the national center for ontological research n.
Skip this Video
Loading SlideShow in 5 Seconds..
The National Center for Ontological Research PowerPoint Presentation
Download Presentation
The National Center for Ontological Research

The National Center for Ontological Research

0 Views Download Presentation
Download Presentation

The National Center for Ontological Research

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

  1. The National Center for Ontological Research A brief guide Barry Smith

  2. Barry Smith • publishing in ontology since 1975 • Handbook of Metaphysics and Ontology, 1991, repr. 2001 • Professor of Philosophy, Buffalo • Director, Institute for Formal Ontology and Medical Information Science, Saarbrücken • Underlying idea: the project of developing ontologies can profit from the theories developed by philosophers over 2500 years of ontological research.

  3. General problems in ontology building • dominance of pragmatism & of need to solve computational problems • lack of coordination • lack of benchmarks • nothing like the challenge evaluations of AI or NLP • no Darwinian mechanism for cumulative improvements • Can we resolve these problems institutionally?

  4. Example of current chaos: HL7-RIM • Animal • Definition: A subtype of Living Subject representing any animal-of-interest to the Personnel Management domain. • LivingSubject • Definition: A subtype of Entity representing an organism or complex animal, alive or not. • “There is no distinction between an activity and its documentation.”

  5. Special circumstances of Europe • ECOR: European Center for Ontological Research • Founded 2004 in Saarbrücken • responding to needs in EU for ontology coordination in e-health domain

  6. ECOR Partners include • Laboratory for Ontology, University of Turin (Maurizio Ferraris) • Laboratory for Applied Ontology, Trento and Rome (Nicola Guarino ) • Istituto di Linguistica Computazionale, Pisa (Nicoletta Calzolari Zamorani) • DFKI (German Research Center for Artificial Intelligence) Competence Center Semantic Web • Bremen Ontology Group (John Bateman) • Foundational Ontology Group, University of Leeds (Brandon Bennett ) • Geneva Centre for Ontology (Kevin Mulligan) • Austrian Centre for Ontological Research (Christian Kanzian)

  7. ECOR • looks beyond the realm of computational artifacts and takes seriously the question: does this ontology correspond to identifiable entities and relations on the side of reality itself ? • does not dictate any philosophical stance with respect to reality, but seeks to do justice to the fact that the very same reality may be sliced in different ways when addressed from different perspectives; • but whatever philosophical stance is taken, it is to be used consistently and rigorously and on the basis of clearly stated principles

  8. NCOR • National Center for Ontological Research • inaugural event in Buffalo • October 27, 2005 • • a consortium of academic, industry and government partners • recognizing need for integration of data across disciplinary boundaries in ways that can support computer processing • two administrative sites: • University at Buffalo • Stanford University

  9. NCOR goals • to advance ontological research, development and evaluation within the United States • to develop challenge evaluations and other tools and measures for quality assurance of ontologies and thereby to establish best practices for ontology development • to provide coordination, infrastructure, and other forms of support for investigators working in the United States on theoretical ontology and on applications in fields such as ontology of the sciences, spatial and cognitive ontology, terminological systems, enterprise ontology and in a variety of defense- and homeland security-related projects • to provide US researchers working in ontology-related areas with specialized support in seeking external funding and in assembling collaborative, interdisciplinary teams both nationally and internationally

  10. NCOR Goals • to engage in massive outreach endeavors designed to broaden the range of institutions and individuals accepting the goals of high quality ontology in both theory and practice • bring ontology groups together in an attempt to stop people reinventing (square) wheels • facilitate data and information integration • via • standardized tools • standardized top-level classes and relations

  11. NCORPartner Institutions • NCOR’s two principal sites in Buffalo and Stanford (Mark Musen) have substantial track record of nationally and internationally funded research in ontology. Both committed to principles-based ontology as a tool for the integration of data and information in scientific research and commercial applications. • NCOR’s partners who will be recruited from academia, industry and government, should have a proven track record of excellence in ontological research and in the application of ontology to solve concrete problems. • NCOR will facilitate a variety of joint projects, ranging from joint supervision of doctoral students to collaborative participation in nationally and internationally funded research networks. • NCOR will develop an internship scheme designed to bring about a cross-fertilization between the academic and industrial aspects of ontological research.

  12. Partners • Research Partners • Structural Informatics GroupUniversity of Washington, Seattle • Gene Ontology ConsortiumJackson Laboratory, Bar Harbor, Maine • Lawrence Berkeley National Laboratories • Industrial Partners • Ontology Works Inc., Baltimore, MD • International Partners • The European Centre for Ontological Research • The Japanese Ontology Forum • Canadian Center for Ontological Research (?) • See also:

  13. principles for high-quality ontological research • NCOR partners: • accept the interdisciplinary character of ontological research • agree to work seriously on developing and testing a range of complementary quality measures, e.g. foundational, software-based, methodological, ... • agree to apply such quality measures systematically to their own work • agree to participate in the dissemination of these quality measures and help others in applying them and to contribute to the development of adequate benchmarks

  14. Example of the methodology in action • The OBO Relation Ontology • “Relations in Biomedical Ontologies” • Smith, Ceusters, Klagges, Köhler, Kumar, Lomax, Mungall, Neuhaus, Rector, and Rosse • Genome Biology, 2005, 6 (5) • brings together representatives of OBO, GALEN, • FMA, BFO communities with biologists, to agree on a common set of formally defined relations

  15. OBO Relation Ontology • • required by all new submissions to Open Biomedical Ontologies library • now managed by National Center for Biomedical Ontology • serves as model for OBO UBO (upper biomedical ontology)

  16. Some basic distinctions:between different kinds of relations • <class, class>: is_a, part_of, ... • <instance, class>: this explosion instance_of the class explosion • <instance, instance>: Mary’s heart part_of Mary

  17. Instance-level relations • part_of • located_in • contained_in • has_participant • has_agent • earlier • . . .

  18. Class-Level Relations Relations in Biomedical Ontologies

  19. We can reason across ontologies developed in relation to entities at different levels of granularity • but only if the top-level categories and associated formal-ontological relations are well-defined and used consistently • Top-level categories for biomedical informatics now being defined under the auspices of OBO-UBO (upper bio-ontology)

  20. is_a • A is_a B =def. every A is a B • human is_a mammal • all instances of the class human are as a matter of necessity instances of the class mammal

  21. part_of • A part_of B =def every A is a part_of some B • part_of= instance-level parthood • (for example between Mary and her heart) • all-someform • Note that A part_of B may be true when B has_part A is false • Consider A = human testis, B = human

  22. NCOR Open Issues (examples) • To be discussed at planning meeting in Buffalo on October 27: • reasoning vs. statistical approaches • syntactical regimentation and semantic interoperability • 2. relation to OWL, W3C, Semantic Web endeavor (focus on logic vs. focus on content)