Ontology Transformations Laurent WOUTERS (EADS Innovation Works, France) Marie-Pierre GERVAIS (Université Paris Ouest, LIP6, France)
Motivation: Example Operating a safety-critical system Ontology Transformations 2 EDOC 2012 Activate fuel jettison Check gears are up Flaps to MAX Pitch to 8° Aircraft ditching procedure: Procedure Stress, fatigue, … System Operator
Motivation: Holistic Model-Based Approach to Testing Ontology Transformations 3 EDOC 2012 Execute Results scenario modifications Model Procedure Stress, fatigue, … System Operator
Motivation: Multiple Domain Experts Ontology Transformations 4 EDOC 2012 Model System Engineers Interaction Experts Cognitive Psychologists Procedure Stress, fatigue, … System Operator
Motivation: Multi-View Visual Modeling Ontology Transformations 5 EDOC 2012 System Engineers Interaction Experts Cognitive Psychologists Modeling Environment for Cognitive Psychologists Modeling Environment for Interaction Experts Modeling Environment for System Engineers Domain-Specific Visual Sentences Common Model Artifact xOWL [1] Transformations OWL [1] xOWL: an Executable Modeling Language for Domain Experts, EDOC 2011
State of the Art: Model Transformations Ontology Transformations 6 EDOC 2012 Input common ontologyOutput visual sentences τ ontology to modelmodel to ontology OWL2 World MOF World Translated input modelVisual sentences model Query/View/Transform [1] (SmartQVT, mediniQVT, ModelMorf) ATLAS Transformation Language [2] Triple Graph Grammars [3] [1] OMG, Meta Object Facility Query/View/Transformation version1.1, 2011 [2] Jouaultand, Kurtev, Transforming Models with ATL MoDELS 2006 [3] Greenyer, Kindler, Comparing Relational Model Transformation Technologies, SoSyM 2010 [4] Silva Parreiras, Staab, Using Ontologies with UML Class-Based Modeling: The Two Use Approach Data & Knowledge Engineering 2010 [5] Djuric, Gasevic, Devedzic, Ontology Modeling and MDA, Journal of Object Technology 2005 Cannot map the whole semantic of OWL [4,5] ODM
State of the Art: Ontology Transformations Ontology Transformations 7 EDOC 2012 [6] W3C, SWRL: A Semantic Web Rule Language Combining OWL and RuleML, 2010 [7] Horrockse et al., OWL Rules: a Proposal and Prototype Implementation, Web Semantics: Science, Services and Agents on the World Wide Web 2005 Cannot operate over classes and relations [7] Semantic Web Rule Language [6] Input common ontologyOutput visual sentences τ’τ’ OWL2 World MOF World
xOWL Rule Language Rule(:CMAttachSubTree_Activity_route13 Antecedents( ClassAssertion(command:Attach ?com) ObjectPropertyAssertion(command:symbol ?com view:Activity) ObjectPropertyAssertion(command:parent ?com ?np) ObjectPropertyAssertion(command:child ?com ?nc) Meta(ObjectPropertyAssertion(view:route13 ?nr ?np)) Meta(ObjectPropertyAssertion(meta:trace ?nr ?or)) Meta(ObjectPropertyAssertion(meta:trace ?nc ?oc)) ) Consequents( ClassAssertion(?oc ?or) ) Ontology Transformations 8 EDOC 2012 OWL2 Axioms Logic Variables
xOWL Transformations A transformation = set of independent xOWL rules (no prioritization) Positive consequents are added to the target Negative consequents are removed from the target A “Meta” ontology is used to store traceability information “Meta” antecedents are matched in the meta ontology “Meta” consequents are added or removed from it Ontology Transformations 9 EDOC 2012 Input ontology Target ontology Meta ontology τ
Validation 3 Steps: Implementation Demonstration on the use case Performance study Implementation: Incremental transformation engine The RETE pattern-matching algorithm is used for matching rules’ antecedents Available under the LGPL license at Ontology Transformations 10 EDOC 2012
System Engineers Interaction Experts Cognitive Psychologists Validation: Application to the Use Case (1) Ontology Transformations 11 EDOC 2012
Validation: Application to the Use Case (2) Ontology Transformations 12 EDOC 2012 component instance-of Common Model Artifact
Validation: Application to the Use Case (2) Ontology Transformations 13 EDOC 2012
Validation: Performance Study Objective: Ensure that ontology transformations have sufficient performances for live incremental transformations Tested the transformations from the use case with ontologies of increasing sizes Correlation is 0.99Correlation between 0.90 and 0.99 Less than 1.5sLess than 10ms Ontology Transformations 14 EDOC 2012
Conclusion Ontology Transformations 15 EDOC 2012 Express ontology transformations with the xOWL Rule Language Execute live incremental ontology transformations Applied to the use case: Supports multiple domain-specific perspectives on a common model artifact Improves the safety of critical systems Perspectives: More expressive rule language with explicit rules prioritization for example. Support the software engineers that have to write the transformations with visual notations for rules.