The Semantic Web Overview

Luis Snchez Fernndez

Departamento de Ingeniera Telemtica Universidad Carlos III de Madrid

Short history of the Web

1990: Creation of World Wide Web infraestructure at CERN by Tim Berners Berners-Lee
HTTP, HTML, first Web client, client, first Web server

1993: Mosaic Mosaic, , first graphic Web client 1994: Netscape Navigator 1996: Commercial use of WWW is generalized 1999: Tim Berners Berners-Lee proposes the Semantic Web 2002: Weblogs and RSS Web 2.0 6th October 2009: at least 8 billion indexable Web pages 23rd September 2010: at least 15 billion indexable Web pages
according to http://www.worldwidewebsize.com/

The problem of information overload

The great success of the web has lead to one of its current problems: information overload
Difficult and time costly to find and update relevant information for people and companies
Ex.: keep an updated state of the art

Company employees can use up to 20% of their working time searching in the Web (Outsell Inc, 2002)

Web problems and pitfalls example: search engines

We make queries and get Web pages that are not related to what we wanted We make queries and do not get Web pages that are related to what we wanted

Another example
Search for images in flickr.com
Java (island (island/ /bird/ bird/coffee/ coffee/programming language) language) vela (Spanish (Spanish) ) (in English: English: candle/ candle/sail) sail)

Web problems and pitfalls example example: : search engines

Some reasons of search engines problems: problems:
polisemy/ polisemy/homonym synonymy multilinguism

In summary summary: : search engines are not based on meanings but in terms (syntactic search) search)

Semantic Search Engine v1.0

We need a notation to identify meanings
example: example: (term (term, , number) number) tuple
(George Bush , 1) Bush senior (George Bush , 2) Bush junior

To attach to (parts of) a document the meaning of terms that are mentioned on it it: : semantic annotation

Semantic Annotation
How somebody knows which is the code of some concept?
By means of vocabularies shared by a community of users
example: example : Wikipedia

How somebody knows that the concept represented by some code is referenced in a given Web page?
semantic annotation the Web is an open world need for trustness mechanisms

Types of vocabularies
Free text Controlled vocabulary Thesauri narrower term relation Informal is-a Formal is-a General Frames Logical (properties constraints ) Value Restrs. Disjointness, Inverse, Part-Of ...

Terms/ glossary

Formal instance

Lassila O, McGuiness D. The Role of Frame-Based Representation on the Semantic Web. Technical Report. Knowledge Systems Laboratory. Stanford University. KSL-01-02. 2001.

Controlled Vocabulary
Example: Example : a catalog
http://www.todocoleccion.net/catalogo.cfm

Glossary
http://www.essentialsofmusic.com/glossary/glossary.html

Thesaurus
Example: Example : UNESCO Thesaurus
http://www2.ulcc.ac.uk/unesco/thesaurus.htm

Informal IsIs-a
IsIs-a: specifies that a concept is narrower than other
A Professors is is-a Human Luis is is-a Professor

Informal IsIs-a: although the relation is usually fulfilled there is not a 100% guarrantee
A Mammal is is-a Not EggEgg-Lying Animal
Platypus

WordNet
Developed at Princeton University Contains nouns, nouns, verbs, verbs, adjetives, adverbs Organized in synsets (synonyms lists + gloss) gloss) Meanings identified by tuples (term, term, number) number) Meanings identified by number Semantic relationships among synsets WordNet Multilingual version: version: EuroWordNet

Formal IsIs-a
A hierarchy of concepts

Logical constraints
Additional properties Axioms that model the relations between concepts and properties in the vocabulary

Ontology definition
Gruber, Borst, 1993:
An ontology is a formal, explicit specification of a shared conceptualization

URIs
Mechanism used to represent namespaces Used to identify resources
anything you can talk about in a Web document

URI components:
URI scheme: http, ftp, urn, ... name

Types of URIs: URL, URN

URIref
URI plus (optionally) fragment identifier Example
http://www.example.org/index.html#section2 URI: http://www.example.org/index.html Fragment identifier: section2

Web problems and pitfalls example: search engines

Questions and answers: answers:
Which are the names of Spanish Comunidades Autnomas? Who is the Spanish political party PSOE president? president ?

10

Semantic Search Engine v2.0

Semantic annotations extension Statements formal representation: representation:
Grin is PSOE president president
isPresident( isPresident (Grian Grian, , PSOE)

Andaluca is an Spanish Comunidad Autnoma

isCCAA(Andaluca) isCCAA (Andaluca)

Semantic Search Engine v3.0

Imagine the following query:
We want to find PSOE members that have positions in some Comunidad Autnoma government

We search in Web (1.0) and find this information:

NOTE: real example by 2005 year

11

Semantic Search Engine v3.0

Manuel Chaves is PSOE president PSOE is a political party Manuel Chaves is Andaluca government president Manuel Chaves is Juan Carlos Rodrguez Ibarra party comrade Andaluca is a Comunidad Autnoma Juan Carlos Rodrguez Ibarra is Extremadura government president Extremadura is a Comunidad Autnoma

Semantic Search Engine v3.0

Domain knowledge:
X, Y, Z, isPartyMember(X,Z) isPartyMember(X,Z) isPartyMember(Y,Z) isPartyComrade(X,Y) X, Y, isPartyComrade(Y,X) isPartyComrade(Y,X) isPartyComrade(X,Y) X, Z, isPartyMember(X,Z) isPartyMember(X,Z) isPartyPresident(X,Z) X, Y, isGovernmentMemberCCAA(X,Y) isGovernmentMemberCCAA(X,Y) isConsejeroCCAA(X,Y) X, Y, isGovernmentMemberCCAA(X,Y) isGovernmentMemberCCAA(X,Y) isPresidentCCAA(X,Y) X, Y, isPresidentCCAA(X,Y) isPresidentCCAA(X,Y) isPresident(X,Y) isCCAA(Y) X, Y, isPartyPresident(X,Y) isPartyPresident(X,Y) isPresident(X,Y) isParty(Y)

12

Algorithm
example example: : PSOE

Note: this is just a sample algorithm

Search in the (Semantic Semantic) ) Web for statements related to entities mentioned in the query Apply logic reasoning to get new statements Repeat Web search over entities (subject, subject, object) object) present in found statements Stop after a number of cycles

Step 1
Search for statements where PSOE appears You get: get:
isPresident( (Chaves,PSOE Chaves,PSOE) ) isPresident isParty isParty(PSOE) (PSOE)

Applying rules

X, Y, isPartyPresident sPartyPresident( (X,Y X,Y) ) isPresident isPresident( (X,Y X,Y) ) isParty(Y) isParty (Y) X, Z, esPartyMember( esPartyMember(X,Z X,Z) ) isPartyPresident isPartyPresident( (X,Z X,Z) ) isPartyPresident isPartyPresident( (Chaves,PSOE Chaves,PSOE) ) isPartyMember( isPartyMember(Chaves,PSOE Chaves,PSOE) )

You get: get:

13

Step 2
Search for statements where Chaves appears You get:
isPresident(Chaves,Andaluca) isPartyComrade(Chaves,Ibarra)

Applying rules
X, Y, Z, isPartyMember(X,Z) isPartyMember(X,Z) isPartyMember(Y,Z) isPartyComrade(X,Y) X, Y, isPartyComrade(Y,X) isPartyComrade(Y,X) isPartyComrade(X,Y)

You get:
isPartyMember(Ibarra,PSOE)

Step 3
Search for statements where Andaluca or Ibarra appear You get: get:
isCCAA isCCAA(Andaluca) (Andaluca) isPresident isPresident(Ibarra, (Ibarra, Extremadura)

Applying rules
X, Y, isPresidentCCAA sPresidentCCAA( (X,Y X,Y) ) isPresidente isPresidente( (X,Y X,Y) ) isCCAA sCCAA(Y) (Y)

You get: get:

isPresidentCCAA isPresidentCCAA( (Chaves,Andaluca Chaves,Andaluca) )

14

Step 4
Search for statements where Extremadura appears You get: get:
isCCAA isCCAA(Extremadura) (Extremadura)

Applying rules
X, Y, isPresidentCCAA( isPresidentCCAA(X,Y) X,Y) isPresident( isPresident (X,Y) X,Y) isCCAA(Y) sCCAA(Y)

You get: get:

isPresidentCCAA( isPresidentCCAA(Ibarra,Extremadura) Ibarra,Extremadura)

Query result
Chaves Ibarra

15

Semantic Search Engine v3.0: Components

Knowledge base Reasoner These are the well known basic components of a knowledgeknowledge-based system

Knowledge Base
Domain knowledge model Information recovered from the Web The domain knowledge model can be built reusing available components
Ex.: a model for locations (concepts like city, country, etc.) can be used in different applications (book a travel, a semantic search application for news items, etc.) Reusable knowledge components => ontologies

16

Ontology components
Classes (ex. Party, Comunidad Autnoma) Instances (ex. PSOE, Andaluca) Properties (ex. isPartyMember) Rules
knowledge creation restrictions reactive rules

Lightweight vs. heavyweight ontologies

What is the Semantic Web

Formal description (machine readable) of Web published contents

17

Technologies
Semantic annotation Ontology engineering Reasoners

Proposal (T. BernersBerners-Lee)

18

Applications
Semantic search
Show user just relevant results Question/answer functionality Multimedia content search

Information integration
Several, distributed sources (ex. databases) integrated by a common domain ontology + mappings

Services semantic description will enable its discovery Web task automatization
Ex. comparing prices between several commercial sites
Filling of forms based on semantics instead of syntax

Material for next session

Protg Ontology editor
http://protege.stanford.edu/

Pellet DL Reasoner
http://clarkparsia.com/pellet

OWL tutorial
http://www.cohttp://www.co-ode.org/resources/tutorials/ProtegeOWLTutorial.pdf

19