9 3 Karan Gosal

Uploaded from authorPOINT Lite
Download as
 PPT
Presentation Description 

No description available

authorSTREAM Premium Service
What's up on authorSTREAM?
Views: 39
Like it  ( Likes) Dislike it  ( Dislikes)
Added: November 28, 2007 This Presentation is Public 
Presentation Category : Entertainment All Rights Reserved
Presentation Transcript

RDFS(FA) Connecting RDF(S) and OWL DL: RDFS(FA) Connecting RDF(S) and OWL DL Jeff Z. Pan and Ian Horrocks


Agenda: Agenda Introduction Issues Background RDF(S) OWL RDF – OWL (Mismatch ?) RDFS(FA) Rules of Thumb Conclusion


What it’s all about : What it’s all about The Semantic Web is about two things. It is about common formats for integration and combination of data drawn from diverse sources, where on the original Web mainly concentrated on the interchange of documents. It is also about language for recording how the data relates to real world objects. http://www.w3.org/2001/sw/


How it’s done: How it’s done What we have: Intelligent agent’s Resources A world Wide Web Consortium Semantic Markup


The Foundations: The Foundations RDF and RDFS. i.e RDF(S) RDFS constructors are primitive OIL +DML OWL OWL Lite OWL DL OWL Full


The Problems:: The Problems: Too Few entailments Contradictions classes Size of the universe Current Approaches Limit the extension of RDFS to a property related subset of Owl with weaker semantics Limit the semantic interconnection


Requirements of an RDF(S) sublanguage:: Requirements of an RDF(S) sublanguage: Ontologies should be RDF graphs Should enable Class URIrefs as property values Metamodelling architecture comparable with UML Semantics should be compatible with the semantics of OWL DL.


Interpretation of V={a,b,c}:: Interpretation of V={a,b,c}: P.Hayes “RDF Semantics” Domain (or universe) of I


Cont..: Cont.. Given a triple [s p o] I([s p o]) = True if s,p,o є V , IS(p) є IP and є IEXT(p); Basically this leads us to the following : “All properties are resources”


A thought :: A thought : Similar deductions lead us to the following thoughts On the one hand, rdfs:Resource is an instance of rdfs:Class; on the other hand, rdfs:Class is a subclass of rdfs:Resource. Hence, Is rdfs:Resource an instance of its subclass?


Mismatch Between RDF(S) and OWL DL:: Mismatch Between RDF(S) and OWL DL: Entailment : Does the following individual axiom Individual(ex:John type(intersectionOf (ex: Student ex: Employee ex:European))) Entail the individual axiom Individual(ex:John type(intersectionOf (ex: Student ex:European))) ?


Contradiction Classes and the size of the Universe: Contradiction Classes and the size of the Universe Designing Contradiction classes takes a lot of time and there is no guarantee that all possible contradiction classes have been excluded from the principles of OWL Full. Individual (elp:Ganesh type (elp:Elephant)) DisjointClasses (elpElephant:elp:Plant)


RDFS(FA): RDFS(FA) Fixed Layered Metamodelling Architecture Addresses the Following :- RDF triples have built-in semantics. Classes and properties, including built-in classes and properties of RDF(S) and its subsequent languages such as OWL, are treated as objects (or resources) in the domain. There are no restrictions on the use of built-in vocabularies.


UML Like (No of Strata = 4): UML Like (No of Strata = 4)


URIrefs of RDFS(FA): URIrefs of RDFS(FA) The built-in class URIrefs of RDFS(FA) are fa : Resource i + 1; fa : Class i+2; fa : Property i+2 ; fa : AbstractProperty i+2; fa : DatatypeProperty; fa : AnnotationProperty:


RDFS(FA) interpretation: RDFS(FA) interpretation Objects Object Pairs Classes


Rules of Thumb on Strata numbers :: Rules of Thumb on Strata numbers : Subscripts represent the stratum they are in. The first rule of thumb is that the subscripts of built in RDFS(FA) vocabulary represent exactly the stratum that they are in. For example, fa : Resource 1  stratum 1 and fa : Class 2  stratum 2.


Second Rule of thumb: Second Rule of thumb Let [s p o .] be an RDFS(FA) axiom. The second rule of thumb is that if p is an instance-of relationship, then o is one stratum higher than s, and p is in the same stratum as o. For example, in the axiom [elp : Ganesh fa : Type1 elp : Elephant :]; elp : Ganesh is an object in stratum 0 (Instance Layer), elp: Elephant, and fa : Type1 are one stratum higher, i.e in stratum 1. s p o Stratum 1 Stratum 2 Ganesh Type1 , Elephant


Third Rule of Thumb: Third Rule of Thumb Let [s p o .] be an RDFS(FA) axiom. The third rule of thumb is that if p is not an instance-of relationship, then s and o should be in the same stratum, and p should be one stratum higher than s and o. For Example, in the axiom [elp : Elephant fa : subClassOf2 elp : Animal :]; elp: Elephant and elp: Animal are in stratum 1 and fa : subClass is in stratum 2; s,o p Elephant , Animal SubClassOf


Fourth Rule of Thumb: Fourth Rule of Thumb Let [s p o .] be an RDFS(FA) axiom, and s, p, and o in strata i, j, and k, respectively. The stratum number of the axiom [s p o .] is min (i; j; k), i.e., the smallest stratum number among those of s, p, and o.


OWLObject Class Diagram: OWLObject Class Diagram


Conclusion: Conclusion RDFS(FA) is syntactically and semantically compatible with OWL DL. Common framework provided Follows the stated requirements Possibility of OWL – FA. Open Problems: Knowledge Access Query Answering in RDFS(FA)


References:: References: http://www.w3.org/2001/sw/ Jeff Z. Pan, Ian Horrocks. RDFS(FA): Connecting RDF(S) and OWL DL. IEEE Transactions on Knowledge and Data Engineering. 19(2): 192-206 (2007) . Jeff Z. Pan, Ian Horrocks : RDFS(FA): A DL-ised Sub-language of RDFS F. Baader and W. Nutt, “Basic Description Logics,” The Description Logic Handbook: Theory, Implementation, and Applications, pp. 43-95, 2003. P. Hayes, “RDF Semantics,” technical report, W3C recommendation, W3C, http://www.w3.org/TR/rdf-mt/, Feb 2004. D. Brickley and R.V. Guha, “RDF Vocabulary Description Language 1.0: RDF Schema,” Feb. 2004 Bernd Brügge, Ralf Pfleghar, Thomas Reicher:OWL: An Object-Oriented Framework for Intelligent Home and Office Applications


Thank You: Thank You