The term "ontology" has been used for a number of years by the artificial intelligence and knowledge representation community but is now becoming part of the standard terminology of a much wider community including information systems modelling and XML. It describes domain knowledge in a generic way and provides an agreed-upon understanding of a domain. A more concise definition might be: An ontology is a system that contains terms, the definitions of those terms, and the specification of relationships among those terms.
These proceeding are the result of the Fifth Agricultural Ontology Service workshop held in Beijing, China from 27 - 29 April 2004. The workshop built on the previous workshops held in Rome, Oxford, Florida and Copenhagen. It was jointly organized by the Scientech Documentation and Information Center of the Chinese Academy of Agricultural Sciences (SDIC/CAAS), China and the Library and Documentation Systems Division of FAO.
The workshop brought together metadata and ontology experts and researchers interested in the AOS project. It provided an opportunity for information management specialists in Asia to learn about the AOS and to strengthen partnerships in building the AOS. Around fifty participants attended the workshop. Most of the attendees were librarians, information managers, and subject specialists in agriculture and development from the Asian continent.
On the first day of the workshop two tutorials were organized:
o Basics of Ontology
o Ontology Tools
During the introductory speeches, Dr Johannes Keizer introduced the AOS Initiative and its goal to create a network for the cooperative management of semantic standards. He presented the progress report and reiterated that we have been heading in the right direction; however, he underlined the need to attract a bigger institutional framework to support the initiative. This was followed by a presentation by Mr Chang Chun, who talked about the role of the AOS project in providing better access to agricultural information in China.
The presentations were divided into four sessions:
Session 1: From Thesauri to Ontologies
The presentations in this session focused on the migration of thesauri to ontologies. The basic structural differences between the two were described, and the possible procedures that could be applied to convert a thesauri to an ontology were illustrated. The rules-as-you go approach was presented as a possible methodology to help the process of transformation. The group discussed the limitations of the existing term-to-term relationships in thesauri and how these relationships can be improved by adding semantic knowledge.
Session 2: Automatic translation using semantic knowledge
The presentations in this session covered the role of domain-ontologies in a multilingual environment. A Phrase-Based T2E Active Reading System was explained in detail. The presentations also covered the impact of standardized terminologies and domain ontologies in multilingual information processing from the perspective of natural language processing.
Session 3: Metadata schemas
Ontologies provide formalized semantics of objects and their relations to each other, and metadata schemas and ontologies are very closely related. The Chinese ontology based metadata model which applies the Functional Requirements of Bibliographic Records (FRBR) was explained in detail. Pedro, a tool to create data entry form based on a data model written as XML Schema was also presented.
Session 4: Use of semantics to enhance access to domain knowledge
This session showed illustrations of applications and application prototypes that exploit semantic information encoded in such resources as documents, thesauri, and ontologies. The objective of this session was also to provide concrete examples and methodologies that can eventually be extended to the development of live applications used in and for information management. The presentations included a demonstration of an application which uses description logics as databases for structuring agricultural information systems.
The group discussed the need for new tools that will facilitate the maintenance of AGROVOC and its conversion to an ontology. The general understanding was that ontology creation is a time consuming process and that it is not possible for a single organization to do all the work on its own; for this reason the need for collaborative efforts was stressed. There was a consensus on the need for speedy update of AGROVOC. Some of the participants who are using the thesaurus expressed that some cleaning of the relationships needs to be done in the current version.
There are several on-going ontology development projects like the Fishery ontology service (FOS) and the Food Nutrition and Agriculture (FNA) projects that are being coordinated and developed by the Library and Documentation Systems Division of FAO. These provided a chance to see how the theory of ontologies is brought to the users through practical applications. Overall the workshop has achieved its goal; the participants said that the workshop enabled them to enhance their understanding of ontologies. They were impressed by some of the applications that demonstrated the power of ontologies. The need for users, domain experts and ontology engineers to work together in close collaboration was also acknowledged. The next steps in creation or development of ontologies should take into account the existing knowledge structures already available.
The participants expressed that these workshops should be an annual event, and needs institutionalization. The objective of these workshops should change from raising awareness to a more scientific and technical environment. The workshops could take place during other Agricultural Conferences and Seminars. Additionally, they should have slots wherein participants have the chance to discuss bilateral or multilateral collaborations.
