dr hoang huu hanh ost hue university
play

Dr. Hoang Huu Hanh, OST - Hue University hanh-at-hueuni.edu.vn - PowerPoint PPT Presentation

Dr. Hoang Huu Hanh, OST - Hue University hanh-at-hueuni.edu.vn Clarification: Cl ifi ti What are Ontologies? Revisited: R i it d How we already have learned to express ontologies Web Ontology Language OWL: W b O l L


  1. Dr. Hoang Huu Hanh, OST - Hue University hanh-at-hueuni.edu.vn

  2.  Clarification: Cl ifi ti  What are Ontologies?  Revisited: R i it d  How we already have learned to express ontologies  Web Ontology Language ‐ OWL: W b O l L OWL  extending expressivity  Semantics of & Reasoning with OWL: f  using the extended expressivity 2

  3.  “Ontology” in Philosophy:  “The metaphysical study of the nature of being “Th t h i l t d f th t f b i and existence”  “Ontology” in Artificial Intelligence:  a shared and common understanding of some domain that can be communicated between people and application systems” – (Gruber) 3

  4. “Ontology (languages)” for the Semantic Web: “O t l (l )” f th S ti W b  We aim at a (XML ‐ based) language to formally describe concepts, instances, relations and axioms, i.e. data+structure t i t l ti d i i d t t t in order to enable machine ‐ processable reasoning on and exchange of data.  Knowledge representation, exchange, combination (inference of new knowledge!) 4

  5.  Concepts Classes + class hierarchy  Concepts: Classes + class ‐ hierarchy  instances  Properties: often also called “Roles” or “Slots”  labeled instance ‐ value ‐ pairs  labeled instance ‐ value ‐ pairs  Axioms/Relations:  relations between classes (disjoint, covers)  inheritance (multiple? defaults?) ( p )  restrictions on slots (type, cardinality)  Characteristics of slots (symm., trans., …)  reasoning tasks:  Classification: Which classes does an instance belong to?  Subsumption: Does a class subsume another one?  Consistency checking: Is there a contradiction in my axioms/instances? axioms/instances? 5

  6.  Web portal W b t l  ontology ‐ based portal for a community of users which have shared interest  Multi ‐ media collections M lti di ll ti  annotating, searching, ontological search instead of keyword search  Corporate Website p  knowledge management  Documentation  engineering & design  engineering & design  Agents & Services, Ubiquitous computing  Interoperability! 6

  7.  Clarification: Cl ifi ti  What are Ontologies?  Revisited: R i it d  How we already have learned to express ontologies  Web Ontology Language ‐ OWL: W b O l L OWL  extending expressivity  Semantics of & Reasoning with OWL: f  using the extended expressivity 7

  8.  RDF: triples for making assertions about resources RDF triples for making assertions abo t reso rces  RDFS extends RDF with “schema vocabulary”, e.g.:  Class, Property  type, subClassOf, subPropertyOf  range, domain  representing simple assertions, taxonomy + typing Vehicle subClassOf subClassOf Company SeaVehicle LandVehicle subClassOf subClassOf NumberOfEngines Hovercraft Number 8

  9. RDFS too weak to describe resources in sufficient detail: RDFS too weak to describe resources in sufficient detail:    No localised range and domain constraints ▪ Can’t say that the range of hasChild is person when applied to persons and elephant when applied to elephants  No existence/cardinality constraints ▪ Can’t say that all instances of person have a mother that is also a person, or that persons have exactly 2 parents  No transitive inverse or symmetrical properties No transitive, inverse or symmetrical properties ▪ Can’t say that isPartOf is a transitive property, that hasPart is the inverse of isPartOf or that touches is symmetrical  No in/equality ▪ Can’t say that a class/instance is the same as some other class/instance, C ’t th t l /i t i th th l /i t can’t say that somthe classes/instances are definitely disjoint/different.  No boolean algebra ▪ Can’t say that that one class is the union, intersection, complement of y p other classes, etc. 9

  10. ??? ??? ???  Semantics+reasoning  Semantics+reasoning OWL ?  Relational Data  Data Exchange D t E h 10

  11.  Clarification: Cl ifi ti  What are Ontologies?  Revisited: R i it d  How we already have learned to express ontologies  Web Ontology Language ‐ OWL: W b O l L OWL  extending expressivity  Semantics of & Reasoning with OWL: f  using the extended expressivity 11

  12.  Both use the same data model:  Both use the same data model: hasAuthor page.html “Dieter Fensel“ Resource Property Value (subject) (subject) (predicate) (predicate) (object) (object)  OWL extends vocabulary and adds axioms 12

  13. 13 RDF OIL DAML+OIL OWL DAML

  14. Two languages developed to satisfy above requirements T l d l d i f b i   OIL (Object Inference Layer): developed by group of (largely) European researchers (several from EU OntoKnowledge project)  DAML ‐ ONT: developed by group of (largely) US researchers (in DARPA DAML programme) Efforts merged to produce DAML+OIL   Development was carried out by “Joint EU/US Committee on Agent Markup Languages”  Extends (“DL subset” of) RDF DAML+OIL submitted to W3C as basis for standardisation   Web ‐ Ontology (WebOnt) Working Group formed  WebOnt group developed OWL language based on DAML+OIL g p p g g  OWL language now a W3C Recommendation (01.02.2004) 14

  15.  OWL Lite  (sub)classes, individuals  (sub)properties, domain, range RDF Schema  intersection  (in)equality (in)equality F ll Full  cardinality 0/1  datatypes  inverse, transitive, symmetric DL  hasValue h l  someValuesFrom  allValuesFrom Lite  OWL DL  OWL DL  OWL Full  OWL Full  Negation (disjointWith, complementOf)  Allow meta ‐ classes etc  unionOf  Full Cardinality Full Cardinality  Enumerated types (oneOf) 15

  16. Full Full Three species of OWL Three species of OWL    OWL DL stays in Description Logic fragment  OWL Lite is “easier to implement” subset of OWL DL DL  OWL Full is union of OWL syntax and RDF OWL Full is union of OWL syntax and RDF OWL DL based on SHIQ Description Logic  Lite  In fact it is equivalent to SHOIN (D n ) DL OWL DL Benefits from many years of DL research   Well defined semantics  Formal properties well understood (complexity, decidability)  Known reasoning algorithms  Implemented systems (highly optimised) OWL f ll h OWL full has all that and all the possibilities ll th t d ll th ibiliti  of RDF/RDFS which destroy decidability 16

  17. DLs are a family of logic based KR formalisms DLs are a family of logic based KR formalisms   Particular languages mainly characterized by:   Set of constructors for building complex concepts and roles from simpler ones  Set of axioms for asserting facts about concepts, roles and individuals Examples:   “Female persons” ▪ Person ⊓ Female  “Non ‐ female persons” Non female persons ▪ Person ⊓  Female  “Persons that have a child” ▪ Person ⊓  hasChild.Person  “Persons all of whose children are female” Persons all of whose children are female ▪ Person ⊓  hasChild.Female  “Persons that are employed or self ‐ eployed” ▪ Person ⊓ (Employee ⊔ SelfEmployed)  “Persons the have at most one father“ Persons the have at most one father ▪ Person ⊓ ≤ 1.hasFather 17

  18.  Inclusion axioms provide necessary conditions:  concept ⊑ definition  Equivalence axioms provide necessary and sufficient conditions: concept ≡ definition { { concept � definition and definition ⊑ concept 18

  19. Dom … Domain I … Interpretation T (T) I = Dom (a class which ANY legal instance is a member of: owl:Thing) ( ┴ ) I = {} ┴  C (  C) I = Dom \C I (C ⊓ D) I = C I  D I C ⊓ D (C ⊔ D) I = C I  D I C ⊔ D (  R.C) I = { x | (x,y)  R I  y  C I }  R.C (  R.C) I = { x | (x,y)  R I  y  C I }  R.C (  nR.C) I = { x | |{ y | (x,y)  R I  y  C I }|  n }  nR.C (  nR.C) I ={ x | |{ y | (x,y)  R I  y  C I }|  n }  nR.C ( ) { | |{ y | ( y) y }| } (=nR.C) I ={ x | |{ y | (x,y)  R I  y  C I }| = n } =nR.C (  nR) I = { x | |{ y | (x,y)  R I }|  n }  nR (  nR) I ={ x | |{ y | (x y)  R I }|  n }  nR  nR (  nR) { x | |{ y | (x,y)  R }|  n } (=nR) I ={ x | |{ y | (x,y)  R I }| = n } =nR 19

  20. in OWL in OWL 20

  21.  OWL allows greater expressiveness, but OWL ll t i b t  OWL (DL/Lite) puts certain constraints on the use of RDF  For instance: a class may not act as an instance of another (meta)class (the same holds for properties)  OWL has got it’s own Class identifier RDFS OWL < owl:Class rdf:ID="River"> < rdfs:Class rdf:ID="River"> <rdfs:subClassOf rdf:resource =" #Stream "/> <rdfs:subClassOf rdf:resource =" #Stream "/> </ rdfs:Class > </ owl:Class > 21

  22. What can you express in RDF/RDFS? What can you express in RDF/RDFS? Not too much… Employee � Person Employee � Person  … class hierarchy, necessary conditions (also rdfs:subClassOf equivalence is expressible because A � B and B � A  A ≡ B) Employee(axel)   … class membership class membership rdf t pe rdf:type … OWL provides more expressive constructs to express the DL features! 22

Recommend


More recommend