Thomas Schmidt

schmidt@informatik.
haw-hamburg.de

Semantic Web –
Introduction and
Problem Statement

• Initial Concepts • Meta Data + RDF

• Initial Problems • Representing Meaning
• Key Perspectives • Ontologies + OWL
• The Role of XML • Evaluating Resources
Thomas Schmidt
Semantic Web: The Idea schmidt@informatik.
haw-hamburg.de

"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."
Tim Berners-Lee, James Hendler, Ora Lassila: “The Semantic Web“.
Scientific American, May 2001

2
Thomas Schmidt
Objectives schmidt@informatik.
haw-hamburg.de

¾ Bring machine processable structure to the bulk of Web

information
¾ Provide a layer of meaningful meta information along with
Web offers to identify their semantics
¾ Provide semantic rules to the community to digest Web
concurrency and allow for conclusions
¾ Offer ways to learn about the reputation of a resource

3
Thomas Schmidt
Semantic Web Layers schmidt@informatik.
haw-hamburg.de

Source: http://www.w3.org/2001/12/semweb-fin/w3csw

4
Thomas Schmidt
Operational Concept schmidt@informatik.
haw-hamburg.de

RI

Ux
R II A2

A1
R III
Ai
Uy
Am
RK

RN

Resources Agents Users 5

Thomas Schmidt
Resources, URIs & Links schmidt@informatik.
haw-hamburg.de

Goal: Understand resources

and their relations
Resources: Anything
addressable by a URI.
Extend resources to carry
a ‘type’ attribute
Links: Relating resources.
Extend links to carry ‘type’
attribute.

6
Thomas Schmidt
Fundamental Problems schmidt@informatik.
haw-hamburg.de

Heterogeneity: Systems, encoding, structures,

languages/expressiveness’, words, meanings, …
Anonymity: Almost all resources in the Web unknown to
recipient
Context: Resources are meaningless without identification
of context
Scale: Peer-to-Peer view has complexity n2, with n =
number of Internet resources
Visions & Expectations: partly naive, partly vague, …
7
Thomas Schmidt
Propagated Visions schmidt@informatik.
haw-hamburg.de

• “The goal of the Semantic Web is to create a universal

medium, which smoothly interconnects personal,
commercial, scientific and cultural data in a machine-
understandable fashion. ”
• “With the Semantic Web we can provide all kinds of
automated services in different domains from future
home and digital libraries to electronic business and
health services.”
• “The Semantic Web taking over completely ones life,
which is the ultimate goal.”
• … space for your profound statement …
8
Thomas Schmidt
Specific Problems schmidt@informatik.
haw-hamburg.de

• Meaningful Buzz Word: Everybody makes up his own

meaning (like ‘Artificial Intelligence’, ‘Chaos Theory’, …).
• The Hype: People without intricate understanding
involved prior to proved results.
• Propagated Visions raise unrealistic expectations.
• Awkward Visions: Sacrificed stuff, people don’t want.
• Ridiculous Personality Cult about Tim Berners-Lee.

9
Thomas Schmidt
What May We Expect? schmidt@informatik.
haw-hamburg.de

• Search & retrieval: Improved search machines and

identification of information
• Data integration: High level tools for rapid/semi
automated data source coupling
• High-level applications in specific fields: Knowledge
management, eLearning, …
• …adaptive distributed systems,
• … original user interfaces for navigation & visualisations

10
Visualisation Example: Thomas Schmidt
schmidt@informatik.
haw-hamburg.de
Cluster Maps

11
Thomas Schmidt
Data-centric Perspective schmidt@informatik.
haw-hamburg.de

• Explore information, discover knowledge

– Find and understand, what resources are about
– Understand and digest content of a web resource
• Evaluate information
– Estimate relevance and reputation
– Judge on precision and correctness
• Integrate information
– Combine data from different sources
– Synchronize different data bases

12
Thomas Schmidt
Structural Heterogeneity schmidt@informatik.
haw-hamburg.de

Different data models with

– Naming / type / integrity conflicts
– Multilateral correspondences
– Missing / redundant / inconsistent data

13
Thomas Schmidt
The Role of XML schmidt@informatik.
haw-hamburg.de

• XML provides standards and transparency

on syntax and encoding
• Plays role of a basic interoperability
mechanism
• Data exchange formally solved with
common DTD/Schema
• XML itself has no semantic definitional
strength
• Meaningful only in communities with
appointed agreements

• Don’t forget: XHTML is mainly without structure!

14
Thomas Schmidt
Semantic Heterogeneity schmidt@informatik.
haw-hamburg.de

• Meta level discrepancies lead to diverging terms

• Data semantics may be divergent € vers. $

15
Thomas Schmidt
How to Solve schmidt@informatik.
haw-hamburg.de
Heterogeneity Problems ?

1. Structural Heterogeneity
o Comparing semantically corresponding data schema entities
o Correlating semantically corresponding data attributes
o Transforming correspondent data types (if possible)
o Special Problem: Aggregation of multilateral correspondences
2. Semantic Heterogeneity
o Detecting semantic correspondences in data

► Semantic is a key issue to solve data heterogeneity

problems
16
Thomas Schmidt
Application-centric schmidt@informatik.
haw-hamburg.de
Perspective

Automated use of applicative resources (e.g. web

services) require answers to:
– What does the application require ?
– How does it work ?
– How is it used ?
Application / community specific approaches:
– OWL-S: Semantic Markup for Web Services
– BPEL: Business Process Execution Language
– …
Focus of the DAML initiative (www.daml.org)
17
Thomas Schmidt
Meta Data (traditional) schmidt@informatik.
haw-hamburg.de

• Provide a (formalized) description about

resources and information
• Commonly organized as (property : values)
maps
• Provide some structure on top of (arbitrary) data,
subject to standardization
• Standards provide definitions on (structured)
properties, occasionally a vocabulary of values
18
Thomas Schmidt
Dublin Core schmidt@informatik.
haw-hamburg.de

• RFC2413 – Simple description scheme (http://dublincore.org)

• Initially minimal consensus from a working group of librarians

Content Intellectual Property Instantiation

Title Creator Date
Subject Publisher Format
Description Contributor Identifier
Type Rights Language
Source
Relation
Coverage
19
Thomas Schmidt
Learning Object Metadata schmidt@informatik.
haw-hamburg.de

• IEEE standard scheme for describing learning objects (LOs)

• Provides a defined, extensible vocabulary in 9 categories

1 General 2 Lifecycle 3 Meta Metadata

General information describing Documenting the history and Groups information about the
the learning object as a whole. the current state of the LO as meta data instance itself.
well as its contributors.

4 Technical 5 Educational 6 Rights

Technical requirements and Allows for a list of educational Intellectual property rights and
characteristics. and pedagogic conditions of use.
characterizations.

7 Relation 8 Annotation 9 Classification

A list of qualified descriptions Comments on educational use Information about this learning
of the relationship between of the learning object. object in relation to a
this instance and other particular classification 20
learning objects. system.
Thomas Schmidt
LOM - General schmidt@informatik.
haw-hamburg.de

1.1 Identifier A globally unique label that identifies this LO.

1.1.1 Catalog The name or designator of the identification or cataloging scheme
for this entry. A namespace scheme.
1.1.2 Entry The value of the identifier within the identification or cataloging
scheme for this entry. A namespace specific string.

1.2 Title Name given to this LO.

1.3 Language The primary human language or languages used within this LO to
communicate to the intended user.
1.4 Description A textual description of the content of this LO.
1.5 Keyword A keyword or phrase describing the topic of this LO.
1.6 Coverage The time, culture, geography or region to which this LO applies.
1.7 Structure Underlying organizational structure of this LO.
1.8 Aggregation level The functional granularity of this LO.
21
Thomas Schmidt
LOM - Educational schmidt@informatik.
haw-hamburg.de

5.1 Interactivity Type Predominant mode of learning supported by this LO.

5.2 Learning Resource Type Specific kind of LO.

5.3 Interactivity Level The degree of interactivity characterizing this LO.

5.4 Semantic Density The degree of conciseness of a LO.

5.5 Intended End User Role Principal user(s) for which this LO was designed.
Context The principal environment within which the learning and use of
5.6 this LO is intended to take place
5.7 Typical Age Range Age of the typical intended user.
Difficulty How hard it is to work with or through this LO for the typical
5.8 intended target audience.
Typical Learning Time Approximate or typical time it takes to work with or through this
5.9 LO for the typical intended target audience.
5.10 Description Comments on how this LO is to be used.
5.11 Language The human language used by the typical intended user.
22
Thomas Schmidt
LOM - Relation schmidt@informatik.
haw-hamburg.de

7.1 Kind Nature of relationship between this LO and the target

LO identified by 7.2.
7.2 Resource The target LO object that this relationship references.
7.2.1 Identifier A globally unique label that identifies the target LO.
7.2.1.1 Catalog The name or designator of the identification or
cataloging scheme for this entry. A namespace
scheme.
7.2.1.2 Entry The value of the identifier within the identification or
cataloging scheme for this entry. A namespace specific
string.

7.2.2 Description Description of the target LO.

23
Thomas Schmidt
Meta Data Extraction schmidt@informatik.
haw-hamburg.de

Subject Predicat

This page is titled

hamster diseases

Object

24
Thomas Schmidt
Resource Description schmidt@informatik.
haw-hamburg.de
Framework (RDF)

• Performs statements about resources

• Statements as triple
– Subject + Predicat + Object
• Maps information directly and unambiguously to a
decentralized model
• URIs used to name objects and concepts
• Graphical representation as semantic nets
• Syntax independent, but usually XML
• Two (syntactically differing) expressions equal if RDF
models coincide
25
Thomas Schmidt
RDF Example schmidt@informatik.
haw-hamburg.de

• Statements:
– Subject: Resource
– Predicat: Property
– Object: Literal

“Simon is the author of hay fever handbook”

http://… http://…

26
Thomas Schmidt
RDF Syntax schmidt@informatik.
haw-hamburg.de

<rdf:RDF>
<rdf:Description about="http://... ">
<dc:author xmlns:dc="http://purl.org/dc">
Simon
</dc:author>
</rdf:Description>
</rdf:RDF>

• XML encoding
• Standard allows for abbreviation
27
Thomas Schmidt
schmidt@informatik.
Roles of XML and RDF haw-hamburg.de

in the Semantic Web

• Expressive
XML
RDF • Syntactically interoperable

• Semantically interoperable

28
Thomas Schmidt
Representing Meaning schmidt@informatik.
haw-hamburg.de

• Words represent meaning

• Dictionaries define the meaning of words
• Problem: Many-to-many relation between words and
meanings
– There may be many words for one meaning
– Words may have many, very distinct meanings
• Solution: Employment of controlled vocabularies
– Pre-selected words used in pre-appointed meanings
• Approaches: use taxonomies and thesauri as present
29
Thomas Schmidt
Taxonomies schmidt@informatik.
haw-hamburg.de

• A taxonomy is a hierarchy of notions, representing a

systematic classification for a collection of entities
• Tree represents specialisation/generalisation
• Association represents an ‘is instance of’ relation
• Examples:
– Linnaeus System (biology)
– ACM Computer Science Index
– Dewey Decimal Classification - DDC
– North American Industry Classification System - NAICS
• Expressiveness: all categories in classified structure of
a given context
30
Dewey Decimal Thomas Schmidt
schmidt@informatik.
haw-hamburg.de
Classification
• Classification of general knowledge within 10
main categories and 10 layers of hierarchy
• Designed by Melvil Dewey in 1873 owned by
the OCLC since 1988
• Version 23 today, ≈ 33.000 entities
• Extensible and processable with minimal
effort
• Some problems with structural evolution of
knowledge and local dependencies (e.g. law)

Specialisation
Categorisation

9 0 0 History & Geography

9 1 0 Geography & Travel
9 1 1 Historical Geography
31
Thomas Schmidt
Thesauri schmidt@informatik.
haw-hamburg.de

• A thesaurus is a classification scheme for terms

• A taxonomy of terms (from a given language) +
additional semantic relations:
– Hierarchical (broader : narrower)
– Equivalence (synonym : antonym)
– Homographic
– Associative
• Useful to extend a controlled vocabulary
• Example: Roget’s Thesaurus (www.gutenberg.net)
• Expressiveness: vocabulary and its basic semantic
relations for (parts of) a given language
32
Thomas Schmidt
Example: WordNet schmidt@informatik.
haw-hamburg.de
http://www.cogsci.princeton.edu/~wn/
Results for "Hypernyms (this is a kind of...)" search of noun "buster"
5 senses of buster
…
buster, broncobuster -- (a person who breaks horses)
=> horseman, equestrian, horseback rider -- (a man skilled in equitation)
=> rider -- (a traveler who actively rides an animal (as a horse or camel))
=> traveler, traveller -- (a person who changes location)
=> person, individual, someone, somebody, mortal, human, soul a human being;
"there was too much for one person to do")
=> organism, being --(a living thing that has (or can develop) the ability
to act or function independently)
=> living thing, animate thing -- (a living (or once living) entity)
=> object, physical object a tangible and visible entity; an entity that can
cast a shadow; ”it was full of rackets, balls and other objects")
=> entity -- (that which is perceived or known or inferred to have
its own distinct existence (living or nonliving))
=> causal agent, cause, causal agency -- (any entity that
causes events to happen)
=> entity -- (that which is perceived or known or inferred to have33
its own distinct existence (living or nonliving))
Thomas Schmidt
TopicMaps schmidt@informatik.
haw-hamburg.de

• Originally form ISO standard (HyTime 19744)

• Goals
– Intelligent information retrieval and subsequent
processing
– Accessing a semantic network of knowledge
– Putting hypertext in semantic relations

34
Thomas Schmidt
Entities of TopicMaps schmidt@informatik.
haw-hamburg.de

• Topic
– in hierarchy
– Topic type
– Public Subject Descriptor with identity attribute
– Scope
• Occurrences: Links to external resources
• Associations: Relations between topics
• Facets: Name-value pairs attributed to Topics or
Associations
35
Thomas Schmidt
TopicMaps - Example schmidt@informatik.
haw-hamburg.de

36
Thomas Schmidt
Ontologies schmidt@informatik.
haw-hamburg.de

• A formal, explicit specification of a shared

conceptualization [Gruber93]
• Define formal semantics for information
• Define real-world semantics
• Pushed in artificial intelligence for knowledge sharing
and re-use
• Description of knowledge domains:
– Standardized terms (Classes, Axioms, etc.)
– Relations between concepts
– Inference rules

37
[Gruber93] http://ksl-web.stanford.edu/KSL_Abstracts/KSL-92-71.html
Thomas Schmidt
Ontology Example schmidt@informatik.
haw-hamburg.de

38
Thomas Schmidt
OWL schmidt@informatik.
haw-hamburg.de

• Web Ontology Language:

– W3C standard
– Semantic markup language for publishing and sharing
ontologies on the web
– Successor of DAML+OIL
• Goals:
– Mapping of Relations between vocables
– Machine processable description of coherences
• Realisation: Extension of RDF
39
Thomas Schmidt
OWL Overview schmidt@informatik.
haw-hamburg.de

• OWL Lite:
– Simple expression of term hierarchies
– Cardinality 0 or 1
• OWL DL (description logics):
– Maximal expressivness while finitely computable
– Some restrictions on nesting
• OWL Full:
– Full expressiveness
– No guarantied computablility

40
Thomas Schmidt
OWL Example schmidt@informatik.
haw-hamburg.de

<owl:Class rdf:ID='‘Snake''>
<rdfs:subClassOf rdf:resource=''#Animal''/>
</owl:Class>
<owl:Class rdf:ID=''Hamster''>
<rdfs:subClassOf rdf:resource=''#Animal''/>
<rdfs:subClassOf>
<owl:Restriction>
<owl:onProperty rdf:resource=''#hasParent''/>
<owl:allValuesFrom rdf:resource=''#Hamster''/>
</owl:Restriction>
</rdfs:subClassOf>
<owl:disjointWith rdf:resource=''#Snake''/>
</owl:Class>
41
Thomas Schmidt
schmidt@informatik.
Redefined Notion of Metadata haw-hamburg.de

In the Semantic Web all available descriptors are needed

as digestive Metadata!
42
Thomas Schmidt
Evaluating Resources schmidt@informatik.
haw-hamburg.de

To judge on received information we need to evaluate:

• Hard states:
– Authenticity of the source/author
– Integrity of data
• Soft states:
– Validity and reliability
– Relevance
– Context
– Trustworthiness

43
Thomas Schmidt
W3C: Web of Trust schmidt@informatik.
haw-hamburg.de

• Recipient determines
group of trustees
• Trust can be inherited
linearly according to rules
• Needs some certification
(PKI, fingerprints …)
• Derived from the CA
approach

Poor Man’s Logic: This approach is hierarchical - chains need

unconditional trustworthiness at their roots!
44
Thomas Schmidt
Alternate Approach: schmidt@informatik.
haw-hamburg.de
Network Analysis

Explicit:
• Evaluate statements on
your issue:
“Dwayne you can trust”
“Kilgore pays promptly”
Implicit:
• Evaluate statements and
relations at hand
• Draw conclusions:
“Donald Knuth is Professor at Stanford, thus I believe him.”
“Tim Berners-Lee is mentioned many times and in ‘Network
Hubs’, he thus must be famous.”
45
Thomas Schmidt
The Problem of Context schmidt@informatik.
haw-hamburg.de

There are two contexts to consider

Context of creation:
– Donald Knuth writes on Surreal Numbers and Diamond Signs
– R. Gernhardt links the words “My Favourite” to M. Reich-Ranickis
Book page from a paragraph titled “Most Awkward Publications”
Context of reception:
– “I know the works of Knuth, but am looking for young talents”
– “1 billion of Chinese think that something is good, but my
Grandma does not”
Problem: Identify contexts and judge on their compatibility/agreement.
46
Thomas Schmidt
The Problem of Time schmidt@informatik.
haw-hamburg.de

Reputation of a resource (person, institution, agent, …) is

a function of time …
Example: Konrad Zuse, the well reputed pioneer of
computer systems, published papers ‘of lesser renown’
on fundamental physics in its later age.
In general: The reputation of a resource is an expectation
about its current behaviour based on information about
or observations of its past behaviour.

47
Thomas Schmidt
The Problem of schmidt@informatik.
haw-hamburg.de
Induced Biases
Implicit:
– Structural inheritance: URI of D. Knuth’s homepage could be of
identical structural formation as some technical staff (is not)
– General problem: How to account for the deep Web
Explicit:
– Trust inflation: People/institutions granting plentiful amounts of
reputations
– Destructive groups: Groups injecting ‘consistent falseness’ on
large scale
– Large players: Players owning many Web sites may enforce
self-exaltation
– Software vendors/pirates: Leading software vendors (or
software pirates) may (self-)reinforce by ‘default settings’
48
Thomas Schmidt
Résumé on Resource schmidt@informatik.
haw-hamburg.de
Evaluation, Reputation & Trust

• Not simple at all

• Would like a global PKI
• Suffers from conceptual unclearness in basic Web
semantics: Contexts, Links, …
• Suffers from the certainty about a persistent
structural chaos in the Web
• Some promising heuristics
• Active research area
49
Thomas Schmidt
References schmidt@informatik.
haw-hamburg.de
ª Semantic Web @ W3C - http://www.w3.org/2001/sw/
ª Semantic Web – http://www.semanticweb.org/
ª Daconta, Obrst, Smith: The Semantic Web, Wiley 2003.
ª B. Thuraisingham: XML Databases and the Semantic Web, CRC
Press, 2002.
ª Ubbo Visser et. al: Web Development, WWW Tutorial May 2004
ª R. Widhalm, Th. Mück: Topic Maps, Springer 2002.
ª D. Fensel: Ontologies, Springer 2004.
ª RDF Primer - http://www.w3.org/TR/rdf-primer/
ª RDF Model & Abstract Syntax - http://www.w3.org/TR/rdf-concepts/
ª Dublin Core – http://dublincore.org
ª DDC - http://www.oclc.org/dewey/default.htm
http://www.gutenberg.net/dirs/1/2/5/1/12513/12513-h/12513-h.htm
50