1 / 35

On Integrating Rules into the Semantic Web

Jan Maluszynski Linköping University Sweden. On Integrating Rules into the Semantic Web. Overview. “Semantic Web” – do we need it? XML and web languages Rules on the web: the RuleML approach The W3C vision of the Semantic Web Towards typed rules Summary. Do we need ”Semantic Web”?.

avent
Télécharger la présentation

On Integrating Rules into the Semantic Web

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Jan Maluszynski Linköping University Sweden On Integrating Rules intothe Semantic Web WFLP'03, Valencia

  2. Overview • “Semantic Web” – do we need it? • XML and web languages • Rules on the web: the RuleML approach • The W3C vision of the Semantic Web • Towards typed rules • Summary WFLP'03, Valencia

  3. Do we need ”Semantic Web”? • Google query: FLP ? gives ~170 000 answers: Forever Living Products Family Living Program Forest Legacy Program Finish Line Performance … I admit, the query is not smart… • Functional Logic Programming Conference ? Michael Hanus Conferences related to FLP …WFLP … WFLP'03, Valencia

  4. Adding semantics to the contents • Annotate web content with a description of its meaning => metadata FLP – is a branch of CS Workshop is a Conference Symposium is a Conference • Use metadata for reasoning/querying FLP conference / CS? => WFLP, FLOPS • Logical foundations? • Various web languages needed: how to define them? what is their semantics? WFLP'03, Valencia

  5. Rules • Can a software agent plan my travel to WFLP ?Constraints/preferences expressed as rules? • Database views • Bussiness rules • … Using them on the web? WFLP'03, Valencia

  6. Web needs standards • Standard notation for data: XML • Standard for creating unique names: Namespaces • Standard for defining web languages (sets of XML elements): XML Schema, (DTD) • … WFLP'03, Valencia

  7. XML person 15 <person age=“25”> <name>John Smith </name> <address> <street> Dalav. 21 </street> <city> Lund </city> </address> </person> name address street city John Smith Dalav. 21 Lund Tags, Attributes, • Standard format facilitates: • Definition of self-describing structured data • Integration from various sources WFLP'03, Valencia

  8. Is XML element a term ? <CD price="10.90$" year="1985"> Empire Burlesque <artist>Bob Dylan</artist> <country>USA</country> <company>Columbia</company> </CD> CD[&{price[10.90$],year[1985]}, Empire Burlesque, artist[Bob Dylan], country[USA], company[Columbia] ] Arity of function symbols is not predefined;unordered subterms WFLP'03, Valencia

  9. Creating unique names • Universal Resource Identifier: URI identifies unique resource special case: URL • Used for creating unique names Namespaces The URL may point to a description of the names (unique meaning?). <contact xmlns:work="http://www.liu.ida.se/"> <name>John Smith</name> <phone>013/175347</phone> <work:phone>013/281400</work:phone> </contact> WFLP'03, Valencia

  10. Defining a web language: syntax • Language: a set of XML elements • Elements ~ trees Define language as a tree-grammar (DTD), place on web could we define language in XML? • XML Schema language: elements are XML documents for any XML Schema document S: • The meaning of S is a set of XML documents L(S) the function L is the semantics of XML Schema • standard on-line validator can check whether any given XML document is in L(S); • The validator implements the semantics of the XML Schema WFLP'03, Valencia

  11. A flavour of XML Schema <xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema"> < xsd: element name=“purchaseOrder” type=“POT” /> < xsd: complexType name=“POT”/> < xsd: sequence> < xsd: element name=“addr” type=“USAddress”/> < xsd: element name=“items” type=“Items”/> < /xsd: sequence> < xsd: attribute name=“orderDate” type=“xsd:date”/> < /xsd: complexType> …. < / xsd: schema> Roughly reflected by a rule: POT  purchaseOrder(USAddress Items) WFLP'03, Valencia

  12. Regular tree grammars for XML languages ERTG: Nonterminals X in N{POT, USAddr,…} Terminals t{pOrder, addr,…}/XML tags Start S in NPOT Rules X t ( R) R - a regular expression over N <pOrder> <addr> ... </addr> <items> <it> ... </it> <it> ... </it> </items> </pOrder> POT   pOrder( USAddr Items) USAddr  addr( Mail Street ZIP City State) Items  items( Item*) Item  it(Artno Quantity) …. pOrder addr… items it it … art … qua … WFLP'03, Valencia

  13. Regular tree languages • RTG: each function symbol has fixed arity ERTG: arguments of a function symbol defined by a regular expression • Regular tree language: a language defined by an RTG • Regular tree languages are closed under: • intersection • union • … Inclusion is decidable • The same holds for languages defined by ERTG WFLP'03, Valencia

  14. Defining web languages: summary • XML schema S defines syntax of L L=L(S) • Saccessible by a URL • A document x can be validated wrt S x  L ? XML Schema validators • Operational semantics: an interpreter for L accessible on the web • Declarative semantics ? WFLP'03, Valencia

  15. RuleML: rules on the web • Rule Markup Language Initiative: => Standardize mark-up for selected rule languages! A small family sufficient to embed various applications (?) • Develop XML schema for XML encoding of a rule language L e.g. for Datalog • Put it on the web: URL • Operational semantics: rule engine(s) • Declarative semantics, XML code XML code XML schema for L L Rules in syntax 1 L Rules in syntax 2 WFLP'03, Valencia

  16. Logic on the web • Define XML syntax for logic languages (different logics) Model-theoretic semantics, Automatic reasoners • Proposed logics include: • Description logics (work in progress DAML+OIL, OWL) • Datalog…. In a chosen logic: • Provide a set of formulae (encoded in XML) specific axioms • Automatic reasoning system answers queries WFLP'03, Valencia

  17. W3C view of the Semantic Web • Layers of tightly connected langugages. XML provides the syntactic basis for all layers • What are the connections between the layers? Rules Ontology RDF Schema RDF XML + NS + XML Schema URI Unicode WFLP'03, Valencia

  18. RDF + RDF Schema • A (still evolving) notation encoded in XML [Beckett D. (ed.) 2002] • For modeling application domains Classes, Subclasses, Properties(Slots), Literals • Basic construct: sentence:Subject Predicate Object • Encoded in XML, • Can be seen as ground atomic formula, • Represented as graph • Model theoretic semantics [P. Hayes (ed.) 2003] • Editors, query tools exist WFLP'03, Valencia

  19. A flavour of RDF Schema rdfs:Resource s = rdfs:subClassOf t = rdf:type s t s t rdfs:Class t xyz:MotorVehicle t t t s s s xyz:Van xyz:Truck t s xyz:PassengerVehicle xyz:MiniVan s WFLP'03, Valencia

  20. Description Logics • A family of KR formalisms, based on FOPL decidable, supported by automatic reasoning systems • Used for modelling of application domains • Classification of concepts and individuals concepts (unary predicates), subconcept (subsumption), roles (binary predicates), individuals (constants), constructors for building concepts, equality … [Baader et al. 2002] WFLP'03, Valencia

  21. R atomic role, A atomic concept C,D  A | (atomic concept) T | (universal concept)  | (bottom concept) A | (atomic negation) C  D | (intersection) R.C |(value restriction)  R.T | limited ex. Quantification Add more constructors: another DL RI I I AI  I TI = I I = Ø (A)I = I \ AI (CD)I = CIDI ( R.C)I = {a  I|b.(a,b) RIb CI } ( R.T)I = {a  I| b.(a,b) RI} Example DL: AL WFLP'03, Valencia

  22. Modelling with DLs • A set of equational axioms T(terminology, T-box)Woman  Person Female Man  Person  Female • Automated check for (C, D concepts): • Satisfiability of C: is there a model I of T such that CI nonempty? Man  Woman ? • Subsumption: CI DI for every model I of T ? Reduces to unsatisfiability C  D • Add assertions about individuals: Man(Peter) … Query the obtained knowledge base K. Ground membership query: Person(Peter)? Find all the individuals in K that are instances of a given concept WFLP'03, Valencia

  23. W3C Ontology Language OWL • An expressive description logic Under development [Dean et al. (eds) 2002] • Encoded in XML • Extending (?) RDFS Some problems with that [Patel-Schneider, Fensel 2002] Ontology: a conceptual model of the application domain; OWL ontology: a set of OWL statements – XML encoding of a set of DL axioms WFLP'03, Valencia

  24. Class : Atomic concept Property: Atomic role A is a subclass of B: A  B P is a subproperty of R: P  R A is a range of R : T   R.A A is a domain of R T   R-.A RDF Schema primitives in DL perspective A,B atomic concepts P,R atomic roles, T universal concept ( R.C)I = {a  I|b.(a,b) RIb CI } R- inverse WFLP'03, Valencia

  25. Ontologies and XML Schema • Ontology: a conceptual model of the application domain • Data will be represented in XML; document structure must be imposed • A systematic approach for translating an ontology into a specific XML Schema [Klein, Fensel et al. 2000]: • represent: a class C by an element definition; • the properties of C correspond to subelements • subclasses modelled by XML Schema extension • multiple inheritance cannot be directly expressed WFLP'03, Valencia

  26. Modeling Ontology with RT Grammar ? Person  p(S (M|W)? T?) TaxP  p(S (M|W)? T) M  p( sex(m) W? T?) W  p( sex(f) M? T?) T  tax(#Pcdata) S  sex(m|f) Man  Person Woman Person Taxpayer  Person S T sex partner tax WFLP'03, Valencia

  27. Adding rule level • Should build-up on the ontology level How ontologies relate to XML data on the web? • What kind of rules? Datalog, (extensions: negation, prioritized rules, fuzzy,…) Query rules for XML-data ….. • What kind of integration? How to add rules on top of ontologies (RDFS, DLs) ? Integrate Description Logic(s) with Datalog WFLP'03, Valencia

  28. Integration of Datalog with DL (1) CARIN [Levy, Rousset 1998], a family of languages Two sets of axioms are provided: (1) T-box in a description logic (2)Rules and ground facts: p1(X1),…,pn(Xn)  q(Y) Xi Y vectors of variables • concept predicates, role predicates, can be used in the body but not in the head; the other: ordinary predicates • all variables of the head appear in the body; no/limited recursion • Ground atomic queries only; special inference procedure needed/provided WFLP'03, Valencia

  29. Integration of Datalog with DL (2) Description Logic Programs [Grosof et al. 2003] • A restricted DL language, where DL axioms can be transformed into logically equivalent Horn clauses., e.g. PQ  Q(X) P(X) • The syntactic restrictions are somewhat complex • DLP is less expressive than DAML+OIL but can model a subset of RDFS The advantage: use of LP engines for query answering; non-monotonic extensions possible WFLP'03, Valencia

  30. Typing Datalog with DLs • Rationale: An ontology describes classes of application data; Rules describe data manipulation; Man  Person Woman Person Taxpayer  Person lives(X,M), couple(X,Y), interestpaid(Y,Z)  taxdeduction(X,Z,M) TYPES: lives: Person, Municipality couple:Man,Woman interestpaid:Person, Amount taxdeduction: Taxpayer, Amount, Municipality X:Person  X:Man Check in DL : (Man  Person)  Taxpayer NO! Clear separation of reasoning in DL from answering Datalog queries; WFLP'03, Valencia

  31. Typechecking of rules T(q1) T(qn) T(r)  T(r(X …Y)) q1(X … Y) … qn(X … Y) r(X … Y) T1(X)…T1(Y) Tn(X)…Tn(Y) T(X)…T(Y) T(X) = T1(X)…Tn(X) T(Y) = T1(Y)…Tn(Y) WFLP'03, Valencia

  32. Conceptual vs. Concrete XML encoding Ontology XML Schema typed typed Datalog rules XML rules ? WFLP'03, Valencia

  33. Towards a typed XML rule language • Query and transformation rule: q1,…, qn c inproc[[ author[X],title[Y] ]], article[[ author[X],title[Y] ]]  multi[X] [[…]] – allows matching term to include additional subterms Xcerpt: Query and Transformation Language [Bry, Schaffert 2002] • Types: approximate the XML databaseby a regular tree language T : • Matching qi ,c against T gives T(qi ), T(c) Matching : intersection of regular tree languages • T(qi ) binds variables: T(Xi ) • Compute T(X)=T(X1 ) …T(Xn), construct T’(c) • Check T’(c)  T(c) WFLP'03, Valencia

  34. Summary • A perspective of the ongoing development • Research issues: • Integration of rules and ontologies • Deriving XML schemata from ontologies • Type checking of rules • Datalog rules typed by DLs • XML transformation rules typed by RTLs derived from XML schemata WFLP'03, Valencia

  35. References F.Baader, D. Calvanese, D. McGuiness, D.Nardi and P. Patel-Schneider (eds) The Description Logic Handbook. Cambridge University Press, 2002. D. Beckett (ed.) RDF/XML Syntax Specification (Revised) W3C November 2002. F.Bry and S. Schaffert. Towards a Declarative Query and Transformation Language for XML and Semistructured Data. Proc. ICLP2002 H.Comon, M.Dauchet, R.Gilleron, F.Jacquemard, D.Lugiez, S.Tison and M. Tommasi. Tree Automata Techniques and Applications. http://www.grappa.univ-lille3.fr/tata M.Dean, D.Connolly, F.van Harmelen, J.Hendler, I.Horrocks, D.L.McGuiness, P.F.Patel-Scheider and L.A.Stein (Eds.). OWL Web Ontology Language, Reference Version 1.0, W3C,Working Draft November 2002 B. N. Grosof, I. Horrocks, R.Volz and S.Decker. Description Logic Programs: Combining Logic Programs with Description Logic. Proc. of WWW2003, Budapest. P.Hayes (ed.) RDF Semantics, W3C Working Draft January 2003 M. Klein, D. Fensel, F. van Harmelen andI. Horrocks, The relation between ontologies and schema-languages: translating OIL specifications in XML Schema.Proc of the Workshop on Ontologies and Problem Solving Methods, ECAI 2000 A.Levy and M-C.Rousset. CARIN: A Representation Language Combining Horn rules and Description Logics. Artificial Intelligence 104(1-2), 1998, 165-209. P.F. Patel-Schneider and D.Fensel. Layering the Semantic Web: Problems and Directions. Proc. ISWC2002, LNCS2342 pp. 16-29 Rule Markup Language Initiative. http://www.ruleml.org WFLP'03, Valencia

More Related