Actions

Ontolog Forum

Revision as of 05:53, 26 April 2013 by imported>PeterYim (Last updated at: 2012-12-17 12:17:53 By user: PeterYim)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Upper Ontology Summit (UOS) - Home Page

To be held as one of a series of meetings at NIST during Interoperability Week

Public Upper Ontology Summit meeting - Wednesday March 15, 2006 1-5 PM EST - details at: Agenda & Proceedings page - UOS Joint Communiqué


Deliverables:

UOS Event Date(s) & Venue

  • The Upper Ontology Summit Event: Wednesday, March 15, 2006, 1:00~5:00pm
  • UOS Preparation Meetings: Tue 14-Mar-2006 (all day) & Wed 15-Mar-2006 (morning)
  • Venue: US National Institute of Standards and Technology (NIST) Campus at Gaithersburg, Maryland, USA.

Project Mission

Develop the mechanism and resources to relate existing upper ontologies to each other in a manner that will increase reuse of knowledge among them, and thereby facilitate semantic interoperability among other ontologies that are linked to them. We want to make the world aware that the technology of upper ontologies has developed to a point suitable for commercial exploitation.

  • The purpose of this panel is to bring together those who recognize the value of open upper ontologies, and in particular, the custodians (developers or maintainers) of the public versions of existing upper ontologies to find a way to interrelate those ontologies in such a way as to provide a freely available common ontology that has sufficient detail to precisely specify meanings of terms and concepts in domain ontologies and which is compatible with each of the starting upper ontologies. One possible method suggested to achieve this is to create a simplified upper ontology which is a compatible subset of all of the linked upper ontologies. The cooperation of the upper ontology custodians is essential, since it is likely that to achieve a compatible subset ontology or other interrelation, some modifications in one or more of the existing upper ontologies will be required. The possible areas of agreement will be discussed among the panelists prior to the March 15th meeting. It is hoped that at that meeting a public announcement, in the form of a joint communique by the co-organizers of this event, can be made that upper ontology builders will be working toward some means of interrelating their ontologies so as to provide a convenient means for users to rapidly develop knowledge bases that can be used with reasoning tools compatible with any of the existing upper ontologies. Another goal will be to develop agreement among developers of mid-level and upper-mid-level domain-spanning ontologies to use the compatible subset ontology as the common high-level ontology that will serve as their common reference ontology for specifying meanings.
  • One near-term benefit could be to enable a large degree of compatibility and interoperability for the many projects developing knowledge bases that include the most common mid-level concepts such as the classes: Person, Organization, Event, Artifact, Document, Plan, Locations, and [[TimeIntervals]] and the relationships: has-parts, spatial relations, and causes, and dozens of others. Using the compatible subset of upper ontologies, domain ontologies or mid-level ontologies spanning domains can be developed and thereby achieve interoperability with the higher-level ontologies as well as enabling interoperability with other domain ontologies. Having such an agreed source of carefully specified meaning can avoid duplication of effort as well as assure interoperability among applications using that common subset ontology, even when the applications do not use the detailed axiomatizations that are available to avoid ambiguity. Having a less complicated high-level ontology for specifying meanings can speed the development of interoperable ontologies by significantly reducing the learning and implementation barriers associated with more highly axiomatized upper ontologies, The less highly axiomatized ontologies will nevertheless be able to be used directly in the more complex systems, and can be extended to take advantage of the greater capabilities of those systems.
  • The Compatible Subset Ontology could be expected to provide immediate benefits by allowing developers of domain ontologies to all reference and reuse the same basic concepts with a greater level of detail and having considerably more precise meanings than the classes and relations now being referenced from Dublin Core or Friend of a Friend. In fact, those formalisms themselves could gain increased precision by referencing the Compatible Subset Ontology. Among the application categories that could benefit from this basic level of interoperability would be Enterprise Architecture ontologies, the developing Semantic Wiki, information exchange models for the Intelligence Community and other federal agencies, general information retrieval, e-commerce and the emerging Web Services network. Ontologies developed by using the compatible subset ontology could be immediately used and subsequently extended within any one of the more capable knowledge representation systems. This will provide a near-term basic level of semantic interoperability and a means to evolve when reusable knowledge and experience is gained within a broad community that agrees to the use of a common paradigm of meaning representation.
  • Compatibility of such a subset with existing upper ontologies would likely require that certain details in each of the starting ontologies would be left out of the common subset. It might also require that some structures in one or more of the existing high-level ontologies would need to be modified in order to permit such a compatible subset to be created (more about that below). The assumption would be that organizations requiring the higher level of functionality, or the built-in knowledge, of one of the linked ontologies, and not needing interoperability with systems using some other linked ontology, would use the more functional ontology in preference to the COSMO. Those not needing such functionality might choose the simpler ontology. In either case, expert assistance in use of the COSMO should be available from the maintainers of the linked ontologies, and perhaps from others.
  • With such a subset, it should be possible to guarantee that an ontology developed using the COSMO could be imported into any one of the linked ontologies and be logically consistent with that ontology. Export of a knowledge base from one of the linked ontologies to the COSMO might involve loss of information, depending on the ontology details. Ontology developers and users who do not yet have sufficient financial support to contract for the assistance of the expert custodians of any of the existing upper ontologies might find the barrier to development of compatible ontologies using the COSMO small enough to justify the effort that might not otherwise seem supportable.

Motivation for Developing a Compatible Subset Ontology

  • The subset COSMO might be considered as an introductory ontology to ease the learning curve for the linked ontologies but would itself have sufficient expressive power for serious applications -- at least quasi-second order (quantification over relations; returning classes as the value of a function), and more expressive than OWL + SWRL. Ontologies developed in OWL or OWL+SWRL would be able to be imported into the COSMO, provided that all semantic relations used in the less expressive ontologies had the same intended meanings as those in the COSMO. This is possible because the creators of knowledge bases in RDF or of ontologies in OWL or OWL+SWRL would be able to read the detailed documentation and axiomatization and would have to agree to accept the meanings expressed by that documentation and detailed axiomatizations, even if their own systems do not use the axioms. Accepting the nuances of meaning that the more expressive ontology specifies for relations would make even the less expressive ontologies accurately compatible with the more expressive ontologies. The upward compatibility of OWL would further reduce the learning barriers to the creation of meaningful domain ontologies by those without detailed understanding of any of the linked ontologies, or even of COSMO itself. A detailed natural-language description of the intended meanings of classes and relations in the COSMO, along with examples of usage, would provide most of the understanding required for accurate use of the included terms, even if the axiomatizations are not consulted. This would be supplemented by consistency checking. If the COSMO became widely used, third-party add-ons or interfaces, either to the COSMO or to the linked ontologies, could further improve the ease of understanding and encourage even wider use.
  • The main virtue of such a subset ontology would be to avoid the need for users to deal with the greatest complexities of any one of the linked ontologies, when it isn't necessary, but nevertheless to preserve compatibility. By reducing the time and complexity barrier for beginning use of the existing ontologies, it would encourage ontology users to build their domain ontologies using meanings that are logically consistent to each of the linked ontologies. Simple RDF triples would acquire precise meanings if the terms in the triples are all present in the COSMO or have their meanings specified by combinations of the terms in the COSMO. If the individual linked ontologies have richer detail for specific classes, the simpler class representations in the COSMO would serve as abstracted "views" of the more detailed concepts.
  • The use of specific concepts present in the compatible subset would commit users to the meanings of those concepts in each of the linked ontologies as well as in the COSMO itself. If interoperability is desired with another system using one of the linked ontologies, and reasoning is performed using that linked ontology, the more detailed meanings in that system would be definitive and controlling. Such meanings should be compatible with the less detailed meanings in the COSMO, but will not necessarily be compatible with the more detailed meanings in the other linked ontologies.
  • Having a single "entry-level" upper ontology should help to advance the technology by providing a common paradigm for experimentation with reasoning methods, and by encouraging development of applications with the expectation that they could be reused as components of a larger modular system, with the COSMO as the common means of communication between modules. Some applications such as a natural-language front end or graphical interfaces could be developed open-source to make the entry of data from non-structured sources easier.
  • The principle can be extended so that concepts in a less expressive Knowledge Classification could be linked to concepts in a more expressive ontology, and the meanings of the terms in the less expressive classification -- a taxonomy or thesaurus -- would thereby be accurately specified, even though the less expressive classification itself cannot distinguish nuances of meaning. When necessary, logical inference on terms in a less expressive ontology could be performed using the more detailed logical specifications.

Deliverable: Plans & Deadlines

  • At least one month prior to the March 15th Upper Ontology Summit meeting the custodians (developers and maintainers) of the participating public upper ontologies should begin discussion of the goals for the meeting and the method to prepare for the meeting. If possible, agreement should be reached prior to the meeting to announce at the meeting an "agreement in principle" to pursue some method to interrelate the existing upper ontologies so as to encourage creation of ontologies that can be imported or converted into one or more of those upper ontologies. The suggested method is to create a "Compatible Subset of Multiple Ontologies", an upper ontology that is less detailed than the existing upper ontologies but which can be immported into or translated into each of the linked upper ontologies.
  • the above-mentioned annoucement to be presented in the formed of a joint communique at the Panel Meeting to the public and to the press. We expect some agreement to be endorsed by the said upper ontology custodians, co-organizer and sponsor organizations, as well as key potential adopters of the work product from this agreement.
  • Development of a Compatible Subset of Multiple Ontologies (COSMO) will not preclude development of other methods to relate the upper ontologies so as to promote interoperability.

The Team

  • Upper Ontology Summit Conveners: the UOS Organizing Committee members, the Key Public Upper Ontology Custodians and UOS Key Participants are referred to, collectively, as the UOS Conveners
  • Invitees to this UOS Event:
    • upper ontology researchers and developers
    • active players in the ontology community
    • potential user organizations
    • potential funders of upper ontology and ontological engineering work
    • representatives from International Standards Bodies
    • leaders of the semantic web community
    • people in acdemia, research and industry working on semantic interoperability, knowledge representation, artificial intelligence ... who are engaged in the research, development and engineering of these subject matters

Discussion Archives & Shared-File Workspace

Resources

Conference Call, Meeting & Workshop