Assadadaad

Semantic Technologies for Knowledge Modeling - 2

He Tan 2014-12-09

Web Ontology Languages (OWL)

OWL

Gives more expressivity Based on Description Logics (DLs) e.g. Mothers that have many children are women who have more than 3 children.

RDFS

Classes and properties Express more generic knowledge e.g. Mothers are female If someone has daughter then that person is parent

RDF

Based on subject-predicate-object triples Allows to express facts e.g. Anna is the mother of Erik

Description Logics (DLs)



First Order Logic is not decidable



This result was established independently by Alonzo Church and Alan Turing in 1936 and 1937

  

Decidable fragments of FOL Satisfiability is decidable and, if possible, of low complexity Well suitable for representation of and reasoning about ontologies  Well defined (model theoretic) semantics  Formal properties well understood (complexity, decidability)  Known reasoning algorithms  Implemented systems (highly optimized)

Pellet

DL Syntax



Describe domain in terms of :



Concepts (classes)





Roles (binary relationships, attributes)



e.g. hasChild, livesIn, hasPet, Birthday



Individuals (instances/objects)



e.g. john, http://example.org/ishockey



e.g. Person, Doctor, ,

DL Semantics



An interpretation function

 



,

is the domain (a set) is a mapping that maps:

Concept name A  subset of  Role name R  binary relation over  Individual name i  element of

concept individual role

.

Knowledge Base



Knowledge base is Tbox + Abox

 

Tbox: concept level Abox: instance level

.

DL family



Smallest propositionally closed DL



boolean: conjunction

Woman  Person Parent  Mother

, disjunction

Female Father

, negation 

ChildlessPerson  Person   Parent

 

qualified restriction: , 

 hasChild.Person hasChild.Female Man ⊑ Person

sub-superclass relationships:

Subsumption

...