1

Introduction to the Semantic Web

(tutorial)

2009 Semantic Technology Conference

San Jose, California, USA
June 15, 2009

Ivan Herman, W3C

[email protected]
 2

Introduction
 3

Let’s organize a trip to Budapest using the

Web!
 4

You try to find a proper flight with …

 5

… a big, reputable airline, or …

 6

… the airline of the target country, or …

 7

… or a low cost one

 8

You have to find a hotel, so you look for…

 9

… a really cheap accommodation, or …

 10

… or a really luxurious one, or …

 11

… an intermediate one …
 12

oops, that is no good, the page is in

Hungarian that almost nobody
understands, but…
 13

… this one could work

 14

Of course, you could decide to trust a

specialized site…
 15

… like this one, or…

 16

… or this one
 17

You may want to know something about

Budapest; look for some photographs…
 18

… on flickr …
 19

… on Google …
 20

… or you can look at mine

 21

… or a (social) travel site

 22

What happened here?

 You had to consult a large number of sites, all
different in style, purpose, possibly language…
 You had to mentally integrate all those information
to achieve your goals
 We all know that, sometimes, this is a long and
tedious process!
 23

 All those pages are only tips of respective icebergs:

 the real data is hidden somewhere in databases, XML files,
Excel sheets, …
 you have only access to what the Web page designers allow
you to see
 24

 Specialized sites (Expedia, TripAdvisor) do a bit

more:
 they gather and combine data from other sources (usually
with the approval of the data owners)
 but they still control how you see those sources
 But sometimes you want to personalize: access the
original data and combine it yourself!
 25

Here is another example…

 26

Another example: social sites. I have a list

of “friends” by…
 27

… Dopplr,
 28

… Twine,
 29

… LinkedIn,
 30

… and, of course, Facebook

 31

 I had to type in and connect with friends again and

again for each site independently
 This is even worse then before: I feed the icebergs,
but I still do not have an easy access to data…
 32

What would we like to have?

 Use the data on the Web the same way as we do
with documents:
 be able to link to data (independently of their presentation)
 use that data the way I want (present it, mine it, etc)
 agents, programs, scripts, etc, should be able to interpret
part of that data
 33

Put it another way…

 We would like to extend the current Web to a “Web
of data”:
 allow for applications to exploit the data directly
 34

But wait! Isn’t what mashup sites are

already doing?
 35

A “mashup” example:
 36

 In some ways, yes, and that shows the huge power

of what such Web of data provides
 But mashup sites are forced to do very ad-hoc jobs
 various data sources expose their data via Web Services
 each with a different API, a different logic, different structure
 these sites are forced to reinvent the wheel many times
because there is no standard way of doing things
 37

Put it another way (again)…

 We would like to extend the current Web to a
standard way for a “Web of data”
 38

But what does this mean?

 What makes the current (document) Web work?

 people create different documents
 they give an address to it (ie, a URI) and make it accessible
to others on the Web
 Steven’s site on Amsterdam
39

(done for some visiting friends)

 40

Then some magic happens…

 Others discover the site and they link to it
 The more they link to it, the more important and
well known the page becomes
 remember, this is what, eg, Google exploits!
 This is the “Network effect”: some pages become
important, and others begin to rely on it even if the
author did not expect it…
 41

This could be expected…

but this one, from the other side of the Globe,
42

was not…
 43

What would that mean for a Web of Data?

 Lessons learned: we should be able to:
 “publish” the data to make it known on the Web
 standard ways should be used instead of ad-hoc approaches
 the analogous approach to documents: give URI-s to the data
 make it possible to “link” to that URI from other sources of
data (not only Web pages)
 ie, applications should not be forced to make targeted
developments to access the data
 generic, standard approaches should suffice
 and let the network effect work its way…
 44

But it is a little bit more complicated

 On the traditional Web, humans are implicitly taken
into account
 A Web link has a “context” that a person may use
 45

Eg: address field on my page:

 46

… leading to this page

 47

 A human understands that this is my institution’s

home page
 He/she knows what it means (realizes that it is a
research institute in Amsterdam)
 On a Web of Data, something is missing; machines
can’t make sense of the link alone
 48

 New lesson learned:

 extra information (“label”) must be added to a link: “this links
to my institution, which is a research institute”
 this information should be machine readable
 this is a characterization (or “classification”) of both the link
and its target
 in some cases, the classification should allow for some
limited “reasoning”
 49

Let us put it together

 What we need for a Web of Data:
 use URI-s to publish data, not only full documents
 allow the data to link to other data
 characterize/classify the data and the links (the “terms”) to
convey some extra meaning
 and use standards for all these!
 50

So What is the Semantic Web?

 51

It is a collection of standard technologies

to realize a Web of Data
 52

 It is that simple…
 Of course, the devil is in the details
 a common model has to be provided for machines to
describe, query, etc, the data and their connections
 the “classification” of the terms can become very complex
for specific knowledge areas: this is where ontologies,
thesauri, etc, enter the game…
 53

In what follows…
 We will use a simplistic example to introduce the
main technical concepts
 The details will be for later during the course
 54

The rough structure of data integration

1. Map the various data onto an abstract data
representation
 make the data independent of its internal representation…
2. Merge the resulting representations
3. Start making queries on the whole!
 queries that could not have been done on the individual data
sets
 55

A simplified bookstore data (dataset “A”)

ID Author Title Publisher Year
ISBN0-00-651409-X id_xyz The Glass Palace id_qpr 2000

ID Name Home Page

id_xyz Ghosh, Amitav http://www.amitavghosh.com

ID Publ. Name City

id_qpr Harper Collins London
 56

1st: export your data as a set of relations

 57

Some notes on the exporting the data

 Relations form a graph
 the nodes refer to the “real” data or contain some literal
 how the graph is represented in machine is immaterial for
now
 Data export does not necessarily mean physical
conversion of the data
 relations can be generated on-the-fly at query time
 via SQL “bridges”
 scraping HTML pages
 extracting data from Excel sheets
 etc.
 One can export part of the data
 58

Another bookstore data (dataset “F”)

A B D E

1 ID Titre Traducteur Original

ISBN0 2020386682 Le Palais A13 ISBN-0-00-651409-X
des
2 miroirs
3

6 ID Auteur
7 ISBN-0-00-651409-X A12

11 Nom
12 Ghosh, Amitav
13 Besse, Christianne
 59

2nd: export your second set of data

 60

3rd: start merging your data

 61

3rd: start merging your data (cont.)

 62

3rd: merge identical resources

 63

Start making queries…

 User of data “F” can now ask queries like:
 “give me the title of the original”
 well, … « donnes-moi le titre de l’original »
 This information is not in the dataset “F”…
 …but can be retrieved by merging with dataset “A”!
 64

However, more can be achieved…

 We “feel” that a:author and f:auteur should be
the same
 But an automatic merge doest not know that!
 Let us add some extra information to the merged
data:
 a:author same as f:auteur
 both identify a “Person”
 a term that a community may have already defined:
 a “Person” is uniquely identified by his/her name and, say,
homepage
 it can be used as a “category” for certain type of resources
 65

3rd revisited: use the extra knowledge

 66

Start making richer queries!

 User of dataset “F” can now query:
 “donnes-moi la page d’accueil de l’auteur de l’originale”
 well… “give me the home page of the original’s ‘auteur’”
 The information is not in datasets “F” or “A”…
 …but was made available by:
 merging datasets “A” and datasets “F”
 adding three simple extra statements as an extra “glue”
 67

Combine with different datasets

 Using, e.g., the “Person”, the dataset can be
combined with other sources
 For example, data in Wikipedia can be extracted
using dedicated tools
 e.g., the “dbpedia” project can extract the “infobox”
information from Wikipedia already…
 68

Merge with Wikipedia data

 69

Merge with Wikipedia data

 70

Merge with Wikipedia data

 71

Is that surprising?
 It may look like it but, in fact, it should not be…
 What happened via automatic means is done every
day by Web users!
 The difference: a bit of extra rigour so that
machines could do this, too
 72

What did we do?

 We combined different datasets that
 are somewhere on the web
 are of different formats (mysql, excel sheet, XHTML, etc)
 have different names for relations
 We could combine the data because some URI-s
were identical (the ISBN-s in this case)
 We could add some simple additional information
(the “glue”), possibly using common terminologies
that a community has produced
 As a result, new relations could be found and
retrieved
 73

It could become even more powerful

 We could add extra knowledge to the merged
datasets
 e.g., a full classification of various types of library data
 geographical information
 etc.
 This is where ontologies, extra rules, etc, come in
 ontologies/rule sets can be relatively simple and small, or
huge, or anything in between…
 Even more powerful queries can be asked as a
result
 74

What did we do? (cont)

 75

The Basis: RDF

 76

RDF triples
 Let us begin to formalize what we did!
 we “connected” the data…
 but a simple connection is not enough… data should be
named somehow
 hence the RDF Triples: a labelled connection between two
resources
 77

RDF triples (cont.)

 An RDF Triple (s,p,o) is such that:
 “s”, “p” are URI-s, ie, resources on the Web; “o” is a URI or
a literal
 “s”, “p”, and “o” stand for “subject”, “property”, and “object”
 here is the complete triple:

(<http://…isbn…6682>,
(<http://…isbn…6682>, <http://…/original>,
<http://…/original>, <http://…isbn…409X>)
<http://…isbn…409X>)

 RDF is a general model for such triples (with

machine readable formats like RDF/XML, Turtle,
N3, RXR, …)
 78

RDF triples (cont.)

 Resources can use any URI, e.g.:
 http://www.example.org/file.xml#element(home)
 http://www.example.org/file.html#home
 http://www.example.org/file2.xml#xpath1(//q[@a=b])
 URI-s can also denote non Web entities:
 http://www.ivan-herman.net/me is me
 not my home page, not my publication list, but me
 RDF triples form a directed, labelled graph
 79

A simple RDF example (in RDF/XML)

<rdf:Description
<rdf:Description rdf:about="http://…/isbn/2020386682">
rdf:about="http://…/isbn/2020386682">
<f:titre
<f:titre xml:lang="fr">Le palais
xml:lang="fr">Le palais des
des mirroirs</f:titre>
mirroirs</f:titre>
<f:original
<f:original rdf:resource="http://…/isbn/000651409X"/>
rdf:resource="http://…/isbn/000651409X"/>
</rdf:Description>
</rdf:Description>

(Note: namespaces are used to simplify the URI-s)

 80

A simple RDF example (in Turtle)

<http://…/isbn/2020386682>
<http://…/isbn/2020386682>
f:titre
f:titre "Le
"Le palais
palais des
des mirroirs"@fr
mirroirs"@fr ;;
f:original
f:original <http://…/isbn/000651409X>
<http://…/isbn/000651409X> ..
 81

“Internal” nodes
 Consider the following statement:
 “the publisher is a «thing» that has a name and an address”
 Until now, nodes were identified with a URI. But…
 …what is the URI of «thing»?
 82

Internal identifier (“blank nodes”)

<rdf:Description
<rdf:Description rdf:about="http://…/isbn/000651409X">
rdf:about="http://…/isbn/000651409X">
<a:publisher rdf:nodeID="A234"/>
<a:publisher rdf:nodeID="A234"/>
</rdf:Description>
</rdf:Description>
<rdf:Description
<rdf:Description rdf:nodeID="A234">
rdf:nodeID="A234">
<a:p_name>HarpersCollins</a:p_name>
<a:p_name>HarpersCollins</a:p_name>
<a:city>HarpersCollins</a:city>
<a:city>HarpersCollins</a:city>
</rdf:Description>
</rdf:Description>

<http://…/isbn/2020386682>
<http://…/isbn/2020386682> a:publisher
a:publisher _:A234.
_:A234.
_:A234
_:A234 a:p_name
a:p_name "HarpersCollins".
"HarpersCollins".

 Syntax is serialization dependent

 A234 is invisible from outside (it is not a “real”
URI!); it is an internal identifier for a resource
 83

Blank nodes: the system can also do it

 Let the system create a “nodeID” internally (you do
not really care about the name…)
<rdf:Description
<rdf:Description rdf:about="http://…/isbn/000651409X">
rdf:about="http://…/isbn/000651409X">
<a:publisher>
<a:publisher>
<rdf:Description>
<rdf:Description>
<a:p_name>HarpersCollins</a:p_name>
<a:p_name>HarpersCollins</a:p_name>
……
</rdf:Description>
</rdf:Description>
</a:publisher>
</a:publisher>
</rdf:Description>
</rdf:Description>
 84

Same in Turtle

<http://…/isbn/000651409X>
<http://…/isbn/000651409X> a:publisher
a:publisher [[
a:p_name
a:p_name "HarpersCollins";
"HarpersCollins";
……
].
].
 85

Blank nodes: some more remarks

 Blank nodes require attention when merging
 blanks nodes with identical nodeID-s in different graphs are
different
 implementations must be careful…
 Many applications prefer not to use blank nodes
and define new URI-s “on-the-fly”
 86

RDF in programming practice

 For example, using Java+Jena (HP’s Bristol Lab):
 a “Model” object is created
 the RDF file is parsed and results stored in the Model
 the Model offers methods to retrieve:
 triples
 (property,object) pairs for a specific subject
 (subject,property) pairs for specific object
 etc.
 the rest is conventional programming…
 Similar tools exist in Python, PHP, etc.
 87

Jena example

//
// create
create aa model
model
Model
Model model=new
model=new ModelMem();
ModelMem();
Resource subject=model.createResource("URI_of_Subject")
Resource subject=model.createResource("URI_of_Subject")
//
// 'in'
'in' refers
refers to
to the
the input
input file
file
model.read(new InputStreamReader(in));
model.read(new InputStreamReader(in));
StmtIterator
StmtIterator iter=model.listStatements(subject,null,null);
iter=model.listStatements(subject,null,null);
while(iter.hasNext())
while(iter.hasNext()) {{
st
st == iter.next();
iter.next();
pp == st.getProperty();
st.getProperty();
oo == st.getObject();
st.getObject();
do_something(p,o);
do_something(p,o);
}}
 88

Merge in practice
 Environments merge graphs automatically
 e.g., in Jena, the Model can load several files
 the load merges the new statements automatically
 89

Example: integrate experimental data

 Goal: reuse of older

experimental data
 Keep data in
databases or XML,
just export key “fact”
as RDF
 Use a faceted
browser to visualize
and interact with the
result

Courtesy of Nigel Wilkinson, Lee Harland, Pfizer Ltd, Melliyal Annamalai, Oracle (SWEO Case Study)
 90

One level higher up

(RDFS, Datatypes)
 91

Need for RDF schemas

 First step towards the “extra knowledge”:
 define the terms we can use
 what restrictions apply
 what extra relationships are there?
 Officially: “RDF Vocabulary Description Language”
 the term “Schema” is retained for historical reasons…
 92

Classes, resources, …
 Think of well known traditional ontologies or
taxonomies:
 use the term “novel”
 “every novel is a fiction”
 “«The Glass Palace» is a novel”
 etc.
 RDFS defines resources and classes:
 everything in RDF is a “resource”
 “classes” are also resources, but…
 …they are also a collection of possible resources (i.e.,
“individuals”)
 “fiction”, “novel”, …
 93

Classes, resources, … (cont.)

 Relationships are defined among classes and
resources:
 “typing”: an individual belongs to a specific class
 “«The Glass Palace» is a novel”
 to be more precise: “«http://.../000651409X» is a novel”
 “subclassing”: all instances of one are also the instances of
the other (“every novel is a fiction”)
 RDFS formalizes these notions in RDF
 94

Classes, resources in RDF(S)

 RDFS defines the meaning of these terms

 (these are all special URI-s, we just use the namespace
abbreviation)
 95

Schema example in RDF/XML

 The schema part:
<rdf:Description
<rdf:Description rdf:ID="Novel">
rdf:ID="Novel">
<rdf:type
<rdf:type
rdf:resource="http://www.w3.org/2000/01/rdf-schema#Class"/>
rdf:resource="http://www.w3.org/2000/01/rdf-schema#Class"/>
</rdf:Description>
</rdf:Description>

 The RDF data on a specific novel:

<rdf:Description
<rdf:Description rdf:about="http://…/isbn/000651409X">
rdf:about="http://…/isbn/000651409X">
<rdf:type rdf:resource="http://…/bookSchema.rdf#Novel"/>
<rdf:type rdf:resource="http://…/bookSchema.rdf#Novel"/>
</rdf:Description>
</rdf:Description>
 96

Further remarks on types

 A resource may belong to several classes
 rdf:type is just a property…
 “«The Glass Palace» is a novel, but «The Glass Palace» is
also an «inventory item»…”
 i.e., it is not like a datatype!
 The type information may be very important for
applications
 e.g., it may be used for a categorization of possible nodes
 probably the most frequently used RDF property…
 (remember the “Person” in our example?)
 97

Inferred properties

(<http://…/isbn/000651409X>
(<http://…/isbn/000651409X> rdf:type
rdf:type #Fiction)
#Fiction)

 is not in the original RDF data…

 …but can be inferred from the RDFS rules
 RDFS environments return that triple, too
 98

Inference: let us be formal…

 The RDF Semantics document has a list of (33)
entailment rules:
 “if such and such triples are in the graph, add this and this”
 do that recursively until the graph does not change
 The relevant rule for our example:

If:
If:
uuu
uuu rdfs:subClassOf
rdfs:subClassOf xxx
xxx ..
vvv
vvv rdf:type
rdf:type uuu
uuu ..
Then
Then add:
add:
vvv
vvv rdf:type
rdf:type xxx
xxx ..
 99

Properties
 Property is a special class (rdf:Property)
 properties are also resources identified by URI-s
 There is also a possibility for a “sub-property”
 all resources bound by the “sub” are also bound by the other
 Range and domain of properties can be specified
 i.e., what type of resources serve as object and subject
 100

Property specification serialized

 In RDF/XML:
<rdf:Property
<rdf:Property rdf:ID="title">
rdf:ID="title">
<rdfs:domain
<rdfs:domain rdf:resource="#Fiction"/>
rdf:resource="#Fiction"/>
<rdfs:range rdf:resource="http://...#Literal"/>
<rdfs:range rdf:resource="http://...#Literal"/>
</rdf:Property>
</rdf:Property>

 In Turtle:
:title
:title
rdf:type
rdf:type rdf:Property;
rdf:Property;
rdfs:domain :Fiction;
rdfs:domain :Fiction;
rdfs:range
rdfs:range rdfs:Literal.
rdfs:Literal.
 101

What does this mean?

 Again, new relations can be deduced. Indeed, if
:title
:title
rdf:type
rdf:type rdf:Property;
rdf:Property;
rdfs:domain :Fiction;
rdfs:domain :Fiction;
rdfs:range
rdfs:range rdfs:Literal.
rdfs:Literal.
<http://…/isbn/000651409X>
<http://…/isbn/000651409X> :title
:title "The
"The Glass
Glass Palace"
Palace" ..

 then the system can infer that:

<http://…/isbn/000651409X>
<http://…/isbn/000651409X> rdf:type
rdf:type :Fiction
:Fiction ..
 102

Literals
 Literals may have a data type
 floats, integers, booleans, etc, defined in XML Schemas
 full XML fragments
 (Natural) language can also be specified
 103

Examples for datatypes

<http://…/isbn/000651409X>
<http://…/isbn/000651409X>
:page_number
:page_number "543"^^xsd:integer
"543"^^xsd:integer ;;
:publ_date
:publ_date "2000"^^xsd:gYear
"2000"^^xsd:gYear ;;
:price
:price "6.99"^^xsd:float
"6.99"^^xsd:float ..
 104

A bit of RDFS can take you far…

 Remember the power of merge?
 We could have used, in our example:
 f:auteur is a subproperty of a:author and vice versa
(although we will see other ways to do that…)
 Of course, in some cases, more complex
knowledge is necessary (see later…)
 105

Example: find the right experts at NASA

 Expertise locater for nearly 70,000 NASA civil
servants, using RDF integration techniques over 6
or 7 geographically distributed databases, data
sources, and web services…

Michael Grove, Clark & Parsia, LLC, and Andrew Schain, NASA, (SWEO Case Study)
 106

How to get RDF Data?

(Microformats, GRDDL, RDFa)
 107

Simple approach
 Write RDF/XML or Turtle “manually”
 In some cases that is necessary, but it really does
not scale…
 108

RDF with XHTML

 Obviously, a huge source of information
 By adding some “meta” information, the same
source can be reused for, eg, data integration,
better mashups, etc
 typical example: your personal information, like address,
should be readable for humans and processable by
machines
 Two solutions have emerged:
 extract the structure from the page and convert the content
into RDF
 add RDF statements directly into XHTML via RDFa
 109

Extract RDF
 Use intelligent “scrapers” or “wrappers” to extract a
structure (hence RDF) from a Web pages or XML
files…
 … and then generate RDF automatically (e.g., via
an XSLT script)
 110

Formalizing the scraper approach: GRDDL

 GRDDL formalizes the scraper approach. For
example:
<html
<html xmlns="http://www.w3.org/1999/">
xmlns="http://www.w3.org/1999/">
<head
<head profile="http://www.w3.org/2003/g/data-view">
profile="http://www.w3.org/2003/g/data-view">
<title>Some
<title>Some Document</title>
Document</title>
<link
<link rel="transformation" href="http:…/dc-extract.xsl"/>
rel="transformation" href="http:…/dc-extract.xsl"/>
<meta
<meta name="DC.Subject" content="Some subject"/>
name="DC.Subject" content="Some subject"/>
...
...
</head>
</head>
...
...
<span
<span class="date">2006-01-02</span>
class="date">2006-01-02</span>
...
...
</html>
</html>

 yields, through dc-extract.xsl:

<>
<>
dc:subject
dc:subject "Some
"Some subject";
subject";
dc:date
dc:date "2006-01-02" ..
"2006-01-02"
 111

GRDDL
 The transformation itself has to be provided for
each set of conventions
 A more general syntax is defined for XML formats
in general (e.g., via the namespace document)
 a method to get data in other formats to RDF (e.g., XBRL)
 112

Example for “structure”: microformats

 Not a Semantic Web specification, originally
 there is a separate microformat community
 Approach: re-use (X)HTML attributes and elements
to add “meta” information
 typically @abbr, @class, @title, …
 different community agreements for different applications
 113

RDFa
 RDFa extends (X)HTML a bit by:
 defining general attributes to add metadata to any elements
 provides an almost complete “serialization” of RDF in
XHTML
 It is a bit like the microformats/GRDDL approach
but fully generic
 114

RDFa example
 For example:
<div
<div about="http://uri.to.newsitem">
about="http://uri.to.newsitem">
<span
<span property="dc:date">March
property="dc:date">March 23,
23, 2004</span>
2004</span>
<span
<span property="dc:title">Rollers hit casino
property="dc:title">Rollers hit casino for
for £1.3m</span>
£1.3m</span>
By
By <span property="dc:creator">Steve Bird</span>. See
<span property="dc:creator">Steve Bird</span>. See
<a href="http://www.a.b.c/d.avi" rel="dcmtype:MovingImage">
<a href="http://www.a.b.c/d.avi" rel="dcmtype:MovingImage">
also
also video
video footage</a>…
footage</a>…
</div>
</div>

 yields, through an RDFa processor:

<http://uri.to.newsitem>
<http://uri.to.newsitem>
dc:date
dc:date "March
"March 23,
23, 2004";
2004";
dc:title
dc:title "Rollers
"Rollers hit casino
hit casino for
for £1.3m;
£1.3m;
dc:creator
dc:creator "Steve
"Steve Bird";
Bird";
dcmtype:MovingImage <http://www.a.b.c/d.avi>.
dcmtype:MovingImage <http://www.a.b.c/d.avi>.
 115

Example: Yahoo’s SearchMonkey

 Search based results may be customized via small
applications
 Metadata in pages (in RDFa, microformats etc) are
reused

Courtesy of Peter Mika, Yahoo! Research, (SWEO Case Study)

Example: RDFa data by the
116

London Gazette
 117

Example: RDFa data by the

London Gazette
 118

Bridge to relational databases

 Data on the Web are mostly stored in databases
 “Bridges” are being defined:
 a layer between RDF and the relational data
 RDB tables are “mapped” to RDF graphs, possibly on the fly
 different mapping approaches are being used
 a number RDB systems offer this facility already (eg,
Oracle, OpenLink, …)
 A survey on mapping techniques has been
published at W3C
 W3C plans to engage in a standardization work in
this area
 119

Linking Data
 120

Linking Open Data Project

 Goal: “expose” open datasets in RDF
 Set RDF links among the data items from different
datasets
 Set up query endpoints
 Altogether billions of triples, millions of links…
 121

Example data source: DBpedia

 DBpedia is a community effort to
 extract structured (“infobox”) information from Wikipedia
 provide a query endpoint to the dataset
 interlink the DBpedia dataset with other datasets on the
Web
 122

Extracting Wikipedia structured data

@prefix
@prefix dbpedia
dbpedia <http://dbpedia.org/resource/>.
<http://dbpedia.org/resource/>.
@prefix
@prefix dbterm
dbterm <http://dbpedia.org/property/>.
<http://dbpedia.org/property/>.

dbpedia:Amsterdam
dbpedia:Amsterdam
dbterm:officialName
dbterm:officialName “Amsterdam”
“Amsterdam” ;;
dbterm:longd
dbterm:longd “4”
“4” ;;
dbterm:longm
dbterm:longm “53”
“53” ;;
dbterm:longs
dbterm:longs “32”
“32” ;;
...
...
dbterm:leaderTitle
dbterm:leaderTitle “Mayor”
“Mayor” ;;
dbterm:leaderName
dbterm:leaderName dbpedia:Job_Cohen
dbpedia:Job_Cohen ;;
...
...
dbterm:areaTotalKm
dbterm:areaTotalKm “219”
“219” ;;
...
...
dbpedia:ABN_AMRO
dbpedia:ABN_AMRO
dbterm:location
dbterm:location dbpedia:Amsterdam
dbpedia:Amsterdam ;;
...
...
 123

Automatic links among open datasets

<http://dbpedia.org/resource/Amsterdam>
<http://dbpedia.org/resource/Amsterdam>
owl:sameAs
owl:sameAs <http://rdf.freebase.com/ns/...>
<http://rdf.freebase.com/ns/...> ;;
owl:sameAs
owl:sameAs <http://sws.geonames.org/2759793>
<http://sws.geonames.org/2759793> ;;
...
...

<http://sws.geonames.org/2759793>
<http://sws.geonames.org/2759793>
owl:sameAs
owl:sameAs <http://dbpedia.org/resource/Amsterdam>
<http://dbpedia.org/resource/Amsterdam>
wgs84_pos:lat
wgs84_pos:lat “52.3666667”
“52.3666667” ;;
wgs84_pos:long
wgs84_pos:long “4.8833333”
“4.8833333” ;;
geo:inCountry
geo:inCountry <http://www.geonames.org/countries/#NL>
<http://www.geonames.org/countries/#NL> ;;
...
...

Processors can switch automatically from one to the other…

 124

The LOD “cloud”, March 2008

 125

The LOD “cloud”, September 2008

 126

The LOD “cloud”, March 2009

 127

Example: mapping application on an iPhone

Courtesy of Chris Bizer and Christian Becker, Freie Universität, Berlin

 128

Example: mapping application on an iPhone

Courtesy of Chris Bizer and Christian Becker, Freie Universität, Berlin

 129

Query RDF Data

(SPARQL)
 130

RDF data access

 How do I query the RDF data?
 e.g., how do I get to the DBpedia data?
 131

Querying RDF graphs

 Remember the Jena idiom:
StmtIterator
StmtIterator iter=model.listStatements(subject,null,null);
iter=model.listStatements(subject,null,null);
while(iter.hasNext())
while(iter.hasNext()) {{
st
st == iter.next();
iter.next();
pp == st.getProperty();
st.getProperty(); oo == st.getObject();
st.getObject();
do_something(p,o);
do_something(p,o);

 In practice, more complex queries into the RDF

data are necessary
 something like: “give me the (a,b) pair of resources, for
which there is an x such that (x parent a) and (b brother x)
holds” (ie, return the uncles)
 these rules may become quite complex
 The goal of SPARQL (Query Language for RDF)
 132

Analyse the Jena example

StmtIterator
StmtIterator iter=model.listStatements(subject,null,null);
iter=model.listStatements(subject,null,null);
while(iter.hasNext())
while(iter.hasNext()) {{
st
st == iter.next();
iter.next();
pp == st.getProperty();
st.getProperty(); oo == st.getObject();
st.getObject();
do_something(p,o);
do_something(p,o);

 The (subject,?p,?o) is a pattern for what we

are looking for (with ?p and ?o as “unknowns”)
 133

General: graph patterns

 The fundamental idea: use graph patterns
 the pattern contains unbound symbols
 by binding the symbols, subgraphs of the RDF graph are
selected
 if there is such a selection, the query returns bound
resources
 134

Our Jena example in SPARQL

SELECT
SELECT ?p
?p ?o
?o
WHERE
WHERE {subject ?p
{subject ?p ?o}
?o}

 The triples in WHERE define the graph pattern,

with ?p and ?o “unbound” symbols
 The query returns all p,o pairs
 135

Simple SPARQL example

SELECT
SELECT ?isbn
?isbn ?price
?price ?currency
?currency ## note:
note: not
not ?x!
?x!
WHERE
WHERE {?isbn a:price ?x. ?x rdf:value ?price. ?x
{?isbn a:price ?x. ?x rdf:value ?price. ?x p:currency
p:currency ?currency.}
?currency.}
 136

Simple SPARQL example

SELECT
SELECT ?isbn
?isbn ?price
?price ?currency
?currency ## note:
note: not
not ?x!
?x!
WHERE
WHERE {?isbn a:price ?x. ?x rdf:value ?price. ?x
{?isbn a:price ?x. ?x rdf:value ?price. ?x p:currency
p:currency ?currency.}
?currency.}

 Returns:
[[<..49X>,33,£], [<..49X>,50,€], [<..6682>,60,€],
[<..6682>,78,$]]
 137

Pattern constraints
SELECT
SELECT ?isbn
?isbn ?price
?price ?currency
?currency ## note:
note: not
not ?x!
?x!
WHERE
WHERE { ?isbn a:price ?x. ?x rdf:value ?price. ?x
{ ?isbn a:price ?x. ?x rdf:value ?price. ?x p:currency
p:currency ?currency.
?currency.
FILTER(?currency ==
FILTER(?currency == € }€ }

 Returns: [[<..409X>,50,€], [<..6682>,60,€]]

 138

Other SPARQL features

 Limit the number of returned results; remove
duplicates, sort them, …
 Optional branches in the query
 Specify several data sources (via URI-s) within the
query (essentially, a merge!)
 Construct a graph combining a separate pattern
and the query results
 Use datatypes and/or language tags when
matching a pattern
 139

SPARQL usage in practice

 SPARQL is usually used over the network
 separate documents define the protocol and the result
format
 SPARQL Protocol for RDF with HTTP and SOAP bindings
 SPARQL results in XML or JSON formats
 Big datasets usually offer “SPARQL endpoints”
using this protocol
 typical example: SPARQL endpoint to DBpedia
 140

SPARQL as a unifying point

 141

Example: integrate Chinese medical data

 Integration of a large number of TCM databases
 around 80 databases, around 200,000 records each
 A visual tool to map databases to the semantic
layer using a specialized ontology
 Form based query interface for end users

Courtesy of Huajun Chen, Zhejiang University, (SWEO Case Study)

 142

Ontologies
(OWL)
 143

Ontologies
 RDFS is useful, but does not solve all possible
requirements
 Complex applications may want more possibilities:
 characterization of properties
 identification of objects with different URI-s
 disjointness or equivalence of classes
 construct classes, not only name them
 can a program reason about some terms? E.g.:
 “if «Person» resources «A» and «B» have the same
«foaf:email» property, then «A» and «B» are identical”
 etc.
 144

Ontologies (cont.)
 The term ontologies is used in this respect:

“defines
“defines the
the concepts
concepts and
and relationships
relationships used
used to
to describe
describe
and
and represent
represent an
an area
area of
of knowledge”
knowledge”

 RDFS can be considered as a simple ontology

language
 Languages should be a compromise between
 rich semantics for meaningful applications
 feasibility, implementability
 145

Web Ontology Language = OWL

 OWL is an extra layer, a bit like RDF Schemas
 own namespace, own terms
 it relies on RDF Schemas
 It is a separate recommendation
 actually… there is a 2004 version of OWL (“OWL 1”)
 and there is an update (“OWL 2”) that should be finalized in
2009
 you will surely hear about it at the conference…
 146

OWL is complex…
 OWL is a large set of additional terms
 We will not cover the whole thing here…
 147

Term equivalences
 For classes:
 owl:equivalentClass: two classes have the same
individuals
 owl:disjointWith: no individuals in common
 For properties:
 owl:equivalentProperty
 remember the a:author vs. f:auteur
 owl:propertyDisjointWith
 For individuals:
 owl:sameAs: two URIs refer to the same concept
(“individual”)
 owl:differentFrom: negation of owl:sameAs
 148

Connecting to French…
 149

Typical usage of owl:sameAs

 Linking our example of Amsterdam from one data
set (DBpedia) to the other (Geonames):

<http://dbpedia.org/resource/Amsterdam>
<http://dbpedia.org/resource/Amsterdam>
owl:sameAs
owl:sameAs <http://sws.geonames.org/2759793>;
<http://sws.geonames.org/2759793>;

 This is the main mechanism of “Linking” in the

Linking Open Data project
 150

Property characterization
 In OWL, one can characterize the behaviour of
properties (symmetric, transitive, functional, inverse
functional…)
 One property may be the inverse of another
 OWL also separates data and object properties
 “datatype property” means that its range are typed literals
 151

What this means is…

 If the following holds in our triples:
:email
:email rdf:type
rdf:type owl:InverseFunctionalProperty.
owl:InverseFunctionalProperty.
<A> :email "mailto:[email protected]""..
<A> :email " mailto:[email protected]
<B> :email ""mailto:[email protected]
<B> :email mailto:[email protected]""..

then, processed through OWL, the following

holds, too:
<A>
<A> owl:sameAs
owl:sameAs <B>.
<B>.

 I.e., new relationships were discovered again

(beyond what RDFS could do)
 152

Classes in OWL
 In RDFS, you can subclass existing classes…
that’s all
 In OWL, you can construct classes from existing
ones:
 enumerate its content
 through intersection, union, complement
 Etc
 153

Classes in OWL (cont)

 OWL makes a stronger conceptual distinction
between classes and individuals
 there is a separate term for owl:Class, to make the
difference (a specialization of the RDFS class)
 individuals are separated into a special class called
owl:Thing
 Eg, a precise classification would be:

ex:Person
ex:Person rdf:type
rdf:type owl:Class.
owl:Class.
<uri-for-Amitav-Ghosh>
<uri-for-Amitav-Ghosh>
rdf:type
rdf:type owl:Thing;
owl:Thing;
rdf:type
rdf:type owl:Person ..
owl:Person
 154

Classes contents can be enumerated

:£
:£ rdf:type
rdf:type owl:Thing.
owl:Thing.
:€ rdf:type owl:Thing.
:€ rdf:type owl:Thing.
:$
:$ rdf:type
rdf:type owl:Thing.
owl:Thing.
:Currency
:Currency
rdf:type
rdf:type owl:Class;
owl:Class;
owl:oneOf
owl:oneOf (:€
(:€ :£
:£ :$).
:$).

 I.e., the class consists of exactly of those

individuals
 155

Union of classes can be defined

:Novel
:Novel rdf:type
rdf:type owl:Class.
owl:Class.
:Short_Story
:Short_Story rdf:type owl:Class.
rdf:type owl:Class.
:Poetry
:Poetry rdf:type
rdf:type owl:Class.
owl:Class.
:Literature rdf:type owl:Class;
:Literature rdf:type owl:Class;
owl:unionOf
owl:unionOf (:Novel
(:Novel :Short_Story
:Short_Story :Poetry).
:Poetry).

 Other possibilities: complementOf,

intersectionOf, …
 156

For example…
If:
:Novel
:Novel rdf:type
rdf:type owl:Class.
owl:Class.
:Short_Story
:Short_Story rdf:type owl:Class.
rdf:type owl:Class.
:Poetry
:Poetry rdf:type
rdf:type owl:Class.
owl:Class.
:Literature rdf:type owl:Class;
:Literature rdf:type owl:Class;
owl:unionOf
owl:unionOf (:Novel
(:Novel :Short_Story
:Short_Story :Poetry).
:Poetry).
<myWork>
<myWork> rdf:type
rdf:type :Novel
:Novel ..

then the following holds, too:

<myWork>
<myWork> rdf:type
rdf:type :Literature
:Literature ..
 157

It can be a bit more complicated…

If:
:Novel
:Novel rdf:type
rdf:type owl:Class.
owl:Class.
:Short_Story
:Short_Story rdf:type owl:Class.
rdf:type owl:Class.
:Poetry
:Poetry rdf:type
rdf:type owl:Class.
owl:Class.
:Literature rdf:type owlClass;
:Literature rdf:type owlClass;
owl:unionOf
owl:unionOf (:Novel
(:Novel :Short_Story
:Short_Story :Poetry).
:Poetry).
fr:Roman
fr:Roman owl:equivalentClass
owl:equivalentClass :Novel
:Novel ..
<myWork>
<myWork> rdf:type
rdf:type fr:Roman
fr:Roman ..

then, through the combination of different terms,

the following still holds:
<myWork>
<myWork> rdf:type
rdf:type :Literature
:Literature ..
 158

What we have so far…

 The OWL features listed so far are already fairly
powerful
 E.g., various databases can be linked via
owl:sameAs, functional or inverse functional
properties, etc.
 Many inferred relationship can be found using a
traditional rule engine
 159

However… that may not be enough

 Very large vocabularies might require even more
complex features
 typical usage example: definition of all concepts in a health
care environment
 a major issue: the way classes (i.e., “concepts”) are defined
 OWL includes those extra features but… the
inference engines become (much) more complex
 160

Property value restrictions

 Classes are created by restricting the property
values on its individuals
 For example: how would I characterize a “listed
price”?
 it is a price (which may be a general term), but one that is
given in one of the “allowed” currencies (say, €, £, or $)
 more formally:
 the value of “p:currency”, when applied to a resource on
listed price, must be of one of those values…
 …thereby defining the class of “listed price”
 161

Restrictions formally
 Defines a class of type owl:Restriction with a
 reference to the property that is constrained
 definition of the constraint itself
 One can, e.g., subclass from this node when
defining a particular class

:Listed_Price
:Listed_Price rdfs:subClassOf
rdfs:subClassOf [[
rdf:type
rdf:type owl:Restriction;
owl:Restriction;
owl:onProperty
owl:onProperty p:currency;
p:currency;
owl:allValuesFrom
owl:allValuesFrom :Currency.
:Currency.
].
].
 162

Possible usage…
If:
:Listed_Price
:Listed_Price rdfs:subClassOf
rdfs:subClassOf [[
rdf:type
rdf:type owl:Restriction;
owl:Restriction;
owl:onProperty
owl:onProperty p:currency;
p:currency;
owl:allValuesFrom
owl:allValuesFrom :Currency.
:Currency.
].
].
:price
:price rdf:type
rdf:type :Listed_Price
:Listed_Price ..
:price
:price p:currency
p:currency <something>
<something> ..

then the following holds:

<something>
<something> rdf:type
rdf:type :Currency
:Currency ..
 163

Other restrictions

 allValuesFrom could be replaced by:

 someValuesFrom
 e.g., I could have said: there should be a price given in at
least one of those currencies
 hasValue, when restricted to one specific value
 Cardinality restrictions: instead of looking at the
values of properties, their number is considered
 eg, a specific property should occur exactly once
 164

But: OWL is hard!

 The combination of class constructions with various
restrictions is extremely powerful
 What we have so far follows the same logic as
before
 extend the basic RDF and RDFS possibilities with new
features
 define their semantics, ie, what they “mean” in terms of
relationships
 expect to infer new relationships based on those
 However… a full inference procedure is hard
 not implementable with simple rule engines, for example
 165

OWL “species”
 OWL species comes to the fore:
 restricting which terms can be used and under what
circumstances (restrictions)
 if one abides to those restrictions, then simpler inference
engines can be used
 They reflect compromises: expressibility vs.
implementability
 166

OWL Full
 No constraints on any of the constructs
 owl:Class is just syntactic sugar for rdfs:Class
 owl:Thing is equivalent to rdfs:Resource
 this means that:
 Class can also be an individual, a URI can denote a property
as well as a Class
 e.g., it is possible to talk about class of classes, apply properties
on them
 etc
 etc.
 Extension of RDFS in all respects
 But: no system may exist that infers everything one
might expect
 167

OWL Full usage

 Nevertheless OWL Full is essential
 it gives a generic framework to express many things
 some application just need to express and interchange
terms (with possible scruffiness)
 Applications may control what terms are used and
how
 in fact, they may define their own sub-language via, eg, a
vocabulary
 thereby ensuring a manageable inference procedure
 168

OWL DL
 A number of restrictions are defined
 classes, individuals, object and datatype properties, etc, are
fairly strictly separated
 object properties must be used with individuals
 i.e., properties are really used to create relationships between
individuals
 no characterization of datatype properties
 …
 But: well known inference algorithms exist!
 169

Examples for restrictions

 The following is not “legal” OWL DL:

<q>
<q> rdf:type
rdf:type <A>.
<A>. ## AA is
is aa class,
class, qq is
is an
an individual
individual
<r>
<r> rdf:type
rdf:type <q>.
<q>. ## error:
error: qq cannot
cannot be
be used
used for
for aa class,
class, too
too
<A>
<A> ex:something
ex:something <B>.
<B>. ## error:
error: properties
properties are
are for
for individuals
individuals only
only
<q>
<q> ex:something
ex:something <s>.
<s>. ## error:
error: same
same property
property cannot
cannot be
be used
used as
as
<p> ex:something “54”.
<p> ex:something “54”. ## object and datatype property
object and datatype property
 170

OWL DL usage
 Abiding to the restrictions means that very large
ontologies can be developed that require precise
procedures
 eg, in the medical domain, biological research, energy
industry, financial services (eg, XBRL), etc
 the number of classes and properties described this way
can go up to the many thousands
 OWL DL has become a language of choice to
define and manage formal ontologies in general
 even if their usage is not necessarily on the Web
 171

OWL 2 defines further species

a.k.a. “profiles”
 Further restrictions on how terms can be used and
what inferences can be expected
 Classification and instance queries in polynomial time:
OWL-EL
 Implementable on top of conventional relational database
engines: OWL-QL
 Implementable on top of traditional rule engines: OWL-RL
 172

Ontology development
 The hard work is to create the ontologies
 requires a good knowledge of the area to be described
 some communities have good expertise already (e.g.,
librarians)
 OWL is just a tool to formalize ontologies
 large scale ontologies are often developed in a community
process
 Ontologies should be shared and reused
 can be via the simple namespace mechanisms…
 …or via explicit import
 173

Must I use large ontologies?

 NO!!!
 Many applications are possible with RDFS and a
just a little bit of OWL
 a few terms, whose meaning is defined in OWL, and that
application can handle directly
 OWL RL is a step to create such a generic OWL level
 Big ontologies can be expensive (both in time and
money); use them only when really necessary!
 174

Ontologies examples
 eClassOwl: eBusiness ontology for products and
services, 75,000 classes and 5,500 properties
 National Cancer Institute’s ontology: about 58,000
classes
 Open Biomedical Ontologies Foundry: a collection
of ontologies, including the Gene Ontology to
describe gene and gene product attributes in any
organism or protein sequence and annotation
terminology and data (UniProt)
 BioPAX: for biological pathway data
 175

Example: improved search via ontology

 Search results are re-ranked using ontologies
 Related terms are highlighted, usable for further
search
 176

Example: improved search via ontology

 Same dataset, different ontology
 (ontology is on non-animal experimentation)
 177

Example: Eli Lilly’s target assessment tool

 Prioritization of
drug target,
integrating data
from different
sources and
formats
 Integration,
search via
ontologies
(proprietary and
public)

Courtesy of Susie Stephens, Eli Lilly (SWEO Case Study)

 178

What have we achieved?

(putting all this together)
 179

Other SW technologies
 There are other technologies that we do not have
time for here
 find RDF data associated with general URI-s: POWDER
 bridge to thesauri, glossaries, etc: SKOS
 use Rule engines on RDF data
 180

Remember the integration example?

 181

Same with what we learned

 182

Example: personalized tourist itinerary

 Integration of
relevant data in
Zaragoza (using
RDF and ontologies)
 Use rules on the
RDF data to provide
a proper itinerary

Courtesy of Jesús Fernández, Mun. of Zaragoza, and Antonio Campos, CTIC (SWEO Use Case)
 183

Available documents, resources

 184

Available specifications: Primers, Guides

 The “RDF Primer” and the “OWL Guide” give a
formal introduction to RDF(S) and OWL
 GRDDL and RDFa Primers have also been
published
 The W3C Semantic Web Activity Homepage has
links to all the specifications:
 http://www.w3.org/2001/sw/
 185

“Core” vocabularies
 There are also a number widely used “core
vocabularies”
 Dublin Core: about information resources, digital libraries,
with extensions for rights, permissions, digital right
management
 FOAF: about people and their organizations
 DOAP: on the descriptions of software projects
 SIOC: Semantically-Interlinked Online Communities
 vCard in RDF
 …
 One should never forget: ontologies/vocabularies
must be shared and reused!
 186

Some books
 G. Antoniu and F. van Harmelen: Semantic Web
Primer, 2nd edition in 2008
 D. Allemang and J. Hendler: Semantic Web for the
Working Ontologist, 2008
 Jeffrey Pollock: Semantic Web for Dummies, 2009
 …

See the separate Wiki page collecting book references:

http://esw.w3.org/topic/SwBooks
 187

Further information
 Planet RDF aggregates a number of SW blogs:
 http://planetrdf.com/
 Semantic Web Interest Group
 a forum developers with archived (and public) mailing list,
and a constant IRC presence on freenode.net#swig
 anybody can sign up on the list:
 http://www.w3.org/2001/sw/interest/
 188

Lots of Tools (not an exhaustive list!)

 Categories:  Some names:
 Triple Stores  Jena, AllegroGraph, Mulgara,
Sesame, flickurl, …
 Inference engines  TopBraid Suite, Virtuoso
 Converters environment, Falcon, Drupal 7,
Redland, Pellet, …
 Search engines  Disco, Oracle 11g, RacerPro,
 Middleware IODT, Ontobroker, OWLIM, Tallis
Platform, …
 CMS  RDF Gateway, RDFLib, Open
 Semantic Web browsers Anzo, DartGrid, Zitgist, Ontotext,
Protégé, …
 Development environments  Thetus publisher, SemanticWorks,
 Semantic Wikis SWI-Prolog, RDFStore…
 …  …
 189

Conclusions
 The Semantic Web is about creating a Web of
Data
 There is a great and very active user and
developer community, with new applications
 witness the size and diversity of this event
 190

By the way: the book is real

 191

Thank you for your attention!

These slides are also available on the Web:

http://www.w3.org/2009/Talks/0615-SanJose-tutorial-IH/