Communities building ontologies: Tensions and Reality

21
1 http://img.cs.man.ac.uk/s tevens Communities building ontologies: Tensions and Reality Robert StevensUniversity of Manchester [email protected]

description

Invited talk at an ontology meeting at harwell 2003

Transcript of Communities building ontologies: Tensions and Reality

Page 1: Communities building ontologies: Tensions and Reality

1http://img.cs.man.ac.uk/stevens

Communities building ontologies: Tensions and Reality

Robert StevensUniversity of Manchester

[email protected]

Page 2: Communities building ontologies: Tensions and Reality

2http://img.cs.man.ac.uk/stevens

Introduction

• CS & domain experts• Why do CS people go on about semantics?• CS version of building ontologies• Some reality• Resolving some tensions

Page 3: Communities building ontologies: Tensions and Reality

3http://img.cs.man.ac.uk/stevens

Two CommunitiesComputer ScientistsBuilding ontologiesKRReasoning

Better Ontologies

BiologistsOntology content

Domain Knowledge

Page 4: Communities building ontologies: Tensions and Reality

4http://img.cs.man.ac.uk/stevens

So what Counts as an ontology? [Deborah McGuinness, Stanford]

Catalog/ID

Thesauri

Terms/glossary

Informal Is-a

FormalIs-a

Formalinstance

Frames(properties)

General Logicalconstraints

Valuerestrictions

Disjointness,Inverse, partof

Gene Ontology

Mouse AnatomyEcoCyc

PharmGKB

TAMBISArom

Page 5: Communities building ontologies: Tensions and Reality

5http://img.cs.man.ac.uk/stevens

Semantics– An operational semantics for a language is

defined by what a sentence in that language will do.

– Denotational semantics is a precise mathematical definition of the objects and relations of language in which each sentence of the language names, or denotes, a mathematical object, such as a function.

– Natural semantics are the loose ordinary language sense, in which the semantics of a statement is its "meaning".

– The term logicist semantics refers to formal models that attempt to represent the natural semantics of some external domain.

Page 6: Communities building ontologies: Tensions and Reality

6http://img.cs.man.ac.uk/stevens

Malate Dehydrogenase Activity• malate dehydrogenase activity • GO:0016615• definition: Catalysis of the reversible

conversion of pyruvate or oxaloacetate to malate using NAD or NADPH.

Page 7: Communities building ontologies: Tensions and Reality

7http://img.cs.man.ac.uk/stevens

What are We Saying?Person

WomanMan

is-ais-a

•Are all instances of Man instances of Person?•Can an instance of Person be both a Manand an instance of Woman?•Can there be any more kinds of Person?

Page 8: Communities building ontologies: Tensions and Reality

8http://img.cs.man.ac.uk/stevens

What are we saying?

• What kinds of class can fill “has chromosome”?• How many “Y chromosome” are present?• Does their have to be a “Y chromosome”?• What properties are sufficient to be a Man and which are simply

necessary?

Y chromosome

Manhas chromosome

Y chromosome

Manhas chromosome

X chromosome

has chromosome

autosomehas chromosome

1

1

44

Page 9: Communities building ontologies: Tensions and Reality

9http://img.cs.man.ac.uk/stevens

What Does the Computer Know?#1234

#5678#9101

#1121

#1121

•Knows that all instances of #5678 are members of #1234•Knows that #5678 & #9101 are disjoint•Knows that #5678 & #9101 are the only kinds of #1234

Page 10: Communities building ontologies: Tensions and Reality

10http://img.cs.man.ac.uk/stevens

malate dehydrogenase

class malate dehydrogenase defined

subClassOf enzymatic_function

restriction onProperty has_reagent_on_side_A has-class malaterestriction onProperty has_reagent_on_side_B has-class oxaloacetaterestriction onProperty has_reagent_on_side_A has-class NADP anionrestriction onProperty has_reagent_on_side_B has-class NADPH

restriction onProperty catalyses has-class

((reducing and (restriction onProperty acts_on has-class NADP))

and (oxidising and (restriction onProperty acts_on has-class malate)

and (restriction onProperty acts_on_donar_group has-class CH-OH group)))restriction onProperty catalyses has-class

((reducing and (restriction onProperty acts_on has-class oxaloacetate))

and (oxidising and (restriction onProperty acts_on has-class NADPH)))

restriction onProperty catalyses to-class

(((reducing and (restriction onProperty acts_on has-class NADP)

and (oxidising and (restriction onProperty acts_on has-class malate)

and (restriction onProperty acts_on_donar_group has-class CH-OH group)))

or ((reducing and (restriction onProperty acts_on has-class oxaloacetate))

and (oxidising and (restriction onProperty acts_on has-class NADPH))))

Page 11: Communities building ontologies: Tensions and Reality

11http://img.cs.man.ac.uk/stevens

Building Ontologies

• No field of Ontological Engineering equivalent to Knowledge or Software Engineering;

• Developing standard methodologies for building ontologies;• Such a methodology would include:

– a set of stages that occur when building ontologies;

– guidelines and principles to assist in the different stages;

– an ontology life-cycle which indicates the relationships among stages.

• Gruber's guidelines for constructing ontologies are well known.

Page 12: Communities building ontologies: Tensions and Reality

12http://img.cs.man.ac.uk/stevens

The Development Lifecycle• Two kinds of complementary methodologies emerged:

– Stage-based, e.g. TOVE [Uschold96] – Iterative evolving prototypes, e.g. MethOntology [Gomez Perez94].

• Most have TWO stages:– Informal stage

• ontology is sketched out using either natural language descriptions or some diagram technique

– Formal stage

• ontology is encoded in a formal knowledge representation language, that is machine computable

• An ontology should ideally be communicated to people and unambiguously interpreted by software

– the informal representation helps the former – the formal representation helps the latter.

Page 13: Communities building ontologies: Tensions and Reality

13http://img.cs.man.ac.uk/stevens

A Provisional Methodology

• A skeletal methodology and life-cycle for building ontologies;• Inspired by the software engineering V-process model;

• The overall process moves through a life-cycle.

The left side charts the processes in building an ontology

The right side charts the guidelines, principles and evaluation used to ‘quality assure’ the ontology

Page 14: Communities building ontologies: Tensions and Reality

14http://img.cs.man.ac.uk/stevens

Methodology

Conceptualisation

Integrating existing ontologies

Encoding

Representation

Identify purpose and scope

Knowledge acquisition

Evaluation: coverage, verification, granularity

Conceptualisation Principles: commitment, conciseness, clarity, extensibility, coherency

Encoding/Representation principles: encoding bias, consistency, house styles and standards, reasoning system exploitation

Ontology in Use

User Model

Conceptualisation Model

Implementation Model

Ontology Learning

Maintenance

Page 15: Communities building ontologies: Tensions and Reality

15http://img.cs.man.ac.uk/stevens

An Ontology Building Life-cycleIdentify purpose and scope

Knowledge acquisition

Evaluation

Language and representation

Available development tools

Conceptualisation

Integrating existing ontologiesEncoding

Building

Ontology Learning

Consistency Checking

Page 16: Communities building ontologies: Tensions and Reality

16http://img.cs.man.ac.uk/stevens

Some Reality & Tensions

• Conceptualisation & encoding conflated• Encoding rarely driven by system requirements• Very difficult to re-use other ontologies• Not really sequential stages, but overlapping• Little support for communal development and comment

Page 17: Communities building ontologies: Tensions and Reality

17http://img.cs.man.ac.uk/stevens

More Reality & Tensions

• The community must need the ontology• The community must build the ontology• There is no one true ontology• Don’t wait for completeness• Don’t wait for correctness• Simple representation gives rapid start• DL too difficult to use

Page 18: Communities building ontologies: Tensions and Reality

18http://img.cs.man.ac.uk/stevens

More Tips from GO

• Policy for updates of concept labels, Ids and definitions• Separate concept labels and Ids• Track obselete terms: “replaces” and “replaced by”• Record provenance of term’s origins• Be open

Page 19: Communities building ontologies: Tensions and Reality

19http://img.cs.man.ac.uk/stevens

Some Solutions

• Simple representation often unsustainable• DL gives better sustainability• Need domain experts to build, but ontologist to refine• DL representation bad for users: Need familiar representation• Reconciling simplicity and sustainability• Applying CS knowledge appropriately

Page 20: Communities building ontologies: Tensions and Reality

20http://img.cs.man.ac.uk/stevens

Migration path

OWL ontologyDAG like ontology

i. Semantic diff

ii. Delivery in familiar format

iii. Consistent and taxonomically complete

iv. Property based and reasonable

i. Communal knowledge acquisition

ii. Provenance of acquisition

iii. Communal commentary

iv. Consensus

i. Report changes

ii. Semantic diff

Communal ontology development

Page 21: Communities building ontologies: Tensions and Reality

21http://img.cs.man.ac.uk/stevens

Things, Symbols & Concepts• Humans require words (or at least symbols) to

communicate efficiently. The mapping of words to things is only indirectly possible. We do it by creating symbols that stand for things.

• The relation between symbols and things has been described in the form of the meaning triangle:

“Jaguar“

Concept

[Ogden, Richards, 1923]