KWTR: ontology maintenance

[edit] Contributors:

Vit Novacek DERI Galway

• What is the state of the art of Semantic Web in your research field?

• • Maintenance and versioning of the Semantic Web data and knowledge is related to ontology acquisition, maintenance, interchange and reuse in dynamic

environments. Typical examples can be:

• • • a hospital, where new patient, drug-related and other similar data have to be processed very often on a daily basis;
    • a research institution, where the knowledge essentially grows and changes over time;
    • HR field, in which efficient means for processing dynamic job-market trends, as well as new job or application-related data, are needed.

• • In such settings, new knowledge occurs continuously and needs to be processed - included into the current knowledge base (i.e. an ontology and/or an instance base). However, just including is not enough - we also have to discover the consequences of changes in our knowledge base and deal with them properly. We should also keep track of different versions of an ontology in time. These issues are studied and respective techniques are being developed within the ontology evolution, maintenance and versioning topic.

• • Ontology evolution process can be described in six dimensions, referring to continuous management of changes in dynamic environments:
    • change capturing - detection and introduction of needed changes;
    • change representation - encoding/classifying changes;
    • semantics of change - the meaning of change, definition of inconsistency;
    • change implementation - change application;
    • change propagation - application of change consequences in the ontology;
    • change validation - assessing the effects of changes, satisfying and/or dealing with the consistency conditions.

• Provide references and short abstracts of three papers you consider as significant in your research field.

1. N. F. Noy and M. Klein. Ontology evolution: Not the same as schema evolution. Knowledge and Information Systems, 5, 2003. CiteSeer
2. Giorgos Flouris, Dimitris Plexousakis, Grigoris Antoniou. Evolving Ontology Evolution. In LNCS 3831, Springer 2006, 14-29. PDF
3. Michel Klein and Natalya F. Noy. A component-based framework for ontology evolution. Technical Report IR-504, Department of Computer Science, Vrije Universiteit Amsterdam, March 2003. CiteSeer

• Please provide one or more examples (either business, or research, or both) in which semantic web has been used (if you can, add some references).
  • The [MarcOnt project] uses SemVersion (syntactic and partially also semantic versioning for ontology evolution).
  • The KAON link research framework and LinkFactory link industrial toolbox implement transaction-based versioning in order to support ontology evolution.
  • The PROMPT and PROMPTDiff link research prototypes implement the merge-based versioning in the context of ontology evolution.
  • The DINO framework (being developed within the Knowledge Web project) uses Semantic Web solutions in order to support dynamic ontology lifecycle and semi-automatic ontology integration in dynamic environments.

• Are there existing tools or demos? Please indicate some of them.
  • See above for the respective tools...

• What are the open problems in your Semantic Web research field? Why?

• • theoretical foundations:
    • proper formalisation of ontology change (e.g. adaptation of the belief revision theory)
    • logical consequences of application of change operators (addition, retraction)
    • relation between syntactic and semantic changes
    • study of inconsistency, its definition, consequences and treatment changing ontologies
  • practical implementations:
    • implementation of change management - mostly using versioning of ontologies (analogical to software version management)
    • diff computation and representation
    • development and implementation of methodologies explicitly and efficiently supporting dynamic in the ontology lifecycle
    • means for efficient knowledge acquisition in dynamic environments - for instance by automatic ontology learning from natural language text

• Provide references and links of the most relevant Semantic Web research projects in your field.

1. SEKT
2. Knowledge Web
3. NEON

• What challenges try these projects to overcome?

• • Methodology of versioning, clasification of basic approaches - formal background
  • Methodology and implementation of dynamic ontology lifecycle and the respective ontology maintenance prototype - incorporation of versioning and (learned) ontology integration into the lifecycle tailored for application to data-intensive and dynamic scenarios
  • Dynamics of networked (distributed) ontologies.

• What are their foreseen benefits (both in market and scientific community)?

• • more efficient knowledge management supported by:

1. well-founded and "safe" methodologies and mechanisms of change implementation in a knowledge base
2. maintenance of ontology versions in time
3. visualisation of the relations between several versions of an ontology

• • facilitation of knowledge acquisition in dynamic environments by:

1. semi-automatic methods of ontology acquisition
2. dynamic machine-assisted knowledge integration techniques
3. methods of consistency checking and automatic resolution/identification of possible inconsistencies

• When, in your opinion, will projects’ results be ready for industry?

2-5 years

• Do you think that it is important to invest (money and time) in these topics? Why?

• • More than important - means for efficient and reasonable change maintenance are crucial for application of the cutting-edge knowledge engineering technologies in practical scenarios, that are very often inherently dynamic. There are many open and practically important research questions, however, even some low-hanging fruit sub-topics, which can be very useful in industrial use - for instance semi-automatic support for integration of ontologies learned from natural language.

• What are, in your opinion, the most relevant Semantic Web challenges that will be solved in the long term (10 years)? Why?

• • dynamic ontology integration and population (needed to support day-to-day dynamics in realistic application scenarios)
  • combination of version management with inference and safe change implementation and propagation in ontologies (the changes are not only managed, but also assessed in the context of the current content of the versioned ontology, which is needed for improved usability of the changing ontologies)