Semantic Technologies and Applications COMP5860M
John Stell
Room 9.15, School of Computing
j.g.stell@leeds.ac.uk
Lecture 2: January 2020
1
Outline
What is an ontology?
Individuals, Classes, Relationships Common mistakes to avoid
Open World vs Closed World
2
What is an ontology?
Definition (Noy & McGuiness)
An ontology defines a common vocabulary for people who need to share information in a domain.
It includes machine-interpretable definitions of basic concepts in the domain and relations among them.
3
Why develop an ontology?
Share common understanding of the structure of information among people or software agents
Make domain knowledge explicit and reusable
Compare domains, applications, contexts
Facilitate collaboration
Provide a model to study a domain
4
An ontology Includes:
Classes (concepts, kinds, types, sorts of thing)
Relationships (properties, roles)
Instances (individuals)
Knowledge Base: Classes + Relationships + Instances
Beware: names are not standard. Description Logic talks about roles rather than properties for example. Some properties may be called attributes.
5
In an ontology
As well as saying what the Classes and Instances and Relationships are, we often:
Say that an individual belongs to a class Dumbo is an Elephant
Say that one class is a subclass of another An Elephant is an Animal
Define new classes out of existing ones and relationships The class of Students studying more than 6 modules
State features that relationships have
If A is married to B then B is married to A
6
An ontology has Classes
Classes (also called concepts)
Play a fundamental role in the ontology;
Define the main concepts in the domain;
Classes are collections;
Class represents a set of individuals (e.g. class University can include many individual universities)
When we define characteristics of classes, they apply to all individuals in that class (e.g. if we say that University is a Higher Education Institution – we apply this to every individual that belongs to the class University
7
Classes have Hierarchy
Define inheritance between classes; Multiple inheritance usually allowed
Allow us to specify common characteristics between classes
Based on their place in the hierarchy, a class can be:
superclass (parent) of another class
subclass (child) of another class
Every subclass inherits the characteristics of its superclass in the domain;
8
Exercise: These are Classes, draw a Class Hierarchy
vehicle
flying fox
chess board
boat water-going-entity means of transport horse dog elephant toy building lorry vessel juggernaut carriage animal fox Lego model building mobile home temporary building
mobile building Hogwarts Castle fictional entity
red panda giant panda flying thing bird
aircraft duck
flying boat skateboard
flying fortress surfboard
board
plank
castle toy
9
Ontologies have Individuals also called instances
These are the specific objects we will see in our world (e.g. in a database including the data from the world)
Allow us to add concrete things from the world
Every individual belongs to at least one class
Individuals inherit the attributes of their class(es), and have specific values (this is what differentiates them from other individuals in the class)
Individuals inherit the links of their class(es), i.e. link to individuals from other classes
but what is an individual can depend on purpose of ontology
10
Is this an individual?
The Lego model of Hogwarts?
The book ‘Jane Eyre’ by Charlotte Bronte?
11
Is this an individual? – it all depends
A class will be a collection of some individuals, but it is possible that individuals are not physical things; maybe we can think of them as collections (e.g. copies of a book) but only in a way that we do not want to model in the particular ontology.
The text of a book could be an individual in one ontology belonging to a class of works by the author
a different ontology might have all physical copies of the same text as individuals in one class; another ontology might have different editions as individuals.
You have to think very carefully about the purpose of the model (the ontology)
12
Ontologies have Relationships / Properties / Roles / Attributes
In OWL relationships are
between individuals (object properties) or
relate an individual to a value (datatype properties).
Elsewhere often called attributes
Allow us to define classes
(e.g. all individuals having certain properties)
Relationships/properties specify the main attributes (by naming them and defining restrictions) which each individual in the class should have
Allow us to define links between classes
Relationships/properties can specify the link between two classes by naming it and defining restrictions
13
Common Mistakes for Beginners
Using a word without thinking carefull what it means and that other people might not understand you. If you just say “the class University” is it clear what the instances of this class are? Might you mean all the students and staff and buildings in one university? Choosing names carefully is important and explaining them too.
Confusing Individuals and Classes. A class is a collection, even if there is only one member.
Not thinking hard enough whether a class with only one member is really needed in an ontology
Confusing class hierarchy with other relationships. E.g. part-of and is-a are quite different
Using diagrams with a network of classes and individuals and relationships without being careful which is which
14
http://www.ebi.ac.uk/ontology-lookup/
Ontologies from European Bioinformatics Institute.
Including among almost 250 ontologies
Basic Formal Ontology
Biological Spatial Ontology
Human Disease Ontology
Information Artifact Ontology
15
Open World and Closed World
Given these facts and nothing else,
mod2345 is a Level2Module
mod3678 is a Level3Module
Ammar is a member of staff who teaches on mod2345
Vania is a member of staff who teaches on mod3678 and mod2345
Does Ammar belong to the class of staff who ONLY teach Level2Modules?
16
Open World and Closed World
“The open world assumption means that we cannot assume something doesn’t exist until it is explicitly stated that it does not exist.
In other words, because something hasn’t been stated to be true, it cannot be assumed to be false – it is assumed that ‘the knowledge just hasn’t been added to the knowledge base’.” [Pizza tutorial]
Under Open World Assumption, we cannot conclude Ammar only teaches Level2Modules.
17
Open World and Closed World
Ontologies typically work with the Open World Assumption.
Relational Databases typically use the Closed World Assumption.
The distinction is important.
It’s not a question of which is right or wrong; there are different ways of interpreting data and we need to be aware of this
18