introduction to semantic web, metadata, knowledge representation, ontologies

Introduction to

the Semantic Web

F. Corno, L. Farinetti - Politecnico di Torino 2

Semantic Web

Web second generation

Web 3.0

“Conceptual structuring of the Web in an explicit

machine-readable way”

(Tim Berners-Lee)

In other words…

…let the machine do most of the work!!!

http://www.w3.org/2001/sw/

http://www.w3.org/People/Berners-Lee/


“Official” introduction The Semantic Web is a web of data. There is

lots of data we all use every day, and its not part

of the web. I can see my bank statements on the

web, and my photographs, and I can see my

appointments in a calendar. But can I see my

photos in a calendar to see what I was doing

when I took them? Can I see bank statement

lines in a calendar?

Why not? Because we don’t have a web of data.

Because data is controlled by applications, and

each application keeps it to itself


Example


“Official” introduction

The Semantic Web is about two things

It is about common formats for integration and combination of data drawn from diverse sources, where on the original Web mainly concentrated on the interchange of documents.

It is also about language for recording how the data relates to real world objects. That allows a person, or a machine, to start off in one database, and then move through an unending set of databases which are connected not by wires but by being about the same thing


An example …

How can a machine distinguish the

meanings … ?

“I am a professor of computer science.”

“I am a professor of computer science,

you may think. Well,…”


Key principles

The Semantic Web is the Web

Same base technologies, evolutionary

Decentralized (incomplete, inconsistent)

Provide explicit statements regarding web resources

Authors, original information providers

Intermediaries (humans and/or machines)

Information consumers determine consequences of the statements

Distributed ‘reasoning’


1989:

WWW

original

proposal


Technology stack (old: pre-2008)


Technology stack (current: 2008)


The real world

Not

yet...!


The real world

Not

yet...! Not always

necessary...

Current “hot” topics


Metadata and

Metadata Standards


Goal of the semantic Web

The Semantic Web will enable

machines to COMPREHEND semantic

documents and data, NOT human

speech and writing

Then, how???

Semantic Web foundation: metadata


Resource and description

Resource

Content, format, …

Access method dependent on format (I can read it if I “know” its language)

Resource description

Independent of the format (I can read “people’s comments” about the resource… provided that I know the language in which the comment is written)



description

resource

this resource

was created on

April 14th, 2009

the title of this

resource is

“Introduction to

the Semantic

Web”

the author of

this resource

is L. Farinetti

this resource is

related to

computer

science,

knowledge

representation

and metadata

the quality of

this resource

is high,

according to F.

Corno

this resource is suitable

for PhD students



Resource Content, format, …

Access method dependent on format (I can read it if I “know” its language)

Standardization (i.e. common language for applications) ??? Practically impossible …

Huge amount of existing information

Hundreds of human languages

Hundreds of computer languages (other word for formats)



Resource description

Independent of the format (I can read “people’s

comments” about the resource… provided that I know

the language in which the comment is written)

Standardization (i.e. common language for

applications) ???

Feasible

Smaller amount of information, possibly new

Solution: define a standard language for writing

comments (“metadata” in semantic web terminology)



this resource

was created on

April 14th, 2009

the title of this

resource is

“Introduction to

the Semantic

Web”

the author of

this resource

is L. Farinetti

this resource is

related to

computer

science,

knowledge

representation

and metadata

the quality of

this resource

is high,

according to F.

Corno

this resource is suitable

for PhD students

Metadata

Field name = field value



description

resource

Date =

2009-04-14

Title =

“Introduction to

the Semantic

Web”

Author =

L. Farinetti

Topic =

{computer

science,

knowledge

representation,

metadata}

Quality = high

Level = PhD students

Rated by F. Corno


Semantic Web main tasks

Metadata annotation

Description of resources using standard

languages

Search

Retrieve relevant information according to

user’s query / interest / intention

Use metadata (and possibly content) in a

“smart” way (i.e. “reasoning” about the

meaning of annotations)


Meaningful metadata annotations

Common language for describing resources

Resource description standards

Common language for description field names

Metadata standards

Common language for description field values

Metadata standards + controlled vocabularies

Semantically rich descriptions to support search

Knowledge representation techniques, ontologies


Common language for describing

resources



resources

Resource Description Framework (RDF)

Resource = URI (retrievable, or not)

RDF is structured in statements

A statement is a triple

Subject – predicate – object

Subject: a resource

Predicate: a verb / property / relationship

Object: a resource, or a literal string



resources

Diagram:

Simple RDF assertion (triple):

triple (hasAuthor, URI, L.Farinetti)

URI L.Farinetti hasAuthor

Author =

L. Farinetti



resources

RDF in XML syntax:

Author =

L. Farinetti

<RDF xmlns=“http://www.w3.org/TR/ … ” >

<Description about=“http://www.polito.it/semweb/intro”>

<Author>L.Farinetti</Author>

</Description>

</RDF>


Common language for field names

Title = ...

Problem

Author = …

Creator, Maker,

Contributor …

Synonymy

Topic = …

Topics, Subject, Subjects,

Argument, Arguments

Singular / plural

Level = …

Difficult to clearly

define concept in a

few words

Educational level,

destination, suitability, …

Date = …

Date of creation, date of

last modification, date of

revision, …

Different concepts:

need for more details


Common language for field names

Solution: metadata standards

Many standardization bodies are involved

Standards may be general

e.g. Dublin Core (DC)

or may depend on goal, context, domain, …

e. g. educational resources (IEEE LOM), multimedia

resources (MPEG-7), images (VRA), people (FOAF,

IEEE PAPI), geospatial resources (GSDGM),

bibliographical resources (MARC, OAI), cultural

heritage resources (CIDOC CRM)


Metadata standards examples


Dublin Core

Dublin Core Metadata Element Set

(DCMES)

Building blocks to define metadata for the

Semantic Web

15 elements, or categories, general enough to

describe most of the published resources

Extra elements and element refinements


DC metadata element set


Example of description using

Dublin Core (in RDF)

A paper in the

“Ariadne” journal


Common language for field values

Problems

Value type

Title =

“Introduction to

the Semantic

Web”

type = string

Date =

2009-04-14

type = date

Author =

L. Farinetti

type = string

“standard” format?

Laura Farinetti, Farinetti

Laura, Farinetti L., …



Problems

Value type

Value restrictions?

freedom vs shared understanding

Quality = high

High, medium, low?

1 to 5?

any value?

Level = PhD students

any value?

list of possible values?

Topic =

{computer

science,

knowledge

representation,

metadata}

any value?

any number of values?



Solution: metadata standards + controlled

vocabularies

Metadata standards

Only some, and partially

Controlled vocabularies

Explicit list of possible values


Examples from IEEE LOM

1484.12.1 - 2002 Learning Object

Metadata (LOM) Standard

Developed by the IEEE Learning Technology

Standards Committee (LTSC)

Standard to describe the “Learning

Objects” in order to guarantee their

interoperability


… + controlled vocabularies

A closed list of named subjects, which can

be used for classification

Metadata field values are

restricted to a list of terms

(selected by experts)

Topic =

{computer

science,

informatics,

knowledge

representation,

metadata}


Semantically rich descriptions to

support search



support search

http://dictybase.org/db/html/help/GO.html

Topic =

{metabolism, …}

Knowledge

Representation


Need for knowledge representation

Semantically rich descriptions need

“understanding” the meaning of a resource

and the domain related to the resource

Disambiguation of terms

Shared agreement on meanings

Description of the domain, with concepts and

relations among concepts


Example: Dublin Core metadata

Metadata of a single paper


Problems

Title usually offers good clues, but it does not necessarily mention all names of all

subjects the user is interested in

it may presuppose knowledge the user does not actually possess

Subject is meant to convey precisely what the document is about, but much depends on how extensive the set of keywords

is, whether all related subjects are mentioned, and whether too many subjects are listed

Metadata does not say much about “how related” a resource is to a given subject


Search results for “topic maps”


Problems

Authors were free to define their own

subject keywords

Results are not “about” topic maps, but

“related to” topic maps

If an author forgets to list “topic maps”, his

paper will never be found


Subject-based classification

Any form of content classification that groups objects by their subjects e.g the use of keywords to classify papers

Metadata fields describe what the objects are about by listing discrete subjects inside a subject-based classification

Important: difference between describing the objects being classified and describing the subjects used to classify them Metadata describe objects

Subject-based classification is the approach to describe subject


Subject-based classification ... “On those remote pages it is written that animals are divided into:

a. those that belong to the Emperor b. embalmed ones c. those that are trained d. suckling pigs e. mermaids f. fabulous ones g. stray dogs h. those that are included in this classification i. those that tremble as if they were mad j. innumerable ones k. those drawn with a very fine camel's hair brush l. others m. those that have just broken a flower vase n. those that resemble flies from a distance"

From The Celestial Emporium of Benevolent Knowledge, Borges



techniques


Taxonomies

Thesauri

Faceted classification

Ontologies

Folksonomies

Others

… Most come from library science


Controlled vocabulary

A closed list of named subjects, which can be

used for classification

Composed of terms: particular name for a

particular concept

similar to keywords

Terms are not concepts

A single term may be the name of one or more

concepts

A single concept may have multiple names

Ambiguity avoided by forbidding duplicate terms



Goal

Prevent authors from defining terms that are meaningless, too broad or too narrow

Prevent authors from misspelling

Prevent different authors from choosing slightly different forms of the same term

The simplest form of controlled vocabulary is a list of terms (or “pick list”)

Topic =

{computer

science,

knowledge

representation,

mtadata, RDF,

topic navigation

maps} topic maps



Reduce ambiguity inherent in normal human languages

Solve the problems of homographs, homonyms, synonyms and polysemes by ensuring

That each concept is described using only one authorized term

That each authorized term in the controlled vocabulary describes only one concept


Problems solved

Synonym

different words with identical or very similar meanings


Problems solved

Synonym


close “Will you please close that door!”

“The tiger was now so close that I could smell it...”

pupil student

opening in the iris of the eye

axes

('æk.səz) plural of axe

('æk.siz) plural of axis


Problems solved

Synonym


student and pupil (noun)

buy and purchase (verb)

sick and ill (adjective)

to get

take (I'll get the drinks)

become (she got scared)

wood

understand (I get it)

a piece of a tree

a geographical area with many trees


Controlled vocabulary examples

Practically no “real” examples

With very little extra effort: taxonomies and

thesauri!

Circuit theory

Electronic circuits

Microwave technology

Electron tubes

Semiconductor materials and devices

Dielectric materials and devices

Magnetic materials and devices

Superconducting materials and devices

…

Blood

Cord blood

Erythrocyte

Leukocyte

Basophil

Eosynophil

Lymphoblast

Lymphocyte

Monocyte

Neutrophil

…


Taxonomy

Subject-based classification that arranges the terms in the controlled vocabulary into a hierarchy

Dates back to Carl Linnæus’s work on zoological and botanical classification (18th century)


Taxonomy

Allow related terms to be grouped together

It is clear that “topic

maps” and “XTM” are

related

Easier to classify

documents

Easier to choose

search keywords


Taxonomies and metadata

Metadata are stored as usual with the resource

The “subject” will contain only controlled terms

Controlled terms belong to a hierarchy, shared by all papers


Taxonomy example: INSPEC

http://www.theiet.org/publishing/inspec/index.cfm


Taxonomy example: INSPEC


INSPEC

journal

article

database


Taxonomy example: anatomy terms

http://www.cbil.upenn.edu/anatomy.php3


Taxonomy example


Taxonomy example

http://www.acm.org/class/1998/ccs98.html


Taxonomy limits

Only two kinds of relationships between terms

Parent = broader term

Child = narrower term

topic navigation maps synonym

no more in use

difference?

synonym XML topic map

difference?


Thesaurus

Extends taxonomies

subjects are arranged in a hierarchy

Other statements can be made about the

subjects

Two ISO standards

ISO2788 for monolingual thesauri

ISO5964 for multilingual thesauri


Thesaurus relationships

BT – broader term Refers to a term with wider or less specific meaning

Some systems allow multiple BTs for one term, while others do not

Inverse property: NT - narrower term

A taxonomy only uses BT and NT

SN – scope note String explaining its meaning within the thesaurus

Useful when the precise meaning of the term is not obvious from context


Thesaurus relationships USE

Another term that is to be preferred instead of this term

Implies that the terms are synonymous

Inverse property: UF

TT – top term The topmost ancestor of this term

The BT of the BT of the BT...

RT – related term A term that is related to this term, without being a

synonym of it or a broader/narrower term


Thesaurus example

http://www.ukat.org.uk/thesaurus/


Thesaurus example

http://www.swinburne.edu.au/corporate/registrar/rms/keywords.htm


Thesaurus example

Library of Congress

Subject Heading

http://www.loc.gov/cds/lcsh.html

W3C

standard:

SKOS


UK Archival Thesaurus


Faceted classification

Proposed by

S.R. Ranganathan in the ‘30s

Facets are the different axes along which

documents can be classified

Each facet contains a number of terms

Usually with a thesaurus organization

Usually a term belongs to one facet only

A document is classified by selecting one term

from each facet


Faceted classification example

http://flamenco.berkeley.edu/


Advantages

Multi-

dimensionality

Persistence

Scalability

Flexibility

http://freeable.polito.it/


Ontology

Model for describing the world that

consists of a set of types, properties, and

relationships

Extends the other subject-based

classification approaches

Has open vocabularies

Has open relationship types (not just BT/NT,

RT and USE/UF)


Ontology structure

Concepts

Relationships

Is-a

Other

Instances


Folksonomy

Internet-mediated social environments

Tags compiled through social tagging

Social tagging

Decentralized practice where individuals and groups create, manage and share tags to annotate digital resources in an online social environment

Generally characterized by non-standard tagging


Other subject-based techniques

Synonym rings

Connect together a set of terms as being

equivalent for search purpose

Similar to UF/USE relationship of thesauri,

but no preferred term


Other subject-based techniques

Authority file

Similar to a synonym ring, but consists of UF/USE

relationships instead of synonym relationships

One term in each synonym ring is indicated as the

preferred term for that subject

e.g. Library of

Congress Name

Authority File



summary

Terminology is rarely used

in a consistent way


are thesauri, thesauri are

ontologies, …

http://www.iesr.ac.uk/profile/vocabs/index.html/#CtrldVocabsList



summary

Ontologies



support search

http://dictybase.org/db/html/help/GO.html

Topic =

{metabolism, …}


Ontologies

An ontology is an explicit description of

a domain

concepts

properties and attributes of concepts

constraints on properties and attributes

individuals (often, but not always)

An ontology defines

a common vocabulary

a shared understanding


“Ontology engineering”

Defining terms in the domain and relations

among them

defining concepts in the domain (classes)

arranging the concepts in a hierarchy

(subclass-superclass hierarchy)

defining which attributes and properties (slots)

classes can have and constraints on their

values

defining individuals and filling in slot values


Why develop an ontology?

To share common understanding of the

structure of information

among people

among software agents

To enable reuse of domain knowledge

to avoid “re-inventing the wheel”

to introduce standards to allow interoperability


An ontology

HNC

HND

Certificate

Diploma

Award

2 years

1 year

Is_a

Is_a

Is_a

Is_a

takes

takes

takes

takes

Is_equivalent_to


A more complex ontology [base.Entity] Person Worker Faculty Professor AssistantProfessor AssociateProfessor FullProfessor VisitingProfessor Lecturer PostDoc Assistant ResearchAssistant TeachingAssistant AdministrativeStaff Director Chair {Professor} Dean {Professor} ClericalStaff SystemsStaff Student UndergraduateStudent GraduateStudent


A more complex ontology Organization Department School University Program ResearchGroup Institute Publication Article TechnicalReport JournalArticle ConferencePaper UnofficialPublication Book Software Manual Specification Work Course Research Schedule


A more complex ontology

Relation Argument 1 Argument 2 ====================================================== publicationAuthor Publication Person publicationDate Publication .DATE publicationResearch Publication Research softwareVersion Software .STRING softwareDocumentation Software Publication teacherOf Faculty Course teachingAssistantOf TeachingAssistant Course takesCourse Student Course age Person .NUMBER emailAddress Person .STRING head Organization Person undergraduateDegreeFrom Person University mastersDegreeFrom Person University doctoralDegreeFrom Person University advisor Student Professor subOrganization Organization Organization ………..


Example of ontology engineering

chair



1.A piece of furniture consisting of a seat, legs, back, and often

arms, designed to accommodate one person.

2.A seat of office, authority, or dignity, such as that of a bishop.

a.An office or position of authority, such as a professorship.

b.A person who holds an office or a position of authority,

such as one who presides over a meeting or administers a

department of instruction at a college; a chairperson.

3.The position of a player in an orchestra.

4.Slang. The electric chair.

5.A seat carried about on poles; a sedan chair.

6.Any of several devices that serve to support or secure, such as

a metal block that supports and holds railroad track in position.

chair



A piece of furniture consisting of a seat, legs, back,

and often arms, designed to accommodate one

person.

chair



chair seat stool bench



Something I can sit on



???




“sittable”




table



“sittable”





“for_sitting”

table

“sittable”

Something designed for sitting


Ontology structure


“for_sitting”

table

“sittable”


Concepts

Some piece of furniture that can

be used to sit on, either by

design or by its shape.

Furniture to sit on

Shorthand name

Synthetic title

Definition “sittable”


Internationalization




Furniture to sit on

Shorthand name

Synthetic title

Definition

Furniture to sit on Furniture to sit on Furniture to sit on Furniture to sit on Furniture to sit on Furniture to sit on



















“sittable”


Relationships


“for_sitting”

table

“sittable”

is_a is_a is_a

is_a

is_a

is_a

room material

wood

is_a

classroom

dining room

is_a is_a


Relationships


“for_sitting”

table

“sittable”

is_a is_a is_a

is_a

is_a

is_a

room material

wood

is_a

classroom

dining room

is_a is_a

made_of

made_of


Ontology building blocks

Ontologies generally describe:

Individuals

the basic or “ground level” objects

Classes

sets, collections, or types of objects

Attributes

properties, features, characteristics, or parameters

that objects can have and share

Relationships

ways that objects can be related to one another


Individuals

Also known as “instances”

can be concrete objects

animals

molecules

trees

or abstract objects

numbers

words


Concepts

Also known as “Classes”

abstract groups, sets, or collections of objects

They may contain

individuals

other classes

a combination of both

Examples

Person: the class of all people

Vehicle: the class of all vehicles


Concepts Can be defined extensionally …

By defining every object that falls under the definition

of the concept

A class C is extensionally defined if and only if for

every class C', if C' has exactly the same members of

C, C and C' are identical

E.g.: DayOfWeek = {Monday, Tuesday, Wednesday,

Thursday, Friday, Saturday, Sunday}

… or intensionally

By defining the necessary and sufficient conditions for

belonging to the concept

E.g.: “bachelor” is an “unmarried man”


Concepts

Defined by

Name: any identifier, usually carefully chosen

Definition: describes the well agreed meaning

of the concept, in a human readable form

Terms (Lexicon): list of terms (synonyms, etc.)

usually adopted to identify the concept


Subsumption

A concept (class) can subsume / be

subsumed by any other class

Subsumption is used to establish class

hierarchies


Class partition

A set of related classes and associated

rules that allow objects to be placed into

the appropriate class

GEOMETRIC

FIGURE

GEOMETRIC POINT

TWO DIMENSIONAL

FIGURE ONE

DIMENSIONAL FIGURE


Class partition

Disjoint partition

A disjoint partition rule guarantees that a

single instance of a class cannot be in more

than one sub-classes

E.g. one specific truck

cannot be in both

4-axle and

6-axle classes

VEHICLE

CAR TRUCK

6-AXLE 4-AXLE


Class partition

Exhaustive partition

every concrete object in the super-class is an

instance of at least one of the partition

classes


Attributes

Describe specific features

Can be complex (e.g.: list of values)

Defined for a class/concept (e.g. car)

Examples:

number-of-doors: 4

number-of-wheels: 4

engine: {3.0L,4.0L}


Relationships

Attributes that relate two or more concepts

two concepts → binary relationship

three concepts → ternary relationship

Domain

the concept(s) from which the relationship departs

Range

the concept(s) to which the relationship applies


Relationships

Examples

Car(MiniMinor) → individual definition

Car(Mini) → individual definition

Successor(Mini,MiniMinor) → relationship

domain range


Commonly used relationships

Subsumption

the most important

is-superclass-of

usually denoted by its inverse is-a

(is-subclass-of)

Meronymy

is-part-of

describes how object are combined together

to form composite objects


Example

http://www.yeastgenome.org/help/GO.html


http://www.webology.ir/2006/v3n3/a28.html

Ontology alignment


License

This work is licensed under the Creative

Commons Attribution-Noncommercial-

Share Alike 3.0 Unported License.

To view a copy of this license, visit

http://creativecommons.org/licenses/by-

nc-sa/3.0/ or send a letter to Creative

Commons, 171 Second Street, Suite 300,

San Francisco, California, 94105, USA.

http://creativecommons.org/licenses/by-nc-sa/3.0/



