State of Semantic Web: Ontology Aspects Rudi Studer FB AI

Download Report

Transcript State of Semantic Web: Ontology Aspects Rudi Studer FB AI

State of Semantic Web: Ontology Aspects
Institute AIFB, University of Karlsruhe
http://www.aifb.uni-karlsruhe.de/WBS
AIFB
Rudi Studer & Raphael Volz
FZI Research Center for Information
Technologies, KM Department (WIM)
http://www.fzi.de/wim
L3S Learning Lab, Hannover/Karlsruhe
http://www.learninglab.de
Ontoprise GmbH, Karlsruhe
http://www.ontoprise.de
DAML PI Meeting, Portland, October 2002
Developing the Semantic Web...

Research Aspects
– Application-oriented vs. basic research
• Objectives and challenges
• Funding strategies
– DARPA, EU, ...

vs.
User & Industrial Application Impact
– Product development
– User take-up
User & Industrial Application Impact in the
Near Future

A lot of real-life applications need a flexible way of
– starting with a small amount of semantics
– enhancing the application gradually with more
and more semantics
Keep entry barrier as low as possible!
User & Industrial Application Impact in the
Near Future : Ontology Management

Usability/Usage
– Connect to well-established modeling paradigm
• exploit industry know-how, e.g. UML, EER
• Keep initial learning effort small
– Light-weight ontologies pay off in a lot of applications
– see e.g. our On-To-Knowledge project experience
• Skill management (Swiss Life, Switzerland)
• Project management (British Telecom, UK)
– Different application domains raise different requirements
User & Industrial Application Impact in the
Near Future : Ontology Management

Implementation Aspects
– Exploit well-known implementation techniques
• to which extent do they meet the requirements
that are set up by OWL / OWL Lite?
– Rely on scalable methods
• what can we take/learn from the DB community?
User & Industrial Application Impact
usage/
usability
EER HTML
KIF
SGML
impl. aspects
User & Industrial Application Impact
usage/
usability
OWL Lite1
OWL Lite2
OWL Lite
OWL
OWL
impl. aspects
Research Topics from an Industry Application
Perspective

Ontology and Metadata Management
– Evolution / Versioning
– Learning / Metadata generation
• Support incremental approaches
– Engineering multiple ontologies / Mapping
– Personalization / Views
First methods and tools are available, but a lot of aspects
have to be clarified for industry applications
Research Topics from an Industry Application
Perspective

Scalability Issues
– Handling hundreds of interconnected ontologies
– Efficient handling of large-scale ontologies
– Metadata repositories
– Appropriate transaction mechanisms
Keep notions simple enough to be able to provide
applicable solutions!
Thank you!
Institute AIFB, University of Karlsruhe
http://www.aifb.uni-karlsruhe.de/WBS
FZI Research Center for Information
Technologies, KM Department (WIM)
http://www.fzi.de/wim
L3S Learning Lab, Hannover/Karlsruhe
http://www.learninglab.de
Ontoprise GmbH, Karlsruhe
http://www.ontoprise.de
AIFB
Rudi Studer & Raphael Volz