Presentation is loading. Please wait.

Presentation is loading. Please wait.

On Relationships among Models, Meta Models and Ontologies Motoshi Saeki Tokyo Institute of Technology Haruhiko Kaiya Shinshu University

Similar presentations


Presentation on theme: "On Relationships among Models, Meta Models and Ontologies Motoshi Saeki Tokyo Institute of Technology Haruhiko Kaiya Shinshu University"— Presentation transcript:

1 On Relationships among Models, Meta Models and Ontologies Motoshi Saeki Tokyo Institute of Technology Haruhiko Kaiya Shinshu University DSM'06@OOPSLA2006, Oct. 22, 2006, Portland

2 Motivation & Approach Meta Model: Abstract Syntax for DSL Semantics of Models and Meta Models? Semantic Consistency & Completeness Using Ontology: Thesaurus + Inference Rules Semantic Concept: Common Words where everybody can have the same and unique interpretation in a domain Structure (Concepts and Relationships) : Thesaurus mechanisms to infer/calculate the semantic properties on the ontology

3 Domain Ontology: Example Specifying semantics of a specific area, e.g. Lift Control System, Insurance Business, etc. Using Class Diagram Word = Class Relationship = Association

4 Basic Idea CD A E B cause Model Domain Ontology (thesaurus part only) aaa bbb semantic mapping

5 Basic Idea CD A E B cause Model Domain Ontology (thesaurus part only) aaa bbb semantic mapping "A causes C" on the Domain Ontology

6 Basic Idea CD A E B cause Model Domain Ontology (thesaurus part only) aaa ccc bbb semantic mapping "aaa causes ccc" for semantic consistency Adding ccc Finding Semantic Inconsistency and Incompleteness Inference on the structure of the ontology

7 Model v.s. Domain Ontology LiftDoor Scheduler 1: request 2: up 3: arrived 4: open Model

8 Model v.s. Domain Ontology Door > Lift > Doors > Lifts > Close > Move > Stop > Open > next LiftDoor Scheduler 1: request 2: up 3: arrived 4: open Model Domain Ontology next : Temporal order of execution

9 Model v.s. Domain Ontology Door > Lift > Doors > Lifts > Close > Move > Stop > Open > next LiftDoor Scheduler 1: request 2: up 3: arrived 4: open Model Domain Ontology semantic mapping

10 Model v.s. Domain Ontology Door > Lift > Doors > Lifts > Close > Move > Stop > Open > next LiftDoor Scheduler 1: request 2: up 3: arrived 4: open Model Domain Ontology semantic mapping ? Inference of causality Model: up causes arrived causes open Ontology: Move next Stop next Open Stop missing!

11 Meta Model v.s. Meta Model Ontology Object Message sendreceive Data carry Meta Model of Sequence Diagram

12 Meta Model v.s. Meta Model Ontology manipulate Object Message sendreceive AssociationClass source destination FunctionData consume produce State describe Object abstraction describe Event associate change-from change-to participate next Data carry Meta Model of Sequence Diagram Meta Model Ontology

13 Meta Model v.s. Meta Model Ontology manipulate Object Message sendreceive AssociationClass source destination FunctionData consume produce State describe Object abstraction describe Event associate change-from change-to participate next Data carry Meta Model of Sequence Diagram Meta Model Ontology semantic mapping

14 Meta Model v.s. Meta Model Ontology manipulate Object Message sendreceive AssociationClass source destination FunctionData consume produce State describe Object abstraction describe Event associate change-from change-to participate next Data carry Meta Model of Sequence Diagram Meta Model Ontology semantic mapping ? Meta Model: Message carry Data, Message ? Message Ontology: Event associate Data, Event next Event Next missing

15 Meta Model v.s. Meta Model Ontology manipulate Object Message sendreceive cause AssociationClass source destination FunctionData consume produce State describe Object abstraction describe Event associate change-from change-to participate next Data carry Meta Model of Sequence Diagram Meta Model Ontology Adding "cause" Semantic Quality = 6/7

16 Syntactic Domain Semantic Domain semantic mapping instantiation Meta Model Model Domain Ontology Meta Model Ontology Inference Rule#2 Inference Rule#1 M1 Layer (in MOF) M2 Layer (in MOF) Relationships

17 Summary & Future Work Providing Semantics for Models and Meta Models using Ontologies How faithfully a model and a meta model reflect their ontologies? Deviations from Ontologies represent semantic inconsistency and incompleteness Future Application Semantic Quality Measurement for Models and Meta Models (DSL) Elaborating a Meta Model Ontology Reasoning Properties on Models and Meta Models Automated Tool for Reasoning, Calculating,...


Download ppt "On Relationships among Models, Meta Models and Ontologies Motoshi Saeki Tokyo Institute of Technology Haruhiko Kaiya Shinshu University"

Similar presentations


Ads by Google