China09SummerSchool7講座6：本體管理與推理II

1、語義網(wǎng)的邏輯基礎(chǔ) Logical Foundation of the Semantic WebZhisheng HuangVrije University Amsterdam， The Netherlands助教： 胡偉 Wei HuSoutheast University 課程時間表Schedule海量語義數(shù)據(jù)推理(Web Scale Reasoning)LarKC:海量語義數(shù)據(jù)推理平臺（LarKC: A Platform for Web Scale Reasoning)結(jié)論和討論 (Conclusion and Discussion) 講座6：本體管理與推理(II)Lecture 6: O

2、ntology Management and Reasoning (II)海量語義數(shù)據(jù)Linking Open Data： In October 2007, datasets consisted of over two billion RDF triples, which were interlinked by over two million RDF links. By May 2009 this had grown to 4.2 billion RDF triples, interlinked by around 142 million RDF links.1 triple:107 Tri

3、plesOWLIMSuez CanalDenny Vrandei AIFB, Universitt Karlsruhe (TH) 9 RDF Store subsecond querying108 TriplesIngentaMoonDenny Vrandei AIFB, Universitt Karlsruhe (TH) 10 109 TriplesEarthDenny Vrandei AIFB, Universitt Karlsruhe (TH) 11 LarKC proposal 1010 Triples 1 triple per web 1 triple per webJupiterD

4、enny Vrandei AIFB, Universitt Karlsruhe (TH) 12 1011 TriplesDistance Sun Pluto Fensel / Harmelen estimate1014 Triples1014 TriplesDenny Vrandei AIFB, Universitt Karlsruhe (TH) 14 LarKC: 一個海量語義數(shù)據(jù)處理平臺The Large Knowledge Collider (大型知識對撞機）A configurable platformfor experimentationby others可布局平臺“Configur

5、able platform”“a configurable platform for infinitely scalable semantic web reasoning”.Enrich current logic-based Semantic Web reasoning with methods from information retrieval, machine learning, information theory, databases, and probabilistic reasoning網(wǎng)絡(luò)科學(xué)與人類智能科學(xué)的結(jié)合Web Science with Human Intellige

6、nceEmploying cognitively inspired approaches and techniques such as spreading activation, focus of attention, reinforcement, habituation, relevance reasoning, and bounded rationalityAchieve scalability through giving up completeness by giving up 100% correctness: trading quality for size often compl

7、eteness is not needed sometimes even correctness is not neededprecision (soundness)recall (completeness)logicIRSemantic Web通過并行計算達(dá)到海量數(shù)據(jù)處理能力Achieve Scalability through Parallelization by parallelisation: cluster computing wide area distribution “Thinkinghome”, “self-computing semantic Web” cloud comp

8、uting 云計算 (Amazon， Google)歐盟第七框架研究課題: LarKCEU 7th framework Project總預(yù)算1千萬歐元：10M budget 歷時3年半： 3.5 years八十個人年： 80 person years3個實例研究： 3 case studies14個合作單位： 14 partners,來自12個國家： 12 countries,來自3大洲： 3 continentsproject nr. FP7 215535The consortium50 people presentThe ConsortiumCombining consortium com

9、petenceIR, CognitionML, OntologiesStatistics, ML, Cognition,DBLogic,DB, Probabilistic InferenceEconomics, Decision Theory課題組成Project WorkpackagesWP1 Conceptual Framework & EvaluationWP 2: Retrieval and SelectionWP5: Collider PlatformWP 9: Exploitation and standardsWP 10: Project ManagementWP 8: Trai

10、ning, dissemination, community buildingWP3: Abstraction and LearningWP4: Reasoning and DecidingWP 6: Use case: Real Time CityWP 7a: Use case: Early Clinical DevelopmentWP 7b: Use case: Carcinogenesis Reference ProductionUse case: Drug Discovery Problem: pharmaceutical R&D in early clinical developme

11、nt is stagnating(Q1Q2Q3)FDA white paper Innovation or Stagnation (March 2004):“developers have no choice but to use the tools of the last century to assess this centurys candidate solutions.” “industry scientists often lack cross-cutting information about anentire product area, or information about

12、techniques that may be used in areas other than theirs”“Show me any potential liver toxicity associated with the compounds drug class, target, structure and disease.”Show me all liver toxicity associated with the target or the pathway.Genetics“Show me all liver toxicity associated with compounds wit

13、h similar structure”Chemistry“Show me all liver toxicity from the public literature and internal reports that are related to the drug class, disease and patient population”LITERATURECurrent NCBI: linking but no inferenceUse Case: Real Time CityOur cities face many challenges Urban Computing is the I

14、CT way to address themHow can we redevelop existing neighborhoods and business districts to improve the quality of life? How can we create more choices in housing, accommodating diverse lifestyles and all income levels?How can we reduce traffic congestion yet stay connected?How can we include citize

15、ns in planning their communities rather than limiting input to only those affected by the next project?How can we fund schools, bridges, roads, and clean water while meeting short-term costs of increased security?Is public transportation where the people are?Which landmarks attract more people?Where

16、 are people concentrating?Where is traffic moving?課題時間表Project Timeline Surveys (plugins, platform) Requirements (use cases)Prototype Internal Release Public Release Final ReleaseUse Cases V1Use Cases V2Use Cases V34206183310如果你對參與開發(fā)感興趣的話How can any other interested party contribute?The Large Knowle

17、dge Collider is an open, and configurable platform.The first public version of the Large Knowledge Collider is available。LarKC will form an early adapters group. LarKC will actively support this group in use the Large Knowledge Collider platform. LarKC 中文論壇Chinese Developer ForumLarKC Chinese Develo

18、per Workshop 2010 will be located with Chinese Web Intelligence Forum 2010 The official page form the 2010 Chinese Developer Forum: 語義操作系統(tǒng)Semantic Operating Systems云計算 Cloud Computing 云計算是指把信息固定儲存在網(wǎng)絡(luò)服務(wù)器上，而只是臨時性給客戶端提供所需信息的一種計算范式， 即把軟件和IT基礎(chǔ)設(shè)施當(dāng)作一種服務(wù)對外提供。Cloud Computing is a paradigm in which informatio

19、n is permanently stored in servers on the Internet and cached temporarily on clients。Realising the ArchitecturePipelineSupportSystemPlug-in RegistryPlug-in ManagerData LayerPlug-in APIData Layer APIRDFStore32LarKC Plug-in API: General Plug-in ModelPlug-ins are identified by a URI (Uniform Resource I

20、dentifier)Plug-ins provide MetaData about what they do (Functional properties): e.g. type = SelecterPlug-ins provide information about their behaviour and needs, including Quality of Service information (Non-functional properties): e.g. Throughput, MinMemory, Cost, + URI getIdentifier()+ QoSInformat

21、ion getQoSInformation() Plug-inFunctional propertiesNon-functional propertiesWSDL descriptionPlug-in description33LarKC Plug-in API: IDENTIFYIDENTIFY: Given a query, identify resources that could be used to answer itSindice Triple Pattern Query RDF GraphsGoogle Keyword Query Natural Language Documen

22、tTriple Store SPARQL Query RDF Graphs+ Collection identify(Query theQuery, Contract contract, Context context) Identifier 34LarKC Plug-in API: TRANSFORM (1/2)Query TRANSFORM: Transforms a query from one representation to another SPARQL Query Triple Pattern QuerySPARQL Query Keyword QuerySPARQL Query

23、 SPARQL Query (different abstraction)SQARQL Query CycL Query+ Set transform(Query theQuery, Contract theContract, Context theContext)QueryTransformer35LarKC Plug-in API: TRANSFORM (2/2)Information Set TRANSFORM: Transforms data from one representation to anotherNatural Language Document RDF GraphStr

24、uctured Data Sources RDF Graph RDF Graph RDF Graph (e.g. foaf vocabulary to facebook vocabulary)+ InformationSet transform(InformationSet theInformationSet, Contract theContract, Context theContext)InformationSetTransformer36LarKC Plug-in API: SELECTSELECT: Given a set of statements (e.g. a number o

25、f RDF Graphs) will choose a selection/sample from this setCollection of RDF Graphs Triple Set (Merged)Collection of RDF Graphs Triple Set (10% of each)Collection of RDF Graphs Triple Set (N Triples)+ SetOfStatements select(SetOfStatements theSetOfStatements, Contract contract,Context context)Selecte

26、r37LarKC Plug-in API: REASONREASON: Executes a query against the supplied set of statementsSPARQL Query Variable Binding (Select)SPARQL Query Set of statements (Construct)SPARQL Query Set of statements (Describe)SPARQL Query Boolean (Ask)+ VariableBinding sparqlSelect(SPARQLQuery theQuery, SetOfStat

27、ements theSetOfStatements, Contract contract, Context context)+ SetOfStatements sparqlConstruct(SPARQLQuery theQuery, SetOfStatements theSetOfStatements, Contract contract, Context context)+ SetOfStatements sparqlDescribe(SPARQLQuery theQuery, SetOfStatements theSetOfStatements, Contract contract, C

28、ontext context)+ BooleanInformationSet sparqlAsk(SPARQLQuery theQuery, SetOfStatements theSetOfStatements, Contract contract, Context context)Reasoner38LarKC Plug-in API: DECIDEDECIDE: Builds the pipeline and manages the control flowScripted Decider: Predefined pipeline is built and executedSelf-con

29、figuring Decider: Uses plug-in descriptions (functional and non-functional properties) to build the pipeline+ VariableBinding sparqlSelect(SPARQLQuery theQuery, QoSParameters theQoSParameters)+ SetOfStatements sparqlConstruct(SPARQLQuery theQuery, QoSParameters theQoSParameters)+ SetOfStatements spa

30、rqlDescribe(SPARQLQuery theQuery, QoSParameters theQoSParameters)+ BooleanInformationSet sparqlAsk(SPARQLQuery theQuery, QoSParameters theQoSParameters)Decider39Released SystemDeciderPlug-in APIPlug-in ManagerQueryTransformerPlug-in APIPlug-in ManagerIdentifierPlug-in APIPlug-in ManagerInfo. SetTran

31、sformerPlug-in APIPlug-in ManagerSelecterPlug-in APIPlug-in ManagerReasonerPlug-in APIPlug-in RegistryPipelineSupportSystemEarly adopters workshop ESWC09， ISWC09participants modified plug-ins, modified workflowsStandard Open Environment: Moving to Sourceforge/SVN40Data Layer APIPipelineSupportSystem

32、Plug-in RegistryRDFStoreRDFStoreRDFStoreRDFDocRDFDocData LayerDeciderPlug-in APIPlug-in ManagerQueryTransformerPlug-in APIPlug-in ManagerIdentifierPlug-in APIPlug-in ManagerInfo. SetTransformerPlug-in APIPlug-in ManagerSelecterPlug-in APIPlug-in ManagerReasonerPlug-in APIApplicationRDFDocPlatform Ut

33、ility FunctionalityAPIsPlug-insExternal systemsExternal data sourcesLarKC Architecture41LarKC Plug-ins Plug-in ManagerQueryTransformerPlug-in APIPlug-in ManagerIdentifierPlug-in API Provide SPARQL end-points Run in separate threads Automatically add meta-data to registry when loaded Communicate RDF

34、data by passing labelled sets or references to labelled sets Parallelisation in progressPlug-in ManagerTransformerPlug-in APIPlug-in ManagerIdentifierPlug-in APIransformerTransformerTransformerPlug-in ManagerIdentifierPlug-in APIPlug-in ManagerIdentifierPlug-in APIPlug-in ManagerSelectorPlug-in APIP

35、lug-in ManagerSelectorPlug-in API Split/Join connectors in progress42DIG Pipeline LocalPlug-in ManagerSimple IdentifierPlug-in APILocalPlug-in ManagerDIGReasonerPlug-in APIDIGDeciderSPARQLResultSPARQL Query43Gate-Cyc DeciderLocalPlug-in ManagerSimple FileReader IdentifierPlug-in APILocalPlug-in Mana

36、gerGate TransformerPlug-in APILocalPlug-in ManagerCycSelectorPlug-in APILocalPlug-in ManagerCycReasonerPlug-in APISPARQLResultSPARQL Query44Gate and Cyc PipelineUrban ComputingUrban CityDeciderSPARQLResultSPARQL QueryLocalPlug-in ManagerSPARQL to GeoQuery TransformerPlug-in APILocalPlug-in ManagerSP

37、ARQL to GeoQuery TransformerPlug-in APILocalPlug-in ManagerGeo LocationIdentifierPlug-in APILocalPlug-in ManagerGeo LocationIdentifierPlug-in APILocalPlug-in ManagerGrowing Data Set SelectorPlug-in APILocalPlug-in ManagerPathFindingReasonerPlug-in APILocalPlug-in ManagerSPARQLEndpointIdentifierPlug-

38、in API45LarKC推理插件一覽Jena ReasonerSPARQL-DL ReasonerPellet ReasonerDIG Interface ReasonerOWLLink ReasonerIRIS rule-based ReasonerPION ReasonerGranular ReasonerStream ReasonerCYC DeciderScripted Decider(s).46Implementation of Reasoner Plug-insReleased:Pellet SPARQL-DL reasonerDIG Interface reasonerIRIS

39、 Rule-based reasoner In Progress:PION reasoner Stream reasoner OWLAPI Reasoner47Using the DIG Plug-in to Reason with Ontologies in LarKC MotivationLarKC Platform & DIG Plug-inInstallation and ExecutationTest ExamplesDeveloper GuideUse Case: PION for Reasoning with Inconsistent OntologiesMotivationAl

40、l popular DL-reasoners such as RACER, FACT+, Pellet, KAON2) provide the DIG interface support.The LarKC platform needs DL/OWL reasoning support. It is convenient for the platform to gain the DL reasoning support via the LarKC DIG interface.To provide an easy approach to wrap non-java-based reasoners

41、 (such as PION, MORE, DION, etc.) with their DIG interface.本體詢問語言：DIG Interface描述邏輯實現(xiàn)組(The Description Logic Implementation Group) 于1999年提出提供了基于描述邏輯的本體描述與查詢操作語言，也為OWLDL提供了一個對應(yīng)于描述邏輯推理機的查詢語言 采用基于HTTP接口的類似SOAP的方法 DIG Interface使用TELL語句結(jié)合其概念描述語言來描述本體，使用ASK語句來對DIG描述的本體進(jìn)行查詢。此外，還提供了包含了推理機標(biāo)識詢問等附加的功能。Concept

42、LanguageTell LanguageExampleResponseAsk LanguageExample：AskResponse LanguageExample： ResponseLarKC Platform and the DIG plug-inLarKC PlatformDIG InterfacePlug-inRacerFACT+KAON2Tasks of the DIG Plug-inTranslate a set of statements (ontology data) into a DIG data. If it is OWL-DL data, the use the OWL

43、2DIG library to translate it into a DIG dataTranslate SPARQL(DL) query into DIG - deal with triple-encoded DL expressions3. Query processing and answer checking4. Translate DIG answers into SPARQL answers61footer14/09/2022LarKC Platform and the DIG plug-inLarKC PlatformDIG InterfacePlug-inExternalDI

44、G ReasonerOntology (URI)/Set of StatementsTellSPARQL queryAskResponseSPARQL AnswerThe DIG plug-in (v0.3)Have been supported Support the DIG interface 1.1.Support Sparqlask and Sparqlselect.DL Expressions (conjunction, disjunction, disjoint, negation)DIG queries (subsumption, instance, instances)Have

45、 been tested withRacer1.7.14PION 2.1.0To be supported soon:Complex DL concept expressions (such as nominal, min, max, etc.) Complex Sparql expressions (such as Filtering, Optional, Regular expressions, sparqlconstruct, sparqldescribe, etc.) Complex DIG queries (role query, functional query, value pa

46、ir query)Why SPARQL-DL?SPARQL is too expressive for a DL reasoner can support.In SPARQL, there is no semantic interpretation for DL expressions such as owl:sameas, owl:disjointwith, etc.SPARQL-DL is a DL-specific SPARQL with some DL primitives, such as type(a, C), SubClassof(C1, C2), DisjointWith(C1

47、,C2), ComplementOf(C1,C2),EquivalentClass(C1,C2),(Sirin and Parsia 2007)Translation of DL expressions into RDF triples Using the OWL-DL method (Patel-Schneider,Hayes, Horrocks 2004).SPARQL-DL Query Example 1?- subClassOf(Wine, PotableLiquid)/ to ask whether or not wine is a subclass of potable liqui

48、d PREFIX rdfs: PREFIX wine: PREFIX food: ASK WHERE wine:Wine rdfs:subClassOf food:PotableLiquid.SPARQL-DL Query Example 2?- subClassOf(Bordeaux, and(SweetWine, TableWine)/ to ask whether or not Bordeaux is a SweetWine and TableWine PREFIX rdfs: PREFIX rdf: PREFIX owl: ASK wine:Bordeaux rdfs:subClass

49、Of _:x. _:x owl:interSectionOf _:y1. _:y1 rdf:first wine:SweetWine. _:y1 rdf:rest wine:TableWine. wine:Bordeaux rdf:type owl:Class.Simple SPARQLSelect Query: Example 3?- subClassOf(?X, Wine)/ to list all subconcepts of Wine PREFIX rdfs: PREFIX wine: SELECT ?X WHERE ?X rdfs:subClassOf wine:Wine.SPARQ

50、L-DL Query Example 4?- subClassOf(Bordeaux, ?X), subClassOf(?X,Wine),subClassOf(?X,?Y). PREFIX rdfs:. . PREFIX wine: SELECT ?X ?Y WHERE wine:Bordeaux rdfs:subClassOf ?X. ?X rdfs:subClassOf wine:Wine. ?X rdfs:subClassOf ?Y. ?Y rdf:type owl:Class. Setting up LarKC with a DIG reasonerDownload the LarKC

51、 platform fromThe wrapped up DIG reasoner plugin is located at: plugins/src/eu/larkc/plugin/reason/dig External DIG ReasonerThe LarKC DIG plug-in requires an external DIG reasoner like Racer. Before starting the test, make sure the external DIG reasoner has been installed at your computer(i.e., loca

52、lhost) and is running at a known port.Test ExamplesThe java program DIGReasonerTest.java at the dig plug-in source directory provides several typical examples how an external DIG reasoner can be called to reason with ontologies at the LarKC platform.Checking the settingBefore executing the test prog

53、ram, you can change the following setting in the program: String hostname= localhost; int port = 8080; /default port for racer String path = /;Claiming ontology dataThe ontology data can be claimed by a code, like this:String ontology = ;or String ontology = ;if the ontology data is located at the l

54、ocal harddisk. Using the test utility programUsing the following test utility to post the query to the external DIG reasoner and get the answer in the test program: ReasonerTest(ontology, query32, hostname, port, path);Developer Guide: Reasoner Interfacepublic interface Reasoner extends Plugin publi

55、c VariableBinding sparqlSelect(SPARQLQuery theQuery,SetOfStatements theSetOfStatements, Contract contract, Context context);public SetOfStatements sparqlConstruct(SPARQLQuery theQuery,SetOfStatements theSetOfStatements, Contract contract, Context context);public SetOfStatements sparqlDescribe(SPARQL

56、Query theQuery,SetOfStatements theSetOfStatements, Contract contract, Context context);public BooleanInformationSet sparqlAsk(SPARQLQuery theQuery,SetOfStatements theSetOfStatements, Contract contract, Context context);To convert a string query into a SPARQLQuerySPARQLQuery sparqlQuery = DataFactory

57、.INSTANCE.createSPARQLQuery(query); To convert an ontology url into a SetOfStatementsRdfGraph graph = DataFactory.INSTANCE.createRemoteRdfGraph(new URIImpl(url.toString();Namely, using an openrdf graph as a SetOfStatements.To claim an external DIG reasoner DIGReasoner reasoner = new DIGReasoner();re

58、asoner.hostname = hostname;reasoner.port = port;reasoner.path = path;To conduct a reasoning taskBooleanInformationSet answer = reasoner.sparqlAsk(sparqlQuery, graph, contract, context);Ignoring the contract and the context for the time being.PIONPION is a system which can get meaningful answers for querying on inconsistent ontologies.PION website:The LarKC DIG plugin requests the external PION system (version 2.1.0 or higher) with SWI-Prolog on your computer. The SWI-Prolog can be downloaded from