Ontologies - LSDIS - University of Georgia

advertisement

Ontologies - What, why and how?

Cartic Ramakrishnan

LSDIS lab

University of Georgia taken from Schulze-Kremer Steffen

What is Ontology

• The study of being

qua being

: the study of

possible

• The study of the

nature

of possible: ontology as the

theory of distinctions

among

possibilia

• The study of the most general characteristics that anything must have in order to count as a (certain kind of)

being

or

entity

.

taken from Schulze-Kremer Steffen

Definitions

• Ontology (capital “o”):

– a

philosophical discipline

.

• An ontology (lowercase “o”):

specific artifact

designed with the purpose of

expressing the intended meaning of a vocabulary

taken from Schulze-Kremer Steffen

What is an ontology?

• A

shared

vocabulary

• Plus … A specification (actually, a

characterization)

of the

intended meaning

of that vocabulary

...i.e., an ontology accounts for the

commitment

of a language to a certain

conceptualization

“An ontology is a specification of a

Models and

Conceptualizations

taken from Schulze-Kremer Steffen

Capturing Intended Meaning

• First order logic is

ontologically neutral

• Logical KBs often rely on

natural language

to convey intended meaning

Red(x) this apple is red

"purple" is a red

John i s a red

Apple(x) this is an apple this is apple taken from Schulze-Kremer Steffen

Intended Models

An ontology consisting of

just

a vocabulary is of little use -

Unintended interpretations

need to be excluded

Models

M(L)

Intended models

I

K

(L)

taken from Schulze-Kremer Steffen

What is a conceptualization?

a b c d e

Scene 1: blocks on a table

Conceptualization of scene 1:

<{a, b, c, d, e }, {on, above, clear, table }> taken from Schulze-Kremer Steffen

What is a conceptualization?

c a b d e

Scene 2: a different arrangement of blocks

The same conceptualization?

taken from Schulze-Kremer Steffen

What is a conceptualization

Conceptualization:

the formal structure of reality as perceived and organized by an agent, independently of:

– the

vocabulary

used (i.e., the

language

used)

– the actual occurence of a specific

situation

• Different situations involving the same objects, described by different vocabularies, may share the same conceptualization.

L

E

L

I apple

same conceptualization

mela taken from Schulze-Kremer Steffen

Ontologies

constrain

the intended meaning

Intended models

I

K

(L)

Conceptualization

C

Commitment

K=<C,

I

>

Language L

Models

M(L)

Ontology taken from Schulze-Kremer Steffen

Levels of

Ontological Depth

(SSK would disagree)

• Lexicon

– Vocabulary with NL definitions

• Simple Taxonomy

• Thesaurus

– Taxonomy plus related-terms

• Relational Model (

NOT DB)

– Unconstrained use of arbitrary relations

• Fully Axiomatized Theory taken from Schulze-Kremer Steffen

ontologies - ontology

• Ontology

– study of

being

as a branch of philosophy

• Ontologies

– result of the analysis of a particular domain of interest (possibly as broad as the universe)

– instantiation of a concrete ontological model of that domain taken from Schulze-Kremer Steffen

Ontologies-ontology

• Ontologies are to a large extent in principle language independent

• Varying scope and content of domain ontologies

– upper-level ontologies (Cyc)

– application ontologies (??)

– task ontologies (??) taken from Schulze-Kremer Steffen

Definitions - Yes, Again!!

• Here are three definitions of domain ontologies

– (i) "System of categories accounting for a particular vision of the world." [ Guarino ]

– (ii) "Specification of a conceptualization."

[ Gruber ]

– (iii) "Concise and unambiguous description of principle relevant entities with their potential, valid relations to each other." [ guess who?

] taken from Schulze-Kremer Steffen

What an Ontology is NOT!!!

not a collection of facts arising from a specific situation not a model of an application domain not a database schema not a knowledge base not a taxonomy not a vocabulary or dictionary not a semantic net

taken from Schulze-Kremer Steffen

Why ontologies?

• Data integration

– Semantic integration of

n

databases

• without the great “o” would require

n*n

integration attempts

• with the great “o” would require n attempts

• Data annotation

– full-fledged ontology not required

• since main purpose is fixed unique reference point in the for of controlled vocabulary taken from Schulze-Kremer Steffen

ENTER GO!!!

• Not ontology but vocabulary

– ISA vs. instance-of

– PART-OF vs “made of”, “belongsto”

– 80% concepts lack explicit defn.

– 700 concepts are orphans

– no clear design principle

– no IC for consistency

– where do new concepts go?

– No grammar rules for to combine concept

References

Guarino, N. (1998). Some Ontological Principles for Designing Upper Level

Lexical Resources.

In:

Proceedings of First International Conference on

Language Resources and Evaluation. Granada, Spain.

Gruber, T. R. (1993). Knowledge Acquisition 5, 199-220.

Schulze-Kremer, S. (1998) Ontologies for molecular biology taken from Schulze-Kremer Steffen

Download