The Semantic Web

1
The Semantic Web

• A Brief Introduction
• Dr. Randy M. Kaplan

2
A Quote

• I have always imagined the information space as
  something to which everyone has immediate and
  intuitive access, not just to browse but to
  create ... Machines become capable of analyzing
  all the data on the web - the content, and links,
  and transaction between people and computers.

2
3
A Quote (continued)

• ... When the Semantic web does emerge the
  day-to-day mechanisms of trade, bureaucracy, and
  our daily lives will be handled by machines
  talking to machines, leaving people to provide
  the inspiration and intuition.
• Tim Berners-Lee, 2000

3
4
What is the Semantic Web

• Semantic meaning
• Semantic Web meaningful web?
• The term semantic web is a vision
• Computers - Software - People
• Find, read, and understand data over the WWW to
  accomplish useful goals for users

4
5
Semantic Web

• Make data anywhere on the web accessible and
  understandable to people and machines
• Informally defined
• Fluid
• Evolving
• Not a working system

5
6
Semantic web

• Ideas
• Machine-readable-data-view
• Data on the web is used by machine
• Used for automation, integration, and reuse (W3C
  2003)

6
7
Semantic web

• Ideas
• Intelligent Agents
• Allow intelligent agents to retrieve and
  manipulate pertinent information

7
8
Semantic web

• Ideas
• Distributed Database View
• The Semantic Web does for data what HTML did for
  textual information systems

8
9
Semantic web

• Ideas
• Distributed Database View
• Represent all databases
• Represent logic rules to link databases together
  to add value

9
10
Semantic web

• Ideas
• Distributed Database View
• Transform the web from a large interlinked book
  to large interlinked database

10
11
Semantic web

• Ideas
• Automated Infrastructure
• Berners-Lee argues that the Semantic web is
  infrastructure and not software

11
12
Semantic web

• Ideas
• Servant of Humanity
• Allow software relieve us of much of the burden
  of finding, extracting, integrating, and indexing
  information on the web

12
13
Semantic web

• Ideas
• Better Annotation
• Web as we know it augmented with annotations that
  are machine-readable and linked

13
14
Semantic web

• Ideas
• Improved Searching
• Possible to access web resources by content
  rather than keywords

14
15
Semantic web

• Ideas
• Web Services
• Semantic Web will be called upon to provide
  services that in turn provide useful behavior
• Agents automate procedures that are currently
  performed manually

15
16
Semantic web

• Recurring Themes
• Indexing and retrieving
• Meta data
• Annotation
• Web as Database
• Machine retrieval of data

16
17
Semantic web

• Recurring Themes
• Web-based services
• Discovery of services
• Agents

17
18
Semantic Web

• Videos
• http//www.youtube.com/watch?vmVFY52CH6Bc
• http//www.youtube.com/watch?vOGg8A2zfWKg

18
19
Semantic web

• Scenario
• During her stay in Honolulu Clara runs into
  several interesting people and they trade vCards.
  In the evening Clara reviews the days activities
  on her digital assistant. The assistant pops up a
  note with a link to one of the vCards that reads,
  This guys profile seems to match the position
  advertisement that Bill put on our intranet. Can
  I notify Bills assistant?
• Clara asked her digital assistant to explain. I
  used his company directory for finding his DAML
  enhanced vita hes got the required skills as a
  statistician who led the data mining group of the
  database department at Montana U for the
  requirement of a researcher who worked in machine
  learning. Clara asked the assistant to provide
  evidence. I checked his affiliation with the
  university of Montana, he is cited several times
  in their web pages reasonably trusted I checked
  his publication records from publishers DAML
  sources and asked bill assistant a rating of the
  journals highly trusted. More details?
• Clara asked the assistant to inform Bills
  assistant

19
20
Semantic web

• Questions about the scenario -
• What sorts of information must the agent have
  access to in order to perform these tasks?
• What sorts of processing must the agent do in
  order to perform these tasks?

20
21
Semantic web

• Foundations
• Resources
• Standardized Addressing
• Small set of commands
• Scalability and large networks

21
22
Semantic web

• Foundations
• Openness, completeness, and consistency
• Web and Semantic Web

22
23
Semantic web

• Resources
• A resource is intended to represent any idea that
  can be referred to
• We think of these as tangible packages of data
• The notion of resource is more general

23
24
Semantic web

• Resources
• Generality of Resources
• A resource can change over time and still be the
  same resource
• A resource can also be a reference to a real or
  fictitious person

24
25
Semantic web
Trust
Logic Proof
Ontology
RDF Schema
RDF
XML Schema
XML
25
26
Semantic Web
Trust
Logic Proof
Ontology
RDF Schema
RDF
XML used to define new languages that are used to
interchange data on the web
XML Schema
XML
26
27
Semantic Web
Trust
Logic Proof
Ontology
RDF Schema
A schema defines a language and is used to make
sure statements in the language are
written correctly
RDF
XML Schema
XML
27
28
Semantic Web
RDF resource description framework
Trust
Logic Proof
Ontology
RDF Schema
RDF
A language that is capable of describing all
sorts of information and meta-data
XML Schema
XML
28
29
Semantic Web
Trust
When designing specific RDF-based application
languages defines specific elements of the
special language
Logic Proof
Ontology
RDF Schema
RDF
XML Schema
XML
29
30
Semantic Web
An ontology defines the vocabulary and usage of
words in the context of the specific vocabulary
Trust
Logic Proof
Ontology
RDF Schema
RDF
XML Schema
XML
30
31
Semantic Web
Logical reasoning is used to establish the
consistency and correctness of datasets
Trust
Logic Proof
Ontology
RDF Schema
RDF
XML Schema
Logical reasoning is also necessary to infer
conclusions which are not necessarily
XML
31
32
Provide authentication of identity
Trust
Logic Proof
Ontology
RDF Schema
RDF
Create evidence that data, service, and agents
are trustworthy
XML Schema
XML
32
33
RDF

• W3Cs candidate language to provide a standard
  way to model, describe, and exchange information
  about resources

33
34
RDF

• Three capabilities
• Describe most kinds of data that will be
  available
• Describe the structural design of data sets
• Describe relationships between bits of data

34
35
RDF

• Uses a simple data model
• Three things
• Resources
• Statements
• Simple sentences in SVO order

35
36
RDF

• Example
• Billy lives in Chicago.
• Billy subject
• Lives verb
• Chicago object

36
37
Statements in RDF

• Statement is sometimes called a triple (because
  it has three parts)
• Subject is called the subject
• Verb is called the predicate (alt. Property)
• Object is called the object (alt. Value)

37
38
Statements in RDF
Subject
Predicate
Object
(person-1, name, William Sparks)
Triple
38
39
Statements in RDF

• A value in a statement can be -
• A number
• A string
• Such a value is called a LITERAL
• A value in a statement may also be another
  resource

39
40
Identifying Resources in RDF

• RDF needs a standard widely used way to identify
  resources
• RDF uses the URI (Uniform Resource Identifier)
• URIs can be used to identify a concept, tangible
  thing, chunk of data

40
41
Identifying Resources in RDF

• Identifying things is complicated
• May never be a general solution
• Names are not unique (many Robert Smith in the
  U.S.)
• Same name - not the same person

41
42
Identifying Resources in RDF

• To identify a resource -
• A URI is paired with or associated with a
  resource and not any other resource
• How this is accomplished in NOT specifed
• May not be a good way to identify a resource

42
43
Identifying Resources in RDF

• Alternative -
• Identify a resource as the same as some other
  resource
• The other resource is well-known
• Use the right predicate (from a trusted source)

43
44
Identifying Resources in RDF

• Alternative
• Identify a resource by its properties and
  relationships with other resources

44
45
Identifying Resources in RDF

• Example using properties
• 2003 New York Yankees had one general manager
• If it were known that a person had been the
  general manager of the Yankees during that time,
  the individual would be uniquely identified

45
46
Identifying Resources in RDF

• Suppose the pairing of a URI with a resource is
  described in a document
• Question How can this document be found?
• RDF provides no standard method for answering
  this question

46
47
Identifying Resources in RDF

• To deal with the problem of locating the
  document, an informal approach is sometimes
  adopted
• Use a URL to retrieve the document
• One must remember that the document may not be
  the actual resource

47
48
Identifying Resources in RDF

• Example
• The type of resource that is an RDF statement is
  identified by its URI
• http//www.w3.org/1999/02/22-rdf-syntax-nsStateme
  nt

48
49
Identifying Resources in RDF

• At this URI there is a document
• The document contains the following RDF fragment
  among others

49
50
Identifying Resources in RDF
lt!-- This is the RDF schema for the RDF data
model as described in The Resource Description
Framework (RDF) Model and Syntax Specifications
http//www.w3c.org/TR/REC-rdf-syntax
--gt ltsClass rdfIDStatement ScommentA
triple consisting of a predicate, a subject, And
an object. /gt ltsClass rdfIDProperty Acomme
ntA name of a property, defining specific
meaning For the property /gt
50
51
Identifying Resources in RDF

• The URI is not required to point to a document
• The fact that this fragment in fact does is useful

51
52
Anonymous Resources

• A resource without an identifying URI reference
  is called an anonymous resource
• It would be analogous to English language
  expressions like,
• That man, a women with a red car
• Anonymous resources are also called blank-nodes

52
53
RDF and Conventional Databases

• Triples may seem different from the model found
  in the conventional relational databases
• The difference is not as great as it appearss

53
54
RDF and Conventional Databases

• Each row of a conventional relational database
  has the same structure as any other
• In relational database lingo, a collection of
  related data items is called a tuple

54
55
RDF and Conventional Databases

• Consider the following table fragment from a
  sample database

Name Phone Email City State
William Sparks 435 555-6789 Wsparks_at_mailto.com Springfield MA
Constance Sim 312 555-2238 Csim_at_coldmail.com Chicago IL
55
56
RDF and Conventional Databases

• In a well designed table, all the cells on a row
  will depend on a single value called the primary
  key

Name Phone Email City State
William Sparks 435 555-6789 Wsparks_at_mailto.com Springfield MA
Constance Sim 312 555-2238 Csim_at_coldmail.com Chicago IL
56
57
RDF and Conventional Databases

• An obvious candidate for this is the persons
  name ... But ...

Name Phone Email City State
William Sparks 435 555-6789 Wsparks_at_mailto.com Springfield MA
Constance Sim 312 555-2238 Csim_at_coldmail.com Chicago IL
57
58
RDF and Conventional Databases

• But ... Names can change and more than one person
  can have the same name so lets give each tuple a
  unique identifier

Name Phone Email City State
William Sparks 435 555-6789 Wsparks_at_mailto.com Springfield MA
Constance Sim 312 555-2238 Csim_at_coldmail.com Chicago IL
58
59
RDF and Conventional Databases

• But ... Names can change and more than one person
  can have the same name so lets give each tuple a
  unique identifier

Name Phone Email City State Person ID
William Sparks 435 555-6789 Wsparks_at_mailto.com Springfield MA person-1
Constance Sim 312 555-2238 Csim_at_coldmail.com Chicago IL person-2
59
60
RDF and Conventional Databases

• How do we get the data in this table in the form
  of a triple?

Name Phone Email City State Person ID
William Sparks 435 555-6789 Wsparks_at_mailto.com Springfield MA person-1
Constance Sim 312 555-2238 Csim_at_coldmail.com Chicago IL person-2
60
61
RDF and Conventional Databases

• How do we get the data in this table in the form
  of a triple? We can break this table into a bunch
  of smaller tables

Name Phone Email City State Person ID
William Sparks 435 555-6789 Wsparks_at_mailto.com Springfield MA person-1
Constance Sim 312 555-2238 Csim_at_coldmail.com Chicago IL person-2
61
62
RDF and Conventional Databases
Name Phone Email City State Person ID
William Sparks 435 555-6789 Wsparks_at_mailto.com Springfield MA person-1
Constance Sim 312 555-2238 Csim_at_coldmail.com Chicago IL person-2
Person ID Name
person-1 William Sparks
person-2 Constance Sim
Person ID Phone
person-1
person-2
435 555-6789
312 555-2238
62
63
RDF and Conventional Databases

• An RDF triple can be written as -
• (resource) - (type of fact) - (value of fact)
• Any set of tables can be rewritten in the form of
  RDF-triples
• The primary key should be converted to a URI
  reference

63
64
RDF and Conventional Databases

• RDF and relational tables seem to be equivalent
• Databases have better performance but require
  regular data
• RDF does not require regular data and can be
  added to by adding a new triple

64
65
RDF and Conventional Databases

• Another interesting feature of RDF is tha it can
  be self describing
• In other words, RDF can be used to describe the
  data that describes the data
• This particular aspect of RDF cannot be matched
  by a conventional database

65
66
Properties as Resources

• An RDF statement associates a resource with a
  property value
• The predicate (property type) denotes the type of
  association
• The object represents the value assigned

66
67
Properties as Resources

• Predicates in RDF are interesting
• Predicates are resources
• Predicates can be the subjects of statements

67
68
Properties as Resources

• An RDF predicate (or property) is also a kind of
  resource
• It can be identified by a URI reference
• A standard RDF property defined in the RDF model

68
69
Properties as Resources

• http//www.w3.org/1999/02/22-rdf-syntax-nstype

69
70
Properties as Resources

• A property identified by this URI reference is
  used to state the TYPE of the resource
• The recommendation includes an example where the
  subject is a person
• This subject will have a type property

70
71
RDF Graphs

• We can represent the RDF notation visually using
  graphs
• In graphs we use a node to represent a resource
  or literal value, an arrow represents a predicate

color
shoe
red
71
72
RDF Graphs

• In RDF we cannot speak about all shoes
• We can only speak about particular resources
• Statements can be made about the class of shoes
• This is not the same as making a statement about
  a generic shoe

72
73
RDF Graphs

• Lets give the shoe an identifier instead of a
  generic name

Shoe-23
color
red
73
74
RDF Graphs

• In RDF graphs circles or ovals represent
  resources and rectangles represent literal values

color
Shoe-23
red
74
75
RDF Graphs

• Resources with Many Statements
• We can have a graph that depicts several
  statements about the same resource

435 555-6789
phone
Person-1
name
William Sparks
email
Wsparks_at_mailto.com
city
state
MA
Springfield
75
76
RDF Graphs

• It goes without saying that city and state are
  somehow related and yet this information is not
  represented in the diagram
• We can correct this situation with the addition
  of a special node - the anonymous node

76
77
RDF Graphs
435 555-6789
phone
Person-1
name
William Sparks
email
Wsparks_at_mailto.com
city
state
MA
Springfield
77
78
RDF Graphs

• person-1 has an address that is something whose
  type is address, whose city is Springfield, and
  whose state is MA
• This something is known by its properties
• It is not known by its intrinsic identity

78
79
RDF Graphs
435 555-6789
phone
Person-1
name
William Sparks
email
Wsparks_at_mailto.com
type
city
state
address
MA
Springfield
79
80
RDF Graphs

• Describing Collections
• RDF provides three types of mechanisms to
  describe collections
• (1) Bag
• (2) Alternative
• (3) Sequence

80
81
RDF Graphs

• Bag
• Contains any number of resources without any
  order
• Duplicates are allowed

81
82
RDF Graphs

• Alternative
• Specifies that any one of the members can be
  selected
• For example -
• A light switch can be on or off

82
83
RDF Graphs

• Sequence
• Contains an ordered collection of resources

83
84
RDF Graphs

• The arcs from a collection node to its members
  are given special labels
• rdf1
• rdf2
• And so on ...

84
85
RDF Graphs

• Container Example

rdf_1
Alternative
state
Light Switch
rdf_2
85
86
rdfStatement

• A special resource
• Can be used to say things about triples
• In the RDF model, a triple is NOT a resource
• Therefore, it cant be the subject of another
  statement

86
87
RDF Graphs
RDFStatement
person-1
type
subject
name
predicate
object
William Sparks
87
88
RDF Graphs

• This graph is saying, There is something whose
  type is RDFStatement ...

RDFStatement
person-1
type
subject
name
predicate
object
William Sparks
88
89
RDF Graphs

• person-1 is the subject ...

RDFStatement
person-1
type
subject
name
predicate
object
William Sparks
89
90
RDF Graphs

• name as a predicate ...

RDFStatement
person-1
type
subject
name
predicate
object
William Sparks
90
91
RDF Graphs

• and William Sparks as an object

RDFStatement
person-1
type
subject
name
predicate
object
William Sparks
91
92
RDF Graphs

• Once we create this node we can use it as the
  subject of other sentencces

RDFStatement
person-1
type
subject
name
predicate
object
William Sparks
92
93
RDF Graphs

• We call this process reifying
• To reify is to make a thing out of something

93
94
Meshing Data and Meta Data

• Supposing you had a data model as shown below

reference
0..n
occurs in
0..n
1..n
publication
person
is written by
one or more
0..n
published
organization
94