Ontology Design Patterns


Monsoon 2018
Instructor: Raghava Mutharaju
IIIT-Delhi
IIIT Delhi

Introduction


  • Distinctive and repetitive invariants across data, objects, processes etc.
  • Patterns emerge from data, ontologies, queries etc.
  • Majority of ontologies share a small set of modelling or publishing challenges that can be solved using a common strategy or an established best practice
  • An ODP should have a motivating requirements, applicability limits, benefits, and shortcomings associated with it
  • ODPs are analogous to software design patterns

Examples


  • Trajectory Pattern
    • Can be used in transportation domain, wildlife monitoring, scientific cruises
  • Information Realization Pattern
    • Describes relation between an information object (book) and its physical realization (paper copy of the book)

ODP Types

  • Several types of ODPs
    • Structural ODPs
    • Correspondence ODPs
    • Content ODPs
    • Reasoning ODPs
    • Presentation ODPs
    • Lexico-Syntactic ODPs
  • Available at ODP Repository
ODP Types
Image source: http://ontologydesignpatterns.org/wiki/Image:Odptypes.jpg

Structural ODPs


  • Logical ODPs and Architectural ODPs make up the Structural ODPs
  • Logical ODPs
    • Composition of logical constructs that solve a problem of expressivity
    • Eg: Expressing n-ary relation between concepts
    • They are further divided into Logical Macros and Transformation patterns
    • Logical Macros are shortcuts to recurrent intuitive logical expressions
    • Transformation patterns translate a logical expression from one language to the other

Structural ODPs


  • Architectural ODPs
    • They have an affect on the overall shape of the ontology
    • They aim to constrain how the ontology should look like
    • They are chosen if there are specific needs such as any computational complexity constraints

Reasoning ODPs


  • They are applications of Logical ODPs to obtain certain reasoning results
  • Eg: Classification (all superclasses for each class in the ontology)
  • They provide hints to the reasoner as to what kind of reasoning has to be performed in order to answer queries

Correspondence ODPs


  • Correspondence ODPs are divided into Reengineering ODPs and Mapping (or Alignment) ODPs
  • Reengineering ODPs
    • They are design solutions that transform a conceptual model into an ontological model
    • The source model can be a non-ontological model such as a thesaurus, UML model etc.
    • Schema reengineering patterns and Refactoring patterns are sub categories of Reengineering ODPs
    • Schema reengineering patterns are rules to transform a non-OWL ontology to an OWL ontology

Correspondence ODPs


  • Refactoring patterns transform (refactor) models in the same form. For example, from one OWL 2 DL ontology to another
    • Eg: An ontology defines an object property to represent preparing coffee between agent and coffee
    • Time of coffee preparation and the tool used should also be added
    • An n-ary Logical ODP needs to be used because the property preparing coffee needs to be associated with agent, coffee, time, tool
    • The n-ary Logical ODP and the mechanism to replace the existing object property together form the Refactoring ODP

Correspondence ODPs


  • Mapping patterns provide design solutions to relate two ontologies without changing the logical types of the ontology entities
    • They provide semantic relations between mappable elements in the ontologies
    • There are 3 basic semantic relations used for mapping: equivalence, containment, and overlap
    • The negative counterparts are also included such as not equivalent, not contained, and no overlap or disjoint

Presentation ODPs


  • They deal with usability and readability of ontologies from a user perspective
  • They are meant to be good practices that support reuse of ontologies
    • Naming ODPs and Annotation ODPs are types of Presentation ODPs
    • Naming ODPs are naming conventions for files, namespaces, and ontological entities (class, property, instace)
    • Annotation ODPs provide annotation properties such as label (description of an entity, synonyms etc.), seeAlso, versionInfo etc.

Lexico-Syntactic ODPs


  • They are linguistic structures or schemas that consist of words following specific order
  • They can be used to associate natural language sentences with other ODPs such as Logical and Content ODPs

Content ODPs (CPs)


  • Small ontologies that mediate between use cases (problem types) and design solutions
  • They are the building blocks to create ontologies
  • An ontology can be built by composing different CPs with appropriate dependencies between them and by expanding the CPs as needed

ODP Description

  • An ODP can be described using the following fields
    • Name
    • Intent
    • Competency questions
    • Alternate names
    • Scenarios: Examples of requirements in natural language that can be modelled using this pattern
    • Diagram: UML diagram representing the pattern
    • Elements: list classes and relations in the pattern
  • Consequences: description of benefits and trade-offs
  • Known uses: examples of realistic ontologies where the pattern can be used
  • Extracted/Reengineered from: reference ontology or conceptual schema from which the pattern has been extracted
  • Related patterns: other patterns that are either a specialization, generalization, composition, or component of this pattern. Also lists patterns that are generally used along with the described pattern
  • Building block: a reference implementation (URI or owl file)

AgentRole ODP

AgentRole ODP
Image source: http://ontologydesignpatterns.org/wiki/Image:Agentrole.jpg
  • Name: Agent Role
  • Intent: To represent agents and the role they play
  • Competency questions
    • what is the role that is played by an agent?
    • which agent plays this role?
  • Consequences: allows designers to make assertions on roles played by agents without involving the agents that play those roles, and vice versa. It does not allow to express temporariness of roles
  • Scenarios: She greeted us all in her various roles of mother, friend, and daughter
AgentRole ODP
Image source: http://daselab.cs.wright.edu/pub2/2017-05-ESWC-Tutorial-Hitzler-Part-1.pdf
AgentRole ODP
Image source: http://daselab.cs.wright.edu/pub2/2017-05-ESWC-Tutorial-Hitzler-Part-1.pdf
AgentRole ODP
Image source: http://daselab.cs.wright.edu/pub2/2017-05-ESWC-Tutorial-Hitzler-Part-1.pdf

Information Realization ODP

Information Realization ODP
Image source: http://ontologydesignpatterns.org/wiki/Image:Informationrealization.jpg
  • Name: Information Realization
  • Description: It represents the relations between information objects like poems, songs, formulas, etc., and their physical realizations like printed books, registered tracks, physical files, etc
  • Intent: To represent information objects and their physical realization.
  • Competency questions
    • what are the physical realizations of this information object?
    • what information objects are realized by this physical object?
  • Consequences: This pattern allows to distinguish information objects from their concrete realizations
  • Scenarios: The book of the "Divina Commedia"

Provenance ODP

Provenance ODP
Image source: http://ontologydesignpatterns.org/wiki/Image:Provenance.png

Provenance ODP in use

Provenance ODP
Image source: http://daselab.cs.wright.edu/pub2/mom-recipes-example.pdf

Plan ODP

Plan ODP
Image source: http://daselab.cs.wright.edu/pub2/mom-recipes-example.pdf

Event ODP

Plan ODP
Image source: http://daselab.cs.wright.edu/pub2/01-chess-example.pdf

Event ODP in use

Plan ODP
Image source: http://daselab.cs.wright.edu/pub2/01-chess-example.pdf

Collections ODP

Collections ODP
Image source: http://ontologydesignpatterns.org/wiki/Image:List.jpg

References


  1. Ontology Design Patterns in a Nutshell. Krzysztof Janowicz et. al. Ontology Engineering with Ontology Design Patterns: Foundations and Applications. IOS Press.
  2. Ontology Design Patterns. Aldo Gangemi and Valentina Presutti. Handbook on Ontologies. 2009