolgiven positive negative examples of a concepta set of background-knowledge predicatesFOIL inductively generates

33、a logical rule for the concept that cover all+and no-LALakersplayForplayForPauteammate(x,y)playFor(y,z)playFor(x,z)teammateKobeBryant Computationally expensive:huge search space large,costly Horn clauses Must add constraints high precision but low recall Inductive Logic Programming:deterministic and

34、 potentially problematic第25页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,August 24,2014S(KB,playFor,LAL)iplayForh(pai(KB,LAL)ipathsPath Ranking Algorithm Lao et al.,11119LALakersplayForPauGasolplayForteammateKobeBryantRandom walks on the graph are used to sam

35、ple pathsPaths are weighted with probability of reaching target from sourcePaths are used as ranking experts in a scoring functionNYKnicksplayInLeagueteamInLeagueopponenth(Pa2(KB,LAL)=0.2h(Pa1(KB,LAL)=0.95第26页/共28页KDD 2014 Tutorial on Constructing and Mining Web-scale Knowledge Graphs,New York,Augus

36、t 24,2014Link prediction with scoring functionsA scoring function alone does not grant a decisionThresholding:determine a threshold(KB,playFor,LAL)is True iff120S(KB,playFor,LAL)Ranking:The most likely relation between Kobe Bryant and LA Lakers is:rel argmaxrrelsS(KB,r,LAL)The most likely team for Kobe Bryant is:obj argmaxeentsS(KB,playFor,e)As prior for extraction models(cf.Knowledge Vault)No calibration of scores like probabilities第27页/共28页感谢您的观看！第28页/共28页