1. Semantic Web Ontologies & Data Models
CS 502 –
Carl Lagoze – Cornell University
Acknowledgements: Eric Miller Dieter Fensel
Cornell CS 502

2. Components of the Semantic Web
Cornell CS 502

3. Knowledge Representation and AI meeting the Web
Knowledge representation is not new
Formalisms
Semantic networks
Frame systems
Frames (objects)
Slots (attributes and properties)
Languages
LOOM
Classic
Cyc-L
Logics
Description Logics (DLs)
Cornell CS 502

4. So why re-invent it for the Web
Not re-invention
Same underlying formalisms (frames, slots, description logic)
But new factors
Lack of central control
Inconsistency, lies, re-interpretations, duplications
Continue change
New facts appear and modify constantly
Massive scale
Tractability
Knowledge expressiveness must be limited or reasoning must be incomplete
Open world context
Contrast to most reasoning systems that assume anything absent from knowledge base is not true
Need to build on existing standards
URI, XML, RDF
Cornell CS 502

5. Why aren’t XML schema good enough?
XML schema is designed to primarily describe structure and constraints on documents
Rich datatypes
Grammar for describing structure of elements (and attributes)
Incomplete modeling of is-a relationships
Top-down inheritance of attributes from superclasses to subclasses with extension and restriction is clumsy
E.g., student as person with a student # and age<28
Bottom-up inheritance of instances to superclasses is not automatic
Multiple inheritance is missing
Cornell CS 502

6. From modeling to implementation
Modeling primitives:
Class;
Slot
ERmodel
OWLspecification
Modeling
Implementation
Relational Model
XMLschema
Data-
base
XML
documents
Cornell CS 502

7. Data model for defining machine-processible semantics of data
Brief review of RDF
Data model for defining machine-processible semantics of data
Three main object types:
Resources
Primitive metadata is URI
Properties
Sub-class of resource
Statements
(:s :p :o)
Graph model that is serializable in XML
Cornell CS 502

8. RDF Schemas A building block for assembling ontology languages
Declaration of vocabularies
Classes and class hierarchies
properties defined by a particular community
characteristics of properties and/or constraints on corresponding values
Provides substructure for inferences based on existing triples
Schema language is an expression of basic RDF model
Schema Type System - Basic Types
Property, Class, SubClassOf, Domain, Range
Minimal (but extensible) at this time
Expressible in the RDF model and syntax
Cornell CS 502

9. Schema Vocabularies
Enables communities to share machine readable tokens and locally define human readable labels.
dc:Creator
“Nom”
rdfs:label
“Author”
rdfs:label
“$100 $a”
rdfs:label
Cornell CS 502

10. Relationships among vocabularies
dc:Creator
marc:100
ms:director
bib:Author
Cornell CS 502

11. Relationships among vocabulary elements
dc:Creator
ms:director
rdfs:
subPropertyOf
rdfs:label
“Director”
dc:Creator
URI:R
ms:director
“John Smith”
Cornell CS 502

12. RDF Schema: Specializing Properties
rdfs:subPropertyOf – allows specialization of relations
E.g., the property “father” is a subPropertyOf the property parent
subProperty semantics
Cornell CS 502

13. Sub-Property Semantics
Cornell CS 502

14. Constraints on Properties
Force objects to be of a certain type
rdfs:domain
Restricts the type of resources that may have a specific property
rdfs:range
Restricts the type of resources that may be the value of a specific property
range
Cornell CS 502

15. Inferences from Constraints
doris
betty
eve
alice
charles
Cornell CS 502

16. Inferences from Constraints
Cornell CS 502

17. Class Hierarchy rdfs:Class rdfs:subClassOf
Resources denoting a set of resources; range of rdf:type
rdfs:subClassOf
Create class hierarchy
rdf:type
rdf:type
rdfs:class
rdfs:subClassOf
rdf:type
rdf:type
rdf:class
rdf:class
Cornell CS 502

18. Sub-Class Inferencing
Cornell CS 502

19. Sub-class Inferencing Example
Cornell CS 502

20. A formal specification of conceptualization shared in a community
What is an Ontology?
A formal specification of conceptualization shared in a community
Vocabulary for defining a set of things that exist in a world view
Formalization allows communication across application systems and extension
Parallel concepts in other areas:
Domains: database theory
Types: AI
Classes: OO systems
Types/Sorts: Logic
Global vs. Domain-specific
Cornell CS 502

21. XML and RDF are ontologically neutral
No standard vocabulary just primitives
Resource, Class, Property, Statement, etc.
Compare to classic first order logic
Conjunction, disjunction, implication, existential, universal quantifier
Cornell CS 502

22. Components of an Ontology
Vocabulary (concepts)
Structure (attributes of concepts and hierarchy)
Logical characteristics of concepts & attributes
Domain and range restrictions
Properties of relations (symmetry, transitivity)
Cornell CS 502

23. On-line lexical reference system, domain-independent
Wordnet
On-line lexical reference system, domain-independent
>100,000 word meanings organized in a taxonomy with semantic relationships
Synonymy, meronymy, hyponymy, hypernymy
Useful for text retrieval, etc.
Cornell CS 502

24. Effort in AI community to accommodate all of human knowledge!!!
CYC
Effort in AI community to accommodate all of human knowledge!!!
Formalizes concepts with logical axioms specifying constraints on objects and classes
Associated reasoning tools
Contents are proprietary but there is OpenCyc
Cornell CS 502

25. Lots of Participants and $$$
Ontologies for the Web
Lots of Participants and $$$
Web Ontology Working Group
Distributed Agent Markup Language
Ontology Inference Layer
OntoWeb
Schemas Project
OWL (Web Ontology Language) – develop standard for encoding ontologies on top of RDF Schema
Cornell CS 502

26. Extending RDF(S) with OWL
Class definition: Conjunction, disjunction, negation
Property constraints: universality, existence, cardinality
Properties of properties: transitivity, symmetry
Class, sub-class definition
Property (attribute), sub-property definition
Domain, range constraints
Cornell CS 502

27. DAML class building operations
disjointWith
No vegetarians are carnivores
sameClassAs (equivalence)
Enumerations (on instances)
The Ivy League is Cornell, Harvard, Yale, ….
Boolean set semantics (on classes)
Union (logical disjunction)
Intersection (logical conjunction)
complimentOf (logical negation)
All non-carnivores are vegetarians
Cornell CS 502

28. DAML property building operations & restrictions
Transitive Property
P(x,y) and P(y,z) -> P(x,z)
SymmetricProperty
P(x,y) iff P(y,x)
Functional Property
P(x,y) and P(x,z) -> y=z
inverseOf
P1(x,y) iff P2(y,x)
InverseFunctional Property
P(y,x) and P(z,x) -> y=z
Cardinality
Cornell CS 502

29. Full use of XML schema data type definitions Examples
DAML+OIL DataTypes
Full use of XML schema data type definitions
Examples
Define a type age that must be a non-negative integer
Define a type clothing size that is an enumeration “small” “medium” “large”
Cornell CS 502

30. DAML+OIL Instance Creation
Create individual objects filling in slot/attribute/property definitions
<Person ref:ID=“William Arms”> <rdfs:label>Bill</rdfs:label> <age><xsd:integer rdf:value=“57”/></age> <shoesize><xsd:decimal rdf:value=“10.5”/></shoesize> </Person>
Cornell CS 502

31. Language Comparison
DTD
XSD
RDF(S)
OWL
Bounded lists (“X is known to have exactly 5 children”) X
Cardinality constraints (Kleene operators)
Class expressions (unionOf, complementOf)
Data types
Enumerations
Equivalence (properties, classes, instances)
Formal semantics (model-theoretic & axiomatic)
Inheritance
Inference (transitivity, inverse)
Qualified contraints (“all children are of type person”
Reification
Cornell CS 502

32. JENA toolkit for manipulating RDF models
Some useful RDF tools
JENA toolkit for manipulating RDF models
RDFSviz for visualizing ontologies expressed as RDF schema
W3C RDF validation service for parsing and view RDF instances
Cornell CS 502