The Structure of Ontologies

1
The Structure of Ontologies

• Martin Volk
• Stockholm University

2

• What Are You?
• I grew up in Rhode Island, a New England state
  which is largely Italian-American and
  French-Canadian, known chiefly for its small
  stature. When I was a kid in our neighborhood,
  the first thing you would ask on encountering a
  newcomer was "whats your name?" The second was
  "What are you" as an invitation to recite your
  ethnic composition in a kind of singsong voice.
• 90 of the kids would say "Italian with a little
  bit of French," or "half-Portuguese, one-quarter
  Italian and one-quarter Armenian." When I would
  chime in with "half Jewish, one quarter Scottish
  and one quarter English," the range of responses
  went from very puzzled looks to "does that mean
  youre not Catholic?" Wherein, I guess, began my
  fascination with classification, and especially
  with the problem of residual categories, or, the
  Other, or not elsewhere classified.
• --Leigh Star
• (quoted from http//weber.ucsd.edu/gbowker/classi
  fication/)

3
Overview of todays lecture

• What is an Ontology?
• How are ontologies structured? A closer look
• Why do Computational Linguists work with / on
  Ontologies?
• Some terminology with respect to the Semantic
  Web.
• Your next task!

4
Remarks about your ontologies

• Protégé is a complex system!
• not suited for synonyms nor for translations.
• Options
• Use of the documentation field
• For translations build a parallel hierarchy
• Trade-off between
• Stockholm-specific ontology
• General university ontology
• Test for classes
• Is the mother node a good answer to What is a X?

5
Remarks about SUIS

• SUIS will use your University ontology for query
  expansion.
• If your query is Who is Päivi Juvonen? then
  SUIS will search for Päivi Juvonen in
  combination with any person function (teacher,
  librarian, cook, rector, director of studies, )
• If your query is Where is the lecture on
  Semiotics? then SUIS will search for semiotics,
  some teaching event and a date.

6
Remarks about SUIS

• SUIS will use your ontology for answering
  questions directly.
• If your query is What is a lecturer? then SUIS
  will use the ontology for replying
• A lecturer is a type of teacher working at a
  university.
• A lecturer is called universitetslektor or
  högskoleadjunkt in Swedish.

7
What is an ontology?

• An ontology holds information about what
  categories exist in the domain, what properties
  they have, and how they are related to one
  another. (Chandrasekaran et al. 1999)

8
Types of ontologies

• Top-level ontology
• Lexical ontology (e.g. WordNet)
• WordNet is a lexical database for English (with
  synonyms and hyperonyms)
• General ontology (e.g. Cyc)
• Cyc is a formalized representation of fundamental
  human knowledge facts, rules of thumb, and
  heuristics for reasoning about the objects and
  events of everyday life
• Domain ontology (e.g. UMLS)
• The Unified Medical Language System
• Task ontology (e.g. CPV)
• The Common Procurement Vocabulary

9
Kinds of Ontologies

• Ontologies may vary not only in their content,
  but also in their structure and implementation.
• Level of description
• An ontology may mean different things
• from simple lexicons or controlled vocabularies,
  to
• categorically organized thesauri, to
• taxonomies where terms are related hierarchically
  and can be given distinguishing properties, to
• full-blown ontologies where these properties can
  define new concepts and where concepts have named
  relationships with other concepts, like "changes
  the effect of" or "buys from".

10
Kinds of Ontologies

• Conceptual scope
• Ontologies differ in respect to the scope and
  purpose of their content. The most prominent
  distinction is between
• the domain ontologies describing specific fields
  of endeavor, like medicine, and
• upper level ontologies describing the basic
  concepts and relationships invoked when
  information about any domain is expressed in
  natural language.
• The synergy among ontologies springs from the
  cross-referencing between upper level ontologies
  and various domain ontologies.

11
Kinds of Ontologies

• Specification language
• To build ontologies a number of possible
  languages can be used, including
• general logic programming languages like Prolog.
• More common are languages that have evolved
  specifically for ontology construction. The Open
  Knowledge Base Connectivity (OKBC) model and
  languages like KIF (and its successor CL --
  Common Logic) have become the bases of other
  ontology languages.
• There are also several languages based on a form
  of logic known as description logics. These
  include Loom and DAMLOIL, which is currently
  being evolved into the Web Ontology Language
  (OWL) standard.
• When comparing ontology languages, language
  expressiveness is usually given up for
  computability and simplicity.

12
Building Ontologies

• Acquire domain knowledge
• Assemble appropriate information resources and
  expertise that will define, with consensus and
  consistency, the terms used formally to describe
  things in the domain of interest.
• Organize the ontology
• Design the overall conceptual structure of the
  domain. This involves identifying the domain's
  principal concrete concepts and their properties,
  identifying the relationships among the concepts,
  creating abstract concepts as organizing
  features, referencing or including supporting
  ontologies, distinguishing which concepts have
  instances, and applying other guidelines of your
  chosen methodology.
• Flesh out the ontology
• Add concepts, relations, and individuals to the
  level of detail necessary to satisfy the purposes
  of the ontology.

13
Ontology (Noy  McGuinness)

• An ontology defines a common vocabulary for
  researchers who need to share information in a
  domain. It includes machine-interpretable
  definitions of basic concepts in the domain and
  relations among them.

14
Why ontologies? (Noy  McGuinness)

• Why would someone want to develop an ontology?
  Some of the reasons are
• To share common understanding of the structure of
  information among people or software agents
• To enable reuse of domain knowledge
• To make domain assumptions explicit
• To separate domain knowledge from the operational
  knowledge
• To analyze domain knowledge

15
(No Transcript)
16
WordNet relations
17
Other relations (Navigli Velardi. 2004.
CL-Journal, p.168)
18
Ontologies(acc. to John Sowa. 2000. Knowledge
Representation)

• (p.51) Logic has no vocabulary for describing
  the things that exist. Ontology fills that gap
  It is the study of existence, of all the kinds of
  entities abstract or concrete that make up
  the world.
• (p. 52) AI systems start with limited ontologies
  (microworlds), a small number of concepts that
  are tailored for a single application.
• Example Chat-80 geographical categories

19
Ontologies

• Sowa (p. 68) All perception begins with
  contrasts light-dark, up-down, hard-soft,
  loud-quiet, sweet-sour. Such contrasts are the
  source of distinctions for generating the
  categories of existence.

20
Examples of Ontologies

• The categories in the Yellow Pages (telephone
  book).
• The Yahoo categories.
• Library cataloging system.
• Good storing in a department store.
• The faculties at a university.
• The departments in a municipal or national
  administration or in a large enterprise.

21
Natural Language Processing (NLP) and Ontologies

• Building ontologies
• extracting terms from corpora as concept
  candidates.
• suggesting synonyms from context analysis.
• suggesting hyperonyms from compound analysis
  (e.g. cruciate ligament is a ligament).
• extending ontologies.
• merging ontologies.
• of two companies / two administrative units.

22
NLP and Ontologies

• Applying ontologies
• mapping concepts found in text to nodes in an
  ontology
• e.g. Riesling ? white wine
• disambiguating word senses
• e.g. SV lag ? EN law or team
• reasoning with the help of ontologies
• e.g. If Riesling is a white wine, then it is an
  alcoholic beverage.
• allowing for cross-language searches
• e.g. head of department ? prefekt

23
Automatic ontology construction

• Find is_a patterns in a corpus
• A computer printer is a computer peripheral
  device
• the HP OfficeJet 4215 is a multitalented,
  multifunction printer
• Use compounds
• laser printer ? printer

24
Automatic ontology construction

• Use appositive NPs
• Your Basic Laser Printer the Brother 1240
• lets you choose your active printer (the printer
  you are about to use)
• Canon launches A4 Photo Printer - The Bubble Jet
  i990
• Use coordination
• The PhotoSmart printer and scanner are 399 each
  
• Printer and Photocopier Troubleshooting
• BUT Hyena's printer and job management functions
  include

25
Merging Ontologies

• (Sowa) Different systems may use different names
  for the same kinds of entities
• even worse, they may use the same names for
  different kinds.

26
Merging Ontologies

• Before two ontologies can be merged it might be
  necessary to introduce new classes to accommodate
  the different structures.
• Often ontologies get aligned rather than merged.

27
The Semantic Web

• a vision by Tim Berners-Lee
• to facilitate the use of web pages by computers
• Ideally specification of content in a formal
  language
• Example
• father(carl_gustav, victoria).
• father(gustav_adolf, carl_gustav).
• grandfather(A,C) -
• father(A,B),
• father(B,C).

28
What is XML?

• XML eXtended Markup Language
• a language for structuring text and data
• the XML structure of a document can be checked
  against a grammar (DTD document type definition)

29
XML
URI Uniform Resource Identifier
provide names
RDF Resource Description Framework
Dublin Core a set of core elements
simple statements
OWL Web Ontology Language
complex logical statements
30
Uniform Resource Identifier (URI)

• If you want to discuss something, you must first
  identify it.
• Example Absolute URIs (from http//en.wikipedia.or
  g )
• http//somehost/absolute/URI/with/absolute/path/to
  /resource.txt
• ftp//somehost/resource.txt
• urna-rose-by-any-other-name (hmm... unregistered
  URN namespace)

31
Uniform Resource Identifier (URI)

• A URL (Uniform Resource Locator web address) is
  one type of a URI.
• Everybody can create a URI.
• A URI is not a set of directions telling your
  computer how to get to a specific file on the
  Web.
• A URI is a name for a "resource" (a thing). This
  resource may or may not be accessible over the
  Internet.

32
RDF (Resource Description Framework)

• RDF is a language for representing information
  about resources in the World Wide Web,
• e.g. representing metadata about Web resources,
  such as the title, author, and modification date
  of a Web page.
• An RDF statement is a lot like a simple sentence,
  except that almost all the words are URIs.
• Each RDF statement has three parts a subject, a
  predicate and an object.

33
RDF example

• Example a Wikipedia page about Tony Benn
• To say that the title of a page is "Tony Benn"
  and its publisher is "Wikipedia" in RDF
• lthttp//en.wikipedia.org/Tony_Benngt
  lthttp//purl.org/dc/elements/1.1/titlegt "Tony
  Benn" .
• lthttp//en.wikipedia.org/Tony_Benngt
  lthttp//purl.org/dc/elements/1.1/publishergt
  "Wikipedia" .
• and in RDF/XML
• ltrdfRDF
• xmlnsrdfhttp//www.w3.org/1999/02/22-rdf-synta
  x-ns
• xmlnsdc"http//purl.org/dc/elements/1.1/"gt
• ltrdfDescription rdfabout"http//en.wikipedia.o
  rg/Tony_Benn"gt
• ltdctitlegtTony Bennlt/dctitlegt
• ltdcpublishergtWikipedialt/dcpublishergt
• lt/rdfDescriptiongt
• lt/rdfRDFgt

34
URI vs Ontologies

• The use of all these URIs is useless if we never
  describe what they mean. This is where schemas
  and ontologies come in. A schema and an ontology
  are ways to describe the meaning and
  relationships of terms.

35
Richer schema capabilities

• RDF Schema provides basic capabilities. Other
  richer schema capabilities that have been
  identified as useful
• cardinality constraints on properties,
• e.g., that a Person has exactly one biological
  father.
• specifying that a given property (such as
  exhasAncestor) is transitive,
• e.g., that if A exhasAncestor B, and B
  exhasAncestor C, then A exhasAncestor C.
• specifying that a given property is a unique
  identifier (or key) for instances of a particular
  class.
• specifying constraints on the range or
  cardinality of a property that depend on the
  class of resource to which a property is applied,
  
• e.g., for a soccer team the exhasPlayers
  property has 11 values, while for a basketball
  team the same property should have only 5 values.
  
• the ability to describe new classes in terms of
  combinations (e.g., unions and intersections) of
  other classes, or to say that two classes are
  disjoint.

36
OWL (Web Ontology Language)

• is a powerful ontology language.
• is based on RDF and RDF Schema.
• The intent of OWL is to provide additional
  machine-processable semantics for resources.
• is a semantic markup language for publishing and
  sharing ontologies on the Web.
• can be used to explicitly represent the meaning
  of terms in vocabularies and the relationships
  between those terms. ? i.e. ontologies

37
OWL

• OWL Lite supports those users needing a
  classification hierarchy and simple constraints.
• For example, it only permits cardinality values
  of 0 or 1. ? simple to provide tool support for
  OWL Lite, and ? quick migration path for thesauri
  and other taxonomies.
• OWL DL (description logics) supports maximum
  expressiveness while retaining computational
  completeness (all conclusions are guaranteed to
  be computable) and decidability (all computations
  will finish in finite time). OWL DL includes all
  OWL language constructs, but with certain
  restrictions
• For example, while a class may be a subclass of
  many classes, a class cannot be an instance of
  another class.
• OWL Full gives maximum expressiveness and the
  syntactic freedom of RDF with no computational
  guarantees.
• For example, in OWL Full a class can be treated
  simultaneously as a collection of individuals and
  as an individual in its own right.

38
The Dublin Core

• The Dublin Core is a set of "elements"
  (properties) for describing documents (and hence,
  for recording metadata).
• The set was originally developed in 1995 at the
  Metadata Workshop in Dublin, Ohio.
• The goal is to provide a minimal set of
  descriptive elements that facilitate the
  description and the automated indexing of
  documents, similar to a library card catalog.
• is meant to be sufficiently simple to be
  understood and used by a wide range of authors.

39
The Dublin Core

• Title A name given to the resource.
• Creator An entity primarily responsible for
  making the content of the resource.
• Subject The topic of the content of the
  resource.
• Description An account of the content of the
  resource.
• Publisher An entity responsible for making the
  resource available
• Contributor An entity responsible for making
  contributions to the content of the resource.

40
The Dublin Core

• Date A date associated with an event in the life
  cycle of the resource.
• Type The nature or genre of the content of the
  resource.
• Format The physical or digital manifestation of
  the resource.
• Identifier An unambiguous reference to the
  resource within a given context.
• Source A reference to a resource from which the
  present resource is derived.
• Language A language of the intellectual content
  of the resource.
• Relation A reference to a related resource.
• Coverage The extent or scope of the content of
  the resource.
• Rights Information about rights held in and over
  the resource.

41
The Dublin Core

• Information using the Dublin Core elements may be
  represented in any suitable language (e.g., in
  HTML meta elements).
• However, RDF is an ideal representation for
  Dublin Core information.

42
XML
URI Uniform Resource Identifier
provide names
RDF Resource Description Framework
Dublin Core a set of core elements
simple statements
OWL Web Ontology Language
complex logical statements
43
Summary

• There are many uses for Natural Language
  Processing with respect to ontologies.
• Ontologies play an important role in the vision
  of the Semantic Web.