evolution in owl 2 ql & owl 2 el ontologies
DESCRIPTION
Evolution in OWL 2 QL & OWL 2 EL Ontologies. Dmitriy Zheleznyakov 28 th of January, 2014, Oslo. Ontology. General rules:. To use ontologies in applications, we need special, formal syntax. . All popes are clerics. Facts:. Benedict XVI is a pope. Explicit knowledge. Ontology. - PowerPoint PPT PresentationTRANSCRIPT
Evolution in OWL 2 QL & OWL 2 EL Ontologies
Dmitriy Zheleznyakov
28th of January, 2014, Oslo
2
Ontology
All popes are clerics
Benedict XVI is a pope
General rules:
Facts:
o To use ontologies in applications,we need special, formal syntax
2
Ontology
o Do ontologies differ from data bases?o Data bases: explicit knowledge onlyo Benedict XVI is a pope
o Ontologies: explicit & implicit knowledgeo Benedict XVI is a popeo Reasoning: Benedict XVI is a cleric
reasoning
Explicit knowledgeImplicit knowledge
o The focus of this work: ontology languages for the Semantic Web
o Web Ontology Language: OWL 2 (W3C Standard)o OWL 2 QLo OWL 2 EL
o Good computational propertieso Efficient schema and data management
o Used in practice
3
Ontology Languages
o Ontology-Based Data Access (OBDA)o provide unified query interface
to heterogeneous data sources
4
OWL 2 QL: Ontology-Based Data Access
o Ontology-Based Data Access (OBDA)o provide unified query interface
to heterogeneous data sourceso EU FP7 project Optique will develop
an OBDA systemo use-case partners: Statoil, Siemens
o Ontologies may change:o new knowledge about domaino new data source is added
o Motivation for our work:o to address the dynamicity of OBDA
systemsby studying evolution of schema and data
4
OWL 2 QL: Ontology-Based Data Access
5
OWL 2 EL: Clinical Science, Bio Ontologieso Ontologies enable communication and
knowledge sharingbetween doctors, scientists, etc.
o SNOMED CT: > 311k termso constantly under development:o 5 modification teams o every 2 weeks
the main team integrates changes, o 2002 2008
SNOMED went 278k 311k termso It is the standard to describe the results of
experiments in the US clinical labs
o Motivation for our work:o to provide techniques that facilitate
ontology development for such a vast community
o To facilitate evolution of ontology-based systemso insertion of knowledgeo deletion of knowledge
o On two levels: o schemao data
o With as little changesas possible
6
Our Goal
Original ontology
To insert
To delete
7
How to Approach the Problem?Original ontology New knowledge
Resulting ontology
1. Define an operator and understand it• a conceptual understanding
of how to evolve ontologies• checking its computational
properties2. Develop an algorithm to
compute the result3. Implement the algorithm
8
Previous Work
AI: 80’s – 90’sPropositional logic, weaker then OWL 2 QL & OWL 2 EL
KR: 2004-2006 2007-2010
[AGM’85][Borgida’85][Dalal’88][Satoh’88][Winslett’90]
Many evolution operators proposed
[Winslett’88][Katsuno&Mendelzon’91]
Model-based operators
Formula-based operators
[Kang&Lau’04][Flouris&al’04][Flouris&al’05][Qi&al’06]
[Liu& al’06][Qi&Du’09][DeGiacomo&al’07-09][Wang&al’10]
Adaptationof some operators
9
Model-based operators
Formula-based operators
General Overview of the Results
OWL 2 QLOWL 2 EL
1 2 3
4 5 6
Work for restriction of OWL 2 QL
Propositionallogic
OWL 2 QL
OWL 2 EL
For OWL 2 QL & EL- inexpressibility- counterintuitiv
e results
- inexpressibility- counterintuitiv
e resultsWorks for OWL 2 QL
Bold operator
Works for OWL 2 QL &
EL
Tunableoperator
Works for OWL 2 QL
Bold operator
Works for OWL 2 QL &
EL
Tunableoperator
Model-based operators
Formula-based operators
Understanding Model-Based Operators
2 3
4 5 6
Work for restriction of OWL 2 QL
Propositionallogic
1
10
OWL 2 QL
OWL 2 EL
11
Understanding Model-Based Operators
o We have shown: operators are determined by three parameterso this gives a three-dimensional space
of operators
o Classical operators fit in this spaceo Novel operators can be easily defined
by changing parameters
11
Understanding Model-Based Operators
o We noticed: operators are determined by three parameterso this gives a three-dimensional space
of operators
o Classical operators fit in this spaceo Novel operators can be easily defined
by changing parameterso We can add new values to dimensions!
o more operators can be defined!
Works for OWL 2 QL
Bold operator
Works for OWL 2 QL &
EL
Tunableoperator
1Model-based operators
Formula-based operators
Inexpressibility of Model-Based Operators
3
4 5 6
Work for restriction of OWL 2 QL
Propositionallogic
12
2
- inexpressibility- counterintuitiv
e results
OWL 2 QL
OWL 2 EL
13
Inexpressibility of Model-Based Operators
Schema:Wives are married to their husbandsPriest cannot be husbands
Facts:Mary is married to John andAdam and Bob are priests
13
Inexpressibility of Model-Based Operators
PriestAdamBob
Mary JohnhasHusband
Facts to add:John is a priest
Under model-based operators:We incorporate new knowledge directly into models
a model:
13
Inexpressibility of Model-Based Operators
PriestAdamBob
Mary JohnhasHusband
John cannot be a husband of Mary anymore!What happens to her?Three options:1. She divorced2. She married some one else3. She married to a former
priest
PriestAdamBobJohn
hasHusband1.
PriestAdamBobJohn
Mary JackhasHusband2.
PriestAdamJohn
Mary BobhasHusband3.
13
Inexpressibility of Model-Based Operators
We showed:all these options cannot be capturedin OWL 2 QL and OWL 2 EL
We need at least disjunction which is not in OWL 2 QL and OWL 2 EL
PriestAdamBobJohn
hasHusband1.
PriestAdamBobJohn
Mary JackhasHusband2.
PriestAdamJohn
Mary BobhasHusband3.
OR
OR
PriestAdamBob
Mary JohnhasHusband
Works for OWL 2 QL
Bold operator
Works for OWL 2 QL &
EL
Tunableoperator
1Model-based operators
Formula-based operators
Bad Behaviour of Model-Based Operators
3
4 5 6
Work for restriction of OWL 2 QL
Propositionallogic
OWL 2 QL
OWL 2 EL 14
2
- inexpressibility- counterintuitiv
e results
15
Bad Behaviour of Model-Based Operators
Facts:Adam and Bob are priests
Facts to add:John is a priest
No schema
Some of model-based operators behave as follows:
15
Bad Behaviour of Model-Based Operators
PriestAdamBob
PriestAdamBobJohn
Expectedresult:
PriestJohn
Actualresult:
Such behaviour is not usefulfor any application
Some of model-based operators behave as follows:
Works for OWL 2 QL
Bold operator
Works for OWL 2 QL &
EL
Tunableoperator
21Model-based operators
Formula-based operators
Restriction of OWL 2 QL
4 5 6Propositional
logicOWL 2 QL
OWL 2 EL 16
3
Work for restriction of OWL 2 QL
17
Restriction of OWL 2 QL
o We found the reason of the bad behaviour of model-base operators:
A binary relation participates in disjointness
o Priest cannot be husbandso What if we forbid
this bad interaction?o We showed:
most of model-based operators work!o this fragment captures (FO part of)
RDFS(another W3C standard)
PriestAdamBob
Mary JohnhasHusband
disjoint with
18
Summing up on Model-Based Operators
o Model-based operatorso suffer from inexpressibilityo tend to lose too much of informationo counterintuitive behaviour
o Our verdict: o model-based operators are not suitable
for the case of OWL 2 QL or OWL 2 EL
o We turned to Formula-based operators!
Works for OWL 2 QL
Bold operator
Works for OWL 2 QL &
EL
Tunableoperator
21 3Model-based operators
Formula-Based Operators
5 6Propositional
logicOWL 2 QL
OWL 2 EL 19
Formula-based operators 4
Work for restriction of OWL 2 QL
- inexpressibility- counterintuitiv
e results
20
Formula-Based Operators
Explicit schema
Implicitschema
o Preserve all the knowledge:both explicit and implicit
o Example: delete Priests are Maleso We do not want to lose info that
Adam is Male
Explicit data
Implicitdata
20
Formula-Based Operators
Explicit schema
Implicitschema
o Preserve all the knowledge:both explicit and implicit
o Example: delete Priests are Maleso How to delete it in such a way
that it will not appear even implicitly?o Deleteo either Priests are Clericso or Clerics are Males
20
Formula-Based Operatorso Preserve all the knowledge:
both explicit and implicito Example: delete Priests are
Maleso How to delete it in such a way
that it will not appear even implicitly?o Deleteo either Priests are Clericso or Clerics are Males
o What to do with a multiple choice?Classical approaches:o Keeping both – impossibleo Combining themo too much of information is losto we proved: it is computationally hard
The resulted schema:either or
Works for OWL 2 QL &
EL
Tunableoperator
4Formula-based operators
21 3Model-based operators
Bold Operator
6Propositional
logicOWL 2 QL
OWL 2 EL 21
Work for restriction of OWL 2 QL
5Works for OWL 2 QL
Bold operator
22
Bold Operator
o Example: delete Priests are Maleso How to delete it in such a way
that it will not appear even implicitly?o Deleteo either Priests are Clericso or Clerics are Males
o What to do with a multiple choice?o We propose: Bold operator.
It picks up one of themo The result is non-deterministic…o … But can be computed
in polynomial time (for OWL 2 QL)
o In the case of OWL 2 EL:o Implicit knowledge can be infiniteo Bold operator does not work
The resulted schema:either or
There is no way to decide which result is better!This is application dependent and should be up to the user
54Formula-based operators
21 3Model-based operators
Tunable Operator
Propositionallogic
OWL 2 QL
OWL 2 EL 23
Work for restriction of OWL 2 QL
6Works for OWL 2 QL
Bold operator
Works for OWL 2 QL &
EL
Tunableoperator
24
Tunable Operator
Explicit schema
Implicitschema
o Tunable operatoro allows to choose
what part of implicit knowledge will be preserved
24
Tunable Operator
o Tunable operatoro allows to choose
what part of implicit knowledge will be preserved
Explicit schema
Implicitschema
25
Tunable Operator
No implicit part
Whole implicit part
26
Summing up on Formula-Based Operators
o Classical Formula-based operatorso suffer from inexpressibilityo tend to lose too much of information
o Bold operator: o works for OWL 2 QLo fails for OWL 2 EL
o Tunable operator:o works for both OWL 2 QL and OWL 2 EL
27
Current Work
o Applying our results to the Optique projecto in progress
o Incorporating evolution in transition systemso IJCAI’2013
o Information hiding & Controlled query evaluationo ISWC 2013o submitted to an international conference