Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

45
1 Comparison and Combination of the Expressive Power of Description Logics and Logic Programs Jidi (Judy) Zhao June 28, 2022

description

Comparison and Combination of the Expressive Power of Description Logics and Logic Programs. Jidi (Judy) Zhao October 9, 2014. Motivation for Extending Description Logics with Horn Logic Rules. By Benjamin Grosof, May, 2003. 2. Examples of LP not representable in DL. - PowerPoint PPT Presentation

Transcript of Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

Page 1: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

1

Comparison and Combination ofthe Expressive Power of

Description Logics and Logic Programs

Jidi (Judy) ZhaoApril 24, 2023

Page 2: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

2

Motivation for Extending Description Logics with Horn Logic

Rules

2By Benjamin Grosof, May, 2003

Page 3: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

3

Examples of LP not representable in DL DL cannot represent “more than one

free variable at a time”. FriendshipBetween(?X,?Y) ← Man(?X) ∧ Woman(?Y). DLs cannot directly support n-ary

predicates Traditional expressive DLs support

transitive role axioms but they cannot derive values of properties

uncleOf (?X,?Z) ←brotherOf(?X,?Y) ∧ parentOf(?Y,?Z).

HomeWorker(?X) ←

Work(?X, ?Y) ∧ Live(?X, ?Z) ∧ Loc(?Y,?W) ∧ Loc(?Z,?W)

X

YWork

Z

Live W

Loc

Loc

Page 4: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

4

Examples of DL not representable in LP

•Horn Logic cannot represent a (1) disjunction or (2) existential in the head.•(1) State a subclass of a complex class expression which is a disjunction. E.g.,

(Human u Adult) v (Man t Woman)•(2) State a subclass of a complex class expression which is an existential. E.g.,

Radio v 9hasPart.Tuner4

Page 5: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

5

Differences between DLs and LPs Description Logics

Open World Assumption (OWA) May exist many models Generally no Unique Name Assumption (UNA) Classical negation

Logic Programs Closed World Assumption (CWA) Only one model Unique Name Assumption (UNA) Negation As Failure (NAF)

5

Page 6: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

6

Semantic Web Layer Cake

URI/IRI

Data interchange:

Rules: RIF

Unifying Logic

Trust

Proof

Ontology:OWL

Crypto

RDFS

User Interface & Applications

XML

Query:

SPARQL

RDF

Page 7: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

7

Different approaches1. approaches reducing description logics to logic programs

A. DLPB. OWL-R DL and OWL 2 RL

2. Homogeneous approachesA. OWL RulesB. SWRL

3. hybrid approaches accessing description logics through queries in logic programsA. AL-Log

Page 8: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

8

Expressiveness of Description Logic Programs (DLP)

Page 9: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

9

DLP comprises basic RDFS & more

by Benjamin Grosof et al.•RDFS subset of DL permits the following statements:

•Subclass, Domain, Range, Subproperty (also SameClass, SameProperty)•instance of class, instance of property

•more DL statements beyond RDFS:•Using Intersection connective (conjunction) in class descriptions•Stating that a property (or inverse) is Transitive or Symmetric•Using Disjunction or Existential in a subclass expression•Using Universal in a superclass expression

Page 10: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

10

DLP

•Figure 1. Relationship between the fragments (profiles) of OWL 1.1•http://www.webont.org/owl/1.1/tractable.html

Page 11: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

11

DLP mappings

Page 12: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

12

OWL 2 RL based on Description Logic Programs

[DLP] is a syntactic profile of OWL 2 DL. allows for scalable reasoning using

rule-based technologies. trades the full expressivity of the

language for efficiency http://www.w3.org/2007/OWL/wiki/Profiles#OWL_2

_RL12

Page 13: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

13

OWL 2 RL•achieved by restricting the use of OWL 2 constructs to certain syntactic positions.•Table 1. Syntactic Restriction on Class Expressions in SubClassOf Axioms

Subclass Expressions Superclass Expressions a class

a nominal class (OneOf)

intersection of class expressions (ObjectIntersectionOf)

union of class expressions (ObjectUnionOf)

existential quantification to a class expressions (ObjectSomeValuesFrom)

existential quantification to an individual (ObjectHasValue)

a class

intersection of classes (ObjectIntersectionOf)

universal quantification to a class expressions (ObjectAllValuesFrom)

at-most 1 cardinality restrictions (ObjectMaxCardinality 1)

existential quantification to an individual (ObjectHasValue)

Page 14: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

14

SWRL A Semantic Web Rule Language

Combining OWL and RuleML SWRL is undecidable SWRL with the restriction of DL Safe

rules is decidable Variables in DL Safe rules bind only to

explicitly named individuals in the ontology.

14

Page 15: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

15

AL-log [Donini et al., 1998]

Provides hybrid reasoning with representational adequacy and deductive power

An AL-log knowledge base K = (Σ, π) Σ is an ALC knowledge base, expressing

knowledge about concepts, roles and individuals. π is a constrained Datalog program

Defines an interface between DL and datalog by allowing Datalog program to “query” DL KB

15

Page 16: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

16

Example 1

FP=Full Professor, FM=Faculty Member, NFP=Nonteaching Full Professor,

AC=Advanced Course, BC=Basic Course, TC=Teaching, CO=Course,ST=Student, TP=Topic.

Page 17: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

17

Page 18: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

18

Conclusion of AL-Log Defines an interface between DL and datalog by

allowing datalog program to “query” DL KB Results of DL satisfiability check used for checking

constraints in query answering AL-log does not allow relational subsystem to

deduce knowledge about the structural subsystem No roles allowed in rule bodies

AL-log extended with roles in rule body by [Rosatti, 1999]

[Eiter et al., 2004] extend the approach for more expressive DLs and more expressive LP language

Page 19: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

19

Uncertainty extension of DL

Page 20: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

20

Motivation for Extending Description Logics with

Uncertainty“Everything is vague to a degree you do not realize till you have tried to make it precise.”

-------Bertrand Russell British author, mathematician, & philosopher

(1872 - 1970)Nobel Prize in Literature,1950

20

Page 21: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

21

Motivation for Extending Description Logics with

Uncertainty (Cont.) Uncertainty is an intrinsic feature of real-

world knowledge and refers to a form of deficiency or imperfection in the information.

The truth of such information is not precisely established.

People work and make decisions with imprecise data in an uncertain world.

21

Page 22: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

2222

URW3 Situation Report: uncertainty ontology

URW3

22

Page 23: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

23

Probability, Possibility and Fuzzy logic

Probabilistic Description Logic: Statistical information e.g. John is a student with the probability 0.6

and a teacher with the probability 0.4 Fuzzy Description Logic:

Express vagueness and imprecision e.g. John is tall with the degree of truth 0.9

Possibilistic Description Logic: Particular rankings and preferences e.g. John prefers an ice cream to a beer

23

Page 24: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

24

Probability, Possibility and Fuzzy logic (Cont.)

Previous work on uncertainty extension to DL can be classified based on (a) the generalization of classical

description logics (b) the supported forms of uncertain

knowledge (c) the underlying semantics (d) their inference problems and reasoning

algorithms.

24

Page 25: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

25

A norm-parameterized fuzzy description logic

[Zhao, Boley, Du, 2009]

Page 26: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

26

Fuzzy Sets Fuzzy sets and set membership is the key

to decision making when faced with uncertainty (Zadeh, 1965).

Fuzzy Logic is particularly good at handling vagueness and imprecision.

Generalize crisp sets to Fuzzy Sets (concepts).

26

Page 27: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

27

Fuzzy values

Cheetahs run very fast. John is young. Mary is old. John is tall.

27

Page 28: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

28

Membership Functions

28

Page 29: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

29

Fuzzy Operations fuzzy intersection (t-norm) fuzzy union (s-norm) fuzzy set complement (negation)

29

Page 30: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

30

A Knowledge Base (KB) <T,A>= a Tbox + an Abox

A TBox (terminology) is a finite set of fuzzy concept inclusion axioms

in FOC

fuzzy concept equivalence axioms

fuzzy DL Knowledge Bases(I)

30

Page 31: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

31

fuzzy role inclusion axioms fuzzy role equivalence axioms

An ABox (Assertion) is a set of fuzzy assertions about individuals fuzzy concept assertions fuzzy role assertions individual inequality

fuzzy DL Knowledge Bases (II)

31

Page 32: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

32

Semantics (I)

32

Semantics given by standard FO model theory and Fuzzy Logic

A fuzzy interpretation I is a tuple (I, •I) I is the domain (a set)•I is a mapping that maps:

Each object (individual/constant) to an element of I

Each unary predicate (classe/concept) C to a membership function of CI: I →[0,1]

Each binary predicate (propertie/role) R to a membership function of RI: I ×I →[0,1]

Page 33: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

33

Semantics (II)

33

Concept Negation

E.g. Concept Conjunction

E.g.

Page 34: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

34

Semantics (III)

34

Concept Disjunction

E.g.

Role Exists Restrictionin FOCexistential quantier: supremum or least upper bound

Page 35: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

35

Semantics (IV) Role Exists Restriction E.g.

Page 36: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

36

Semantics (V) At-least Number Restriction

in FOC

Inverse Role

Page 37: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

37

Semantics (VI)

37

Page 38: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

38

Reasoning Procedure

Page 39: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

39

Probabilistic reasoning in terminological

logics(Jaeger,1994) Propositional concept language (PCL)

Syntax: Terminological axioms Probabilistic terminological axioms Probabilistic assertions

Semantics: The probability measure that interprets an individual will

be defined by Jeffrey’s rule.

A C or A C

( | )P C D p( )P a C p

Page 40: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

40

Probabilistic reasoning in terminological

logics(Jaeger,1994) Reasoning Tasks:

(1)derive additional conditional probabilities.

(2) derive additional probabilistic assertions.

The former codifies statistical information that will be gained generally by observing a large number of individual objects and checking their membership of the various concepts.

The latter expresses a degree of belief in a specific proposition. Its value most often will be justified only by a subjective assessment of likelihood.

Page 41: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

41

Probabilistic reasoning in terminological

logics(Jaeger,1994) Example: TBox

PTBox

PABox

Page 42: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

42

Probabilistic reasoning in terminological

logics(Jaeger,1994) Reasoning on TBox and PTBox:

( _ | _ )( _ _ )

( _ )( _ _ )( _ | ) ( )( _ | _ ) ( _

u Flying bird Bird Antarctic birdP Flying bird Bird Antarctic bird

P Bird Antarctic birdP Flying bird Antarctic birdP Antarctic bird Bird P BirdP Antarctic bird Flying bird P Flying bir

)

( _ | ) ( )(1 ( _ | _ )) ( _ )

(1 ( _ | )) ( )( _ _ )(1 ) ( _ )

( _ )(1 ( _ |

dP Antarctic bird Bird P Bird

P Antarctic bird Flying bird P Flying birdP Antarctic bird Bird P Bird

P Antarctic bird Flying bird P Flying birdP Flying birdP Antarctic bird B

)) ( )( _ ) ( _ _ )

(1 ( _ | )) ( )( _ ) ( _ | _ ) ( _ )

(1 ( _ | )) ( )0.95 0.2*0

ird P BirdP Flying bird P Antarctic bird Flying bird

P Antarctic bird Bird P BirdP Flying bird P Flying bird Antarctic bird P Antarctic bird

P Antarctic bird Bird P Bird

.01 0.958

1 0.01

Page 43: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

43

Probabilistic reasoning in terminological

logics(Jaeger,1994) Reasoning on KB:

According to Jeffrey’ rule,

Present a naive method for computing the probability of new knowledge

( _ )( _ )* ( _ | _ )( _ )* ( _ | _ )

0.9*0.2 0.1*0.9580.2758

P Opus Flying birdP Opus Antarctic bird u Flying bird Antarctic birdP Opus Bird Antarctic bird u Flying bird Bird Antarctic bird

Page 44: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

44

Research Challenges in DL Extensions

Syntax and Semantics Decidability Reasoning algorithms for

possible extensions Soundness and completeness Complexity/efficiency Effective methods for

reasoning under uncertainty

44

Page 45: Comparison and Combination of the Expressive Power of Description Logics and Logic Programs

45

Questions?