2nd Review OnToKnowledge

1

OntoEdit
Ontology Engineering Environment

• 2nd Review On-To-Knowledge
• Amsterdam, 2001-08-14/15
• York Sure, AIFB
• Rudi Studer, AIFB

On-To-Knowledge
2
(No Transcript)
3
WP-1.2, Deliverable 3

• OntoEdit is a graphical Ontology Engineering
  Environment
• Helps creating, modifying and browsing of
  ontologies
• Supports the whole ontology development process
  from kick-off to maintenance phase

4
Modeling with OntoEdit

• Modeling at a conceptual level
• As much as possible independent of a concrete
  representation language
• Using GUIs as representing views on conceptual
  structures
• Concepts
• Concept hierarchy
• Relations
• Instances

5
Scalability

• Ontologies are expected to be O(103) classes
• OntoEdit comfortably deals with ontologies up to
  that range

6
Interoperability

• OntoEdit uses homemade OXML as regular data
  storage format
• Various export filters are available, including
  OIL-Core and DAMLOIL (March 2001)
• Direct import and export of ontologies from/to
  Sesame

7
Extensibility

• Flexible and expandable plug-in framework
• Implement functions by extensions
• Partners may expand and add functionalities

8
Screenshot
9

Methodology
Employed and Evaluated Version

• 2nd Review On-To-Knowledge
• Amsterdam, 2001-08-14/15
• York Sure, AIFB
• Rudi Studer, AIFB

On-To-Knowledge
10
WP-5 Workpackage project plan
Employment Evaluation
Expansion Revision
Final version
Employment Evaluation
Expansion Revision
Final version

• Baseline version

Baseline version

• Based on CommonKADS
• Literature study
• Systematization of existing practical experiences

• Employment of baseline version in the set of case
  studies
• Feedback on KM guidelines
• Plus Reuse of ontologies

• Revision and Expansion based on feedback

• General and comprehensive KM guidelines
• Practically illustrated by real-life industry
  cases
• Guide book

11
SwissLife Skills Management

• Top-down ontology build from scratch
• High quality engineered ontology
• Ontology workshop
• Training of three domain experts as ontology
  engineers
• Usage of a standard tool for draft taxonomy

12
SwissLife Skills Management

• Major challenges
• Handle quick growth of ontology
• Avoid over-motivation
• Manage cooperative ontology engineering
• Train domain experts

13
SwissLife Access to IAS

• Bottom-up semi-automatic approach to build
  ontology
• Use NLP IR-techniques to gather weak ontology
• Major challenges
• Finding appropriate clusters
• Lack of time requires fully automatic approach
  for ontology development

14
British Telecom Call Center Support

• Feasibility study
• Target area / actors / stakeholders identified
  (users providers)
• Major challenges
• Maintenance of ontologies
• Knowledge ages quickly
• Update/insert/delete has to be simple and
  effective
• Time-critical environment

15
EnerSearch Virtual Organization

• Lessons learned from initial phase
• Manual lightweight ontology is built quickly,
  but already shows results
• Semi-automatic approach needed for further
  expansion to a heavyweight ontology

16
Next Steps

• Gradual expansion and revision in ongoing case
  studies
• Feedback is crucial
• Push additionally to pull

• Integrate semi-automatic and automatic ontology
  extraction related topics
• Handling of maintenance process