openhpi 4.3 - how do i define a formal model of an ontology

18
This file is licensed under the Creative Commons Attribution-NonCommercial 3.0 (CC BY-NC 3.0 ) Dr. Harald Sack Hasso Plattner Institute for IT Systems Engineering University of Potsdam Spring 2013 Semantic Web Technologies Lecture 4: Knowledge Representations I 03: How Do I Define a Formal Model of an Ontology

Upload: harald-sack

Post on 16-Jan-2015

379 views

Category:

Education


3 download

DESCRIPTION

 

TRANSCRIPT

Page 1: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

This file is licensed under the Creative Commons Attribution-NonCommercial 3.0 (CC BY-NC 3.0)

Dr. Harald Sack

Hasso Plattner Institute for IT Systems Engineering

University of Potsdam

Spring 2013

Semantic Web Technologies

Lecture 4: Knowledge Representations I03: How Do I Define a Formal Model of an Ontology

Page 2: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

2

Lecture 4: Knowledge Representations I

Open HPI - Course: Semantic Web Technologies

Page 3: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

316

03 How do I define a formal model of an ontology?

Open HPI - Course: Semantic Web Technologies - Lecture 4: Knowledge Representations I

Page 4: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

416• in propositional logic the world consists simply of facts and

nothing else (statements of assertions)

• Example for propositional logic assertions and deductions:

• If it rains, the road will get wet.• If the moon is made out of green cheese, then cows can fly.

• If Oliver is in love, then he will be happy.

• The world consists out of objects and properties that distinguish one object from another.

• Between objects are relations. Some relations are unique, i.e. functions.

Propositional Logic

Page 5: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

516• In First Order Logic (FOL) quantors allow assertions about sets

of objects, without naming the objects explicitely.

• All humans are mortal.• Socrates is a human.• Socrates is mortal.

• FOL is perfectly suited for the description of ontologies, but...

• FOL is rather expressive,

• therefore also rather bulky for modelling,

• difficult to achieve consense in modelling and

• rather complex to proof (correctness and completeness of assertions)

• Therefore: look for some well suited fragment of FOL!

First Order Logic

Page 6: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

616

Description Logics

Description Logics are a family of languages for knowledge representation. Most description logics are a subset of First Order Logic, but in difference to FOL most description logics are decidable. Therefore, it is possible to make logical deductions based on description logics, i.e. to create new knowledge from existing knowledge.

Page 7: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

616

Description Logics

Description Logics are a family of languages for knowledge representation. Most description logics are a subset of First Order Logic, but in difference to FOL most description logics are decidable. Therefore, it is possible to make logical deductions based on description logics, i.e. to create new knowledge from existing knowledge.

Lecture

Lecture„Semantic Web Technologies“

TBox terminological knowledge Knowledge about concepts of a domain (classes, attributes, relations…)

ABox assertional knowlegde knowledge about instances / entities

Knowledge Base

Page 8: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

716 • Concepts (unary predicates),

• represent entities / classes

• e.g., Person, Course, Student, Lecturer, Seminar, ...

• Roles (binary predicates, properties)

• represent properties / relations

• e.g., participatesAt, givesLecture, isGivenByLecturer, …

Student: { x | Student(x)}

participatesAt: {(x,y)|participatesAt(x,y)}

Description Logics - A Brief Summary

Page 9: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

816 • Individuals (constants, individual entities, concept assertion)

• e.g., Alice, Bob, SemanticWeb

• Syntax: Student(Alice)

• Operators / Constructors (to construct complex representations of concepts / roles)

• Expressivity is limited:

• Satisfiability and Subsumption is decidable and

• (preferably) of low complexity

• Syntax: participatesAt(Alice, SemanticWeb)

Description Logics - A Brief Summary

Page 10: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

916• Fundamental operators:

• Conjunction (⊓),

• Disjunction (⊔),

• Negation (⌐)

• restricted form of Quantification (∀,∃)

• represents Basic Description Logic ALC• Attributive Language with Complement

Description Logics

Page 11: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

1016

Attributive Language with Complement - ALC• Atomic Types

• concept names A, B, ...

• special concepts

• ⊤ - Top (universal concept)

• ⊥ - Bottom concept

• role names R,S, ...

• Constructors

• Negation: ¬C

• Conjunction: C ⊓ D

• Disjunction: C ⊔ D

• Existential Quantifier: ∃R.C

• Universal Quantifier: ∀R.C

Page 12: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

1016

Attributive Language with Complement - ALC• Atomic Types

• concept names A, B, ...

• special concepts

• ⊤ - Top (universal concept)

• ⊥ - Bottom concept

• role names R,S, ...

• Constructors

• Negation: ¬C

• Conjunction: C ⊓ D

• Disjunction: C ⊔ D

• Existential Quantifier: ∃R.C

• Universal Quantifier: ∀R.C

∃attends.Lecture

Defines a Class

property / role

range restriction

Page 13: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

1116 • Class Relations

• Inclusion C ⊑ D

• E.g., Man ⊑ Human

• Equality C ≣ D

• E.g., Frau ≣ Woman

• Class Constructors

• E.g., Seminarist ≡ Person ⊓ (∃participatesAt.Seminar ⊔ ∃manages.Seminar)

Attributive Language with Complement - ALC

Page 14: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

1216 • Terminological Knowledge (TBox)

• Axioms describing the structure of the represented domain(conceptional schema)

• Human ⊑ ∃hasParent.HumanOrphan ≣ Human ⊓ ¬∃hasParents.Alive

• Assertional Knowledge (ABox)

• Axioms describing specific situations (data)

• Orphan(harrypotter)hasParent(harrypotter, jamespotter)

Attributive Language with Complement - ALC

Page 15: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

1316 Operator / Constructor Syntax LanguageLanguage

Conjunction A ⊓ B

FL

S*

Value Restriction ∀ R .C FL

S*

Existential Quantification ∃ R

FL

S*

Top (Universal Concept) ⊤

AL*

S*Bottom (Most Special Concept) ⊥

AL*

S*Negation (C) ⌐ A

AL*

S*

Disjunction C ⊔ D AL*

S*

Existential Restriction ∃ R .C

AL*

S*

Cardinality Restriction (N) (≤ n R) (≥ n R)

AL*

S*

Set of Individuals (O) {a1,…,an}

AL*

S*

Hierarchy of Relations R ⊆ S HH

Inverse Relation R-1 II

Qualified Cardinality Restriction (≤ n R.C) (≥ n R.C) QQ

Description Logics

Page 16: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

1416 • Semantics is determined via Interpretation (ΔI, I)

• ΔI … Domain of Discourse, ΔI ≠ ∅

• Interpretation Function:

• I :A " AI ⊆ ΔI , A ... atomic concept

• I :R " RI ⊆ ΔI x ΔI , R … atomic role/property

$I # # = # ΔI⊥I # # =# ∅(¬A)I # # =# ΔI \ AI # # #(C ⊓ D)I # =# CI ∩ DI # # #(∀R.C)I # =# {a ∈ ΔI | ∀ b.<a,b> ∈ RI ⇒ b ∈ CI}

(∃R.$)I # =# {a ∈ ΔI | ∃ b.<a,b>∈ RI}#

Semantics of Description Logics

Page 17: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

1516

• Ontologies can also be modelled via database or softwaremodelling technologies, as e.g.

• UML, ER-Model, …

Seminar

- Titel: String - Semester: String - Begin: Date - End: Date - …

Person

- GivenName: String - FamilyName: String - …

participatesAtnn

givesLecture1n

How should we represent Ontologies?

Page 18: OpenHPI 4.3 - How DO I Define a Formal Model of an Ontology

Semantic Web Technologies , Dr. Harald Sack, Hasso-Plattner-Institut, Universität Potsdam

1616

04 Ontology TypesOpen HPI - Course: Semantic Web Technologies - Lecture 4: Knowledge Representations I