1. INF385T: Information Modeling — Class Semantic Web Vocabulary Systems RDF, RDFS, OWL Slides for November 30, Karen Wickett School of Information University of Texas at Austin

2. RDF
The Resource Description Framework (RDF) is a language for representing information about resources in the World Wide Web. [W3C - RDF Primer]
The RDF specification is published by the W3C and is considered a fundamental part of the Semantic Web.
Some projects:
dbpedia
bibframe
Europeana / DPLA

3. Resource Description Framework
RDF gives a general model for describing things - by saying that some thing has a particular property with a particular value.
These descriptions are intended to be processed by software and shared across applications and computing environments without loss of meaning.
RDF by itself is typically expressed with a graphical convention.
There are several serialization syntaxes for RDF:
RDF/XML, Trig, Turtle, N3, JSON-LD

4. Ontologies and the Semantic Web
RDF lets us make statements about individual resources and their properties.
RDFS lets us describe the vocabulary we’re using
in terms of domain/range and class/subclass relationships
But we may need to say more than we can we with RDFS.
For example:
That two URIs refer to the same individual
That two URIs refer to the same property
That a property is symmetric or transitive
Cardinality and participation constraints.
These are beyond the expressiveness of RDFS. We need an ontology language.

5. But first... what is ontology?
In philosophy and metaphysics:
The science or study of being/reality/existence [in general]
- dictionary definitions of the most general philosophical sense of ontology
The favorite (informatic) definition:
“An ontology is an explicit specification of a conceptualization of a domain”.
Tom Gruber (1993).

6. Ontologies in contemporary informatics
Comes from a CS/AI background, but is evolving, based on a more sophisticated understanding of needs and opportunities in informatics
Objectives:
designing information systems (same as conceptual modeling)
interoperating information systems
integrating data
Inferencing and analysis

7. Ontologies and the Semantic Web
The W3C “Layer Cake” diagram:

8. The Web Ontology Language (OWL)
From the W3C (
OWL 2 is a Semantic Web language designed to represent rich and complex knowledge about things, groups of things, and relations between things.
OWL documents, known as ontologies, can be published in the World Wide Web and may refer to or be referred from other OWL ontologies.
OWL 2 ontologies provide classes, properties, individuals, and data values and are stored as Semantic Web documents.
OWL 2 ontologies can be used along with information written in RDF
OWL 2 ontologies themselves are primarily exchanged as RDF documents.
So, OWL is a language for creating ontologies.
(not itself an ontology)

9. RDF Data Model
The basic structure of a single descriptive assertion in RDF is a triple, an ordered set of these three elements
subject (thing)
predicate (property)
object (value)
In RDF each these elements should be referred to with a URI.
However: “Objects in RDF statements may be either URIrefs, or constant values (called literals) represented by character strings, in order to represent certain kinds of property values.” – RDF Primer

10. Uniform Resource Identifiers (URIs)
This is a convention for generating “compact strings” that unambiguously refer to a resource.
Resources here can be anything.
Webpages, people, concepts, books, planets, organizations,
species (e.g. Equus ferus), individual animals (e.g. Secretariat), etc. Anything.
This is a world of things with properties.
We name the things (subject) and the general properties (predicate) with URIs.
We name the property values (object) with URIs or typed literals.

11. RDF Graph Languague In the RDF graph language: predicate subject
object
An RDF graph is a collection of triples.

12. example -- to express in RDF
A simple bibliographic description expressed using simple Dublin Core*:
identifier = b128935
title = The Intellectual Foundation of Information Organization
creator = Elaine Svenonius
date = 2000
publisher = MIT Press
language = english
type = text *

13. expressing DC in RDF The description as a (simplifed) RDF graph.
dc:creator
“b128935”
Elaine Svenonius
dc:publisher
MIT Press
dc:identifier
dc:date
2000
dc:title
“The Intellectual...”
dc:language
English
dc:type
text

14. expressing DC in RDF RDF graph with URIs.
“b128935”
“The Intellectual...”

15. Querying RDF
SPARQL is a query language for RDF where you specify a “graph pattern” that individual triples are matched against.
To return all resources (and their titles) where dc:date is “2000”, you would issue the following query
PREFIX dc:
SELECT ?x ?title
WHERE{
?x dc:date “2000”.
?x dc:title ?title }

16. RDF description as triples
< < <”b128935”>.
< < <”The Intellectual Foundation of Information Organization”>.
< < <
< < <
< < <
< < <
< < <

17. in a triple store Subject Predicate Object http://example.com/b128935
"b128935"
"The Intellectual Foundation of Information Organization"

18. Example in RDF/XML Syntax
<rdf:RDF xmlns:rdf=" xmlns:dc=" <rdf:Description rdf:about=" <dc:title>The Intellectual Foundation of Information Organization</dc:title> <dc:creator rdf:resource=“ <dc:date rdf:resource=“ <dc:publisher rdf:resource=“ Category:Massachusetts_Institute_of_Technology”/> <dc:language rdf:resource=“ <dc:type rdf:resource=“ </rdf:Description> </rdf:RDF>

19. RDF in FOL How would we translate the example RDF graph into FOL?
What features of FOL are needed to express the assertions in the graph?
Recall that the graph is a collection of triples of the form <subject, predicate, object>.

20. Syntax and Semantics
Specify the syntax and semantics of the fragment of FOL necessary to express the RDF graph.
What do we need in FOL to say everything in the last few slides?
Note from the RDF Concepts and Abstract Syntax:
The assertion of an RDF graph amounts to asserting all the triples in it, so the meaning of an RDF graph is the conjunction (logical AND) of the statements corresponding to all the triples it contains.

21. A fragment of note
EC - the Existential Conjunctive fragment of first order logic.
This fragment of FOL contains predicates, a single (existential) quantifier, and conjunction.
Intuitively, the assertions in an RDF graph are making existential statements.
This fragment is well-known and computationally tractable.

22. Inference system for EC
Given the fragment of FOL that corresponds to RDF, what kinds of inferences can be made?
How would an inference be carried out?
Derivation Rule:
You can derive a sentence as a conclusion when that sentence appears as a conjunct in the premises.

23. Metatheory of the EC fragment
This is a fragment of FOL. So the results on soundness and completeness hold here.
However, the expressiveness here is less than that of FOL.
Connectives:
There is no negation. Therefore no way no deny any particular assertion.
There is no universal quantification
There is no way to express the conditional.
This simplifies inferencing, but also reduces it - No Modus Ponens.
Predication:
There are only dyadic predicates.
Therefore no (direct) way to express 3-place relations.

24. “Readers who are familiar with conventional logical semantics may find it useful to think of RDF as a version of existential binary relational logic in which relations are first-class entities in the universe of quantification.
Such a logic can be obtained by encoding the relational atom R(a,b) into a conventional logical syntax, using a notional three-place relation Triple(a,R,b).”
-W3C Semantics

25. RDF and RDFS
RDFS (RDF Schema) is a set of RDF properties that let us describe vocabularies

26. Core elements of RDFS rdf:type rdfs:class rdfs:subclass rdfs:domain
connects a resource with a class
rdfs:class
states that something is a class
rdfs:subclass
states that one class is a subclass of another
rdfs:domain
rdfs:range

27. expressing DC in RDF The description as a (simplified) RDF graph.
dc:creator
“b128935”
Elaine Svenonius
dc:publisher
MIT Press
dc:identifier
dc:date
2000
dc:title
“The Intellectual...”
dc:language
English
dc:type
text

28. Creating a FRBR-compliant description
The RDF graph on the previous slide is not consistent with FRBR.
Why not?

29. FRBR group 1 entity-relationship model
FRBR entities include works, expressions, manifestations, and items.
Chart from Tillet, 2004

30. FRBR Group 1 entity types
A work in the FRBR model is “a distinct intellectual or artistic creation.”
An expression is analogous to a text: “the intellectual or artistic realization of a work” in a form, be it textual, sound, image, musical notation, whatever.
The expression encompasses the intellectual but not the physical form (e.g., typeface and layout are not part of the expression).
A manifestation in FRBR is the realization of an expression in a physical medium.
The same expression can be embodied in different manifestations.
All copies that are produced as part of the same set are the same manifestation.
An item in FRBR refers to the actual physical copy of a manifestation.

31. Creating a FRBR-based description
In order to fit our description to the FRBR conceptual model, we need to express facts about the attributes we’re using and the types of things we’re dealing with.
We need a notation that interoperates with RDF.
RDF Schema provides a language for describing RDF vocabularies.
What follows is based on the RDF vocabulary for FRBR published here:

32. FRBR in RDFS -- Classes Work Endeavor Item Expression Manifestation
subClassOf
subClassOf
Work
Item
subClassOf
subClassOf
Expression
Manifestation
ResponsibleEntity
subClassOf
subClassOf
Person
if A is a subclass of B then
all instances of A are also instances of B
Corporate Body

33. FRBR in RDFS -- Properties
Endeavor
subClassOf
subClassOf
Work
Item
subClassOf
subClassOf
domain
range
isExemplifiedBy
isRealizedBy
domain
Expression
Manifestation
range
domain
isEmbodiedBy
range
if the domain of the property P is D
then for every P(x,y), x  D
if the domain of the property P is R
then for every P(x,y), y  R

34. FRBR in RDFS -- More Properties
Endeavor
range
ResponsibleEntityOf
subClassOf
subClassOf
Work
subClassOf
Item
subClassOf
domain
range
range
Manifestation
domain
isRealizedBy
range
isExemplifiedBy
domain
Expression
CreatorOf
range
domain
domain
isEmbodiedBy
range
ResponsibleEntity
domain
ProducerOf
subClassOf
subClassOf
Person
Corporate Body

35. FRBR in RDFS -- Sub-Properties
RelatedEndeavor
Endeavor
range
ResponsibleEntityOf
subClassOf
subClassOf
Work
subClassOf
Item
subClassOf
domain
range
range
Manifestation
domain
RealizedBy
range
ExemplifiedBy
domain
Expression
CreatorOf
range
domain
domain
EmbodiedBy
range
ResponsibleEntity
domain
ProducerOf
subClassOf
subClassOf
Person
Corporate Body
if P is a subproperty of Q
then whenever P(x,y) holds, Q(x,y) holds.

36. FRBR in RDFS with RDF instance
ResponsibleEntityOf
RelatedEndeavor
range
domain
ResponsibleEntity
Endeavor
domain
subClassOf
domain
ProducerOf
subClassOf
CreatorOf
subClassOf
subClassOf
range
range
Item
subClassOf
Work
Corporate Body
Person
subClassOf
range
Manifestation
domain
ExemplifiedBy
domain
Expression
RealizedBy
range
range
domain
EmbodiedBy
EmbodiedBy
w128935
RealizedBy
ExemplifiedBy
m128935
e128935
i128935

37. FRBR in RDFS with RDF instance
ResponsibleEntityOf
RelatedEndeavor
range
domain
ResponsibleEntity
Endeavor
domain
subClassOf
domain
ProducerOf
subClassOf
CreatorOf
subClassOf
range
Item
range
subClassOf
Work
Corporate Body
Person
subClassOf
range
subClassOf
Manifestation
domain
ExemplifiedBy
domain
Expression
RealizedBy
range
range
domain
EmbodiedBy
type
type
type
type
w128935
RealizedBy
EmbodiedBy
ExemplifiedBy
e128935
i128935
m128935

38. FRBR in RDFS with RDF instance
ResponsibleEntityOf
RelatedEndeavor
range
domain
ResponsibleEntity
Endeavor
domain
subClassOf
domain
ProducerOf
subClassOf
CreatorOf
subClassOf
range
Item
range
subClassOf
Work
subClassOf
Corporate Body
range
subClassOf
Manifestation
Person
domain
ExemplifiedBy
domain
Expression
RealizedBy
range
range
type
type
domain
EmbodiedBy
type
type
type
type
Elaine Svenonius
MIT Press
CreatorOf
ProducerOf
type
EmbodiedBy
w128935
RealizedBy
ExemplifiedBy
e128935
m128935
i128935
type
Karen Wickett
Owns

39. A-Box and T-Box
In artificial intelligence and knowledge representation systems there is a notion of separation between the A-Box and T-Box.
The A-Box is assertional and contains information about individuals, their attributes and relationships.
also called a world description.
The T-Box is terminological and contains information about the vocabulary -- domains and ranges of properties, constraints on relationships, class/subclass structure

40. T-Box and A-Box T-BOX A-BOX Work ResponsibleEntity Endeavor Item
ResponsibleEntityOf
RelatedEndeavor
range
domain
ResponsibleEntity
Endeavor
T-BOX
domain
subClassOf
domain
ProducerOf
subClassOf
CreatorOf
subClassOf
range
Item
range
subClassOf
Work
subClassOf
Corporate Body
range
subClassOf
Manifestation
Person
domain
ExemplifiedBy
domain
Expression
RealizedBy
range
range
type
type
domain
EmbodiedBy
type
type
type
type
Elaine Svenonius
MIT Press
CreatorOf
ProducerOf
A-BOX
type
EmbodiedBy
w128935
RealizedBy
ExemplifiedBy
e128935
m128935
i128935
type
Karen Wickett
Owns