Semantic Web and RDF

1
Semantic Web and RDF

• Numan Salati
• numan_at_cc.gatech.edu
• November 16,2006

2
Introduction

• Tim Berners-Lees vision of the web
• The Semantic Web is an extension of the
  current web in which information is given
  well-defined meaning, better enabling computers
  and people to work in cooperation
• The vision is to transform the current web into a
  vast distributed knowledge base.

3
Current web vs. Semantic web

• Current web
• HTTP how to transfer data
• URI how to locate and address data
• HTML how to display data
• Syntactic web Computers do the presentation of
  information and humans do the linking and
  interpretation of information.

4
Limitations of current web

• Searching for information
• Primarily keyword based Cant do complex queries
  which require background knowledge
• High recall, low precision
• Extracting information
• E.g. What is the price of this book?
• E.g. What does this image mean?
• Combining information
• E.g. Make an airline and hotel reservation based
  on my calendar information

5
What is the problem?

• Most of the information is unstructured and
  semantic content is accessible to humans not
  computers.
• Instead of publishing all the information in
  natural language, some more and more machine
  processable data.

6
Semantic web solution

• Machine processable -gt Structured knowledge
• Need to define data/metadata in more structured
  way so that vast amounts of knowledge can be
  shared, reused, searched, reasoned about and
  consequently made accessible to automated
  processing

7
Approaches

• Simple Semantic Annotation(tagging)
• del.icio.us, Flickr
• But we want more expressibility
• Formally specify terms in our domain (i.e.
  meaning is unambiguous)
• Relate terms to related to each other and to
  other terms in other domains
• Express constraints on the meaning of terms

8
How to represent knowledge?

• Ontologies
• A formal, explicit specification of a shared
  conceptualization
• formal ? Machine Processable
• explicit specification ? everything (concepts,
  relations, axioms etc are explicitly defined)
• shared ? consensual Knowledge
• conceptualization ? abstract model of some
  phenomenon in the world

9
Ontology

• Ontology
• Concepts (vocabulary) Facts, constrains,
  relations (background knowledge)

10
Anatomy of an ontology language

• Classes/Terms/Concepts
• Grouping (a set) of individuals with common
  properties
• E.g. Person, Car, School
• Relations
• Connect/relate concepts to each other
• E.g. Person owns Car
• Instances/Individual/Objects
• Numan owns a Corvette.
• Constraints/Axioms
• Additional statements about the domain
• E.g. Every persons owns at least one car

11
Ontology languages

• A continuum of ontology languages
• Glossary
• Thesauri
• Taxonomies
• Relational Schema
• EER
• RDFS
• Logic based
• OWL, DAMLOIL (DL based)
• Prolog, RuleML (Rule based)
• KIF (FOL based)

Increasing order of expressiveness
12
Ontology languages

• These languages vary in the degree of formalism
  and expressiveness
• Reasoning becomes easier on very formal and
  expressive languages.
• Tradeoff between expressibility and computability.

13
Reasoning

• Types of reasoning
• Subsumption
• check whether a concept is more general than
• Equivalence
• check whether two concepts are equivalent
• Consistency
• check whether a concept is meaningful
• Membership
• check whether an individual i is a member of
• Closed-world vs. Open-world assumption
• Close-world Everything you dont know is false
• Open-world Everything you dont know is
  undecidable

14
Adding semantics

• Ontologies provide a vocabulary of terms. Use
  ontologies to specify meaning of annotations. New
  terms can be formed by combining existing ones.
• Ontologies provide us background knowledge
  (information models)
• This knowledge can be reasoned about and shared
  with humans and/or software agents (think
  automation).

15
Semantic web Layer cake
16
Main Components

• To make data machine processable we need
• unambiguous names for resources URI
• a common data model to access, connect, describe
  the resources RDF
• access to that data SPARQL
• define common vocabularies RDFS, OWL,
• reasoning logics OWL, Rules (SWRL)

17
Resource Description Framework (RDF)

• RDF is a specification standard by W3C to
  represent meta data on the web.
• RDFS extends RDF with schema vocabulary
• Class, SubClassOf
• Property, subPropertyOf
• range, domain restrictions

18
RDF concepts

• Resources (identified by URIs)
• A URI identifies a resources, but does not
  necessarily point to it
• Correspond to nodes in a graph
• E.g. http//www.w3.org/, http//myhomepage.org/jo
  hn,
• Properties (identified by URIs)
• Correspond to labels of edges in a graph
• Binary relation between two resources
• E.g. http//www.example.org/hasName,
• Literals
• concrete data values
• E.g. 25", New York"

19
RDF Concepts (contd.)

• Triple data model
• ltsubject,predicate,objectgt
• Subject Resource or blank node
• PredicateProperty
• Object Resource, literal or blank node
• Example
• ltjohn,father-of,billgt
• Labeled, directed graphs
• Nodes resources, literals
• Labels properties
• Edges statements

20
Resources

• A resource may be
• Web page (e.g. http//www.w3.org/)
• A person (e.g. www.johnswebsite.net)
• A book (e.g. urnisbn0-345-33971-1)
• Anything denoted with a URI
• A URI is an identifier and not a location on the
  Web. Used to distributed concepts and avoid
  collisions
• RDF allows making statements about resources
  (meta-data)
• http//www.w3.org/ has the format text/html
• urnisbn0-345-33971-1 has the author Tolkien

21
Literals

• Plain literals
• E.g. foo
• Typed literals
• E.g. "hello"xsdstring, "1"xsdinteger
• Recommended datatypes XML Schema datatypes
• Datatype mechanism extensible in XSD
• Only as object of a triple, e.g.
• ( lthttp//blah.org/johngt,
• lthttp//blah.org/hasNamegt,
• JohnSmithxsdstring )

22
Blank node and Reification

• Reification statements about statements
• Mary claims that Johns name is John Smith.
• Blank Nodes
• Unnamed resources
• More complex constructs

23
RDF Vocabulary

• Classes
• rdfProperty rdfStatement rdfXMLLiteral rdfSeq
  rdfBag, rdfAlt rdfList
• Properties
• rdftype rdfsubject rdfpredicate rdfobject
  rdffirst rdfrest, rdf_n, rdfvalue
• Resources
• rdfnil
• Typing using rdftype
• ltA,rdftype,Bgt A belongs to class B
• All properties belong to class rdfProperty
• ltP,rdftype,rdfPropertygt P is a property
• Example ltrdftype,rdftype,rdfPropertygt
  rdftype is a property

24
RDF containers

• Used for grouping values
• Three types
• Bag The lecture is attended by John, Mary and
  Chris
• SeqRDF-Concepts is edited by Graham and
  Jeremy (in that order)
• Limitation How would you we know whether Graham
  and Jeremy are the only editors? (i.e a closed
  set)
• AltThe source code for the application may be
  found at ftp1.example.org, ftp2.example.org,
  ftp3.example.org

25
RDF Schema (RDFS)

• RDF gives a formalism for meta data annotation,
  and a way to write it down in XML, but it does
  not give any special meaning to vocabulary such
  as subClassOf or type
• Interpretation is an arbitrary binary relation
• I.e., ltPerson,subClassOf,Animalgt has no special
  meaning
• RDF Schema defines schema vocabulary that
  supports definition of ontologies
• gives extra meaning to particular RDF
  predicates and resources (such as subClasOf)
• this extra meaning, or semantics, specifies how
  a term should be interpreted

26
RDFS Examples

• These terms are the RDF Schema building blocks
  (constructors) used to create vocabularies
• ltPerson,type,Classgt
• lthasColleague,type,Propertygt
• ltProfessor,subClassOf,Persongt
• ltCarole,type,Professorgt
• lthasColleague,range,Persongt
• lthasColleague,domain,Persongt

27
RDFS vocabulary

• RDFS Classes
• rdfsResource
• rdfsClass
• rdfsLiteral
• rdfsDatatype
• rdfsContainer
• rdfsContainerMembershipProperty
• RDFS Properties
• rdfsdomain
• rdfsrange
• rdfssubPropertyOf
• rdfssubClassOf
• rdfsmember
• rdfsseeAlso
• rdfsisDefinedBy
• rdfscomment
• rdfslabel

28
RDFS example
29
RDF/XML

• RDF can be represented as XML (i.e. xml
  serialization)
• Ugly and tedious to write by hand
• Use tools to generate it.
• Other sytanx turtule, N3

30
Querying RDF

• SPARQL
• RDF Query language
• Uses SQL-like syntax
• Example
• PREFIX dc http//purl.org/dc/elements/1.1/
• SELECT ?title
• WHERE lthttp//example.org/book/book1gt dctitle
  ?title

31
More SPARQL

• PREFIX
• Prefix mechanism for abbreviating URIs
• SELECT
• Identifies the variables to be returned in the
  query answer
• FROM
• Name of the graph to be queried
• Can have multiple graphs
• WHERE
• Query pattern as a list of triple patterns

32
Queries

• Return the full names of all people in the
  graph
• PREFIX vCard lthttp//www.w3.org/2001/vcard-rdf/3
  .0gt
• SELECT ?fullName
• WHERE ?x vCardFN ?fullName

33
Queries

• Return the relation between John and Mary
• PREFIX lthttp//example.org/datagt
• SELECT ?p
• WHERE john ?p mary

34
Tools

• Databases (Triple Stores)
• Oracle Triple store
• Kowari
• Sesame
• Graphic editors
• Protégé
• GraphViz
• Parsers
• Jena semantic web toolkit
• Protégé API

35
Who is using RDF?

• Mozilla/FireFox bookmarks
• Biological Databases
• RSS
• Web Admins
• Publishing Industry

36
References

• Tim Berners-Lee, James Hendler and Ora Lassila
  The semantic web http//www.scientificamerican.c
  om/article.cfm?articleID00048144-10D2-1C70-84A980
  9EC588EF21catID2
• OWL Web Ontology Language Guide.
  http//www.w3.org/TR/owl-guide/
• RDF Resource Description Framework
  http//www.w3.org/RDF/
• SPARQL Query language for RDF http//www.w3.org/T
  R/rdf-sparql-query/
• Protégé Ontology editor http//protege.stanford.e
  du/