Transcript
Page 1: Automating Instance Migration in Response to Ontology Evolution

Automating Instance Migration in Response to Ontology Evolution

Mark Fischer – Queen’s Juergen Dingel – Queen’s Maged Elaasar – Carleton Steven Shaw – IBM

Page 2: Automating Instance Migration in Response to Ontology Evolution

2

Agenda Overview Approach

Comparing Two Ontologies Creating a Transformation

Oital Analyzing a Transformation

Case Study Future Work Conclusion

MODELS 2013 Workshop

Page 3: Automating Instance Migration in Response to Ontology Evolution

3

Overview Migration

Move individuals from one ontology to another.

Motivation This setup reflects the

way IBM’s Design Management tool stores models as Ontologies.

MODELS 2013 Workshop

Page 4: Automating Instance Migration in Response to Ontology Evolution

4

Overview Developed:

Automated:

MODELS 2013 Workshop

Page 5: Automating Instance Migration in Response to Ontology Evolution

5

Approach We let the Migration be performed via a

Transformation

Creating this transformation is hard. Add steps to make it easier

What would help? Some way of comparing two ontologies An easy way to write a transformation Ways to test/analyze transformations for correctness

MODELS 2013 Workshop

Page 6: Automating Instance Migration in Response to Ontology Evolution

6

Comparing Two Ontologies There are many competing ways to compare ontologies

For creating these sorts of transformations, only those parts of an ontology that may effect individuals are of any interest. We are interested in Axioms For any axiom, C, the axiom and all other axioms it is

influenced by is called the context of C

MODELS 2013 Workshop

Page 7: Automating Instance Migration in Response to Ontology Evolution

7

Comparing Two Ontologies

MODELS 2013 Workshop

Page 8: Automating Instance Migration in Response to Ontology Evolution

8

Comparing Two Ontologies: Original

MODELS 2013 Workshop

Page 9: Automating Instance Migration in Response to Ontology Evolution

9

Comparing Two Ontologies: Updated

MODELS 2013 Workshop

Page 10: Automating Instance Migration in Response to Ontology Evolution

10

Creating a Transformation Use domain-specific language

We created Oital

About Oital Syntax based off of the Manchester Owl syntax Becomes a form of documentation Has an integrated development environment called

Oital-T

MODELS 2013 Workshop

Page 11: Automating Instance Migration in Response to Ontology Evolution

11

Oital An Oital transformation

consists of: Actions which delete or create

individuals and their properties TransformationClasses which

define a category of individual based off of a query

The order of actions does matter

TransformationClasses change depending on their context

MODELS 2013 Workshop

Page 12: Automating Instance Migration in Response to Ontology Evolution

12

Analyzing a Transformation We currently support a form of Abstract

Interpretation How does it help?

Lets you isolate specific properties of the input and output of a transformation

Example: Abstract Interpretation of Class Membership can answer the following questions

Does every individual which is a member of a removed class get migrated so that it is a member of an existing class?

Which classes are guaranteed to have no individuals? Are individuals being migrated into more restrictive classes?

MODELS 2013 Workshop

Page 13: Automating Instance Migration in Response to Ontology Evolution

13

Case Study Use IBM’s Ontology encoding of UML 2.1.1, UML 2.2, and UML

2.4.1 to recreate their migration using this approach. UML 2.4.1 has:

255 Named Classes 801 Anonymous Classes (enumerated, union, complement,

intersection, restriction) 594 properties

Comparing UML 2.1.1 and UML 2.2: # of must investigate axioms: 38 # of should investigate axioms: 118 # of ok axioms: 4361

MODELS 2013 Workshop

Page 14: Automating Instance Migration in Response to Ontology Evolution

14

When to use this approach It is often faster to migrate manually

Transformation are general and can make no assumptions about any specific set of individuals

When does this approach make most sense? When ontology developers and users are different people. When there are many users (applications) using the evolving

ontology When there is no way of predicting how an ontology will be

used

MODELS 2013 Workshop

Page 15: Automating Instance Migration in Response to Ontology Evolution

15

Future Work More analysis!

Abstract interpretation isn’t the only helpful form of analysis possible.

Continue development on Oital-T Discover usage patterns for Oital

Integrate them into the language or tool to insure ease of use.

Case study. Continue with IBM UML case study

MODELS 2013 Workshop

Page 16: Automating Instance Migration in Response to Ontology Evolution

16

Conclusion Of great importance to the efficient use of an

ontology is the ability to easily effect change.

The approach described here facilitates a way of keeping certain types of ontological artifacts up to date in a way that is, potentially, very scalable.

MODELS 2013 Workshop

Page 17: Automating Instance Migration in Response to Ontology Evolution

17

References Natalya F. Noy and Michel Klein. Ontology evolution: Not the same as

schema evolution. Knowledge and Information Systems, 6(4):428–440. Peter Plessers, Olga De Troyer, and Sven Casteleyn. Understanding

ontology evolution: A change detection approach. Web Semantics: Science, Services and Agents on the World Wide Web, 5(1):39–49, 2007.

Asad Masood Khattak, Zeeshan Pervez, Sungyoung Lee, and Young-Koo Lee. After effects of ontology evolution. 5th International Conference on Future Information Technology. IEEE, 2010.

Matthew Horridge and Peter F. Patel-Schneider. OWL 2 Web Ontology Language Manchester Syntax. W3C Working Group Note. Dec 11, 2012.

Sean Bechhofer, Frank van Harmelen, Jim Hendler, Ian Horrocks, Deborah L. McGuinness, Peter F. Patel-Schneider, and Lynn Andrea Stein. Owl web ontology language reference. February 2004.

MODELS 2013 Workshop


Top Related