>>>All the presentations are provided as PDF (Portable Document Format) files. You will need Adobe's free ACROBAT READER to access these presentations. If it is not installed on your computer, please download it [click on the Acrobat logo], to your hard drive, then double-click on the .exe file and follow the instructions.<<<

Title: AOS INITIATIVE TO CREATE A NETWORK FOR THE COOPERATIVE MANAGEMENT OF SEMANTIC STANDARDS

Creator: Johannes Keizer, Food & Agriculture Organization of the United Nations (FAO)
Format: PDF, 17 pp.

Title: BETTER ACCESS TO AGRICULTURAL INFORMATION IN CHINA: THE ROLE OF THE AGRICULTURAL ONTOLOGY SERVICE PROJECT

Creator: Chang Chun, Chinese Academy of Agricultural Sciences, Beijing, China
Format: PDF, 40 pp.

Title: BUILDING A RICH ONTOLOGY FROM AGROVOC

Creator: Dagobert Soergl, College of Information Studies, University of Maryland
Format: PDF, 35 pp.

Abstract

This presentation will start with a few examples to illustrate the usefulness of ontologies; it will then develop a set of concept relationships to be included in an ontology, based on an in-depth analysis of the BT/NT and RT relationships in AGROVOC, determining what they really mean. Lastly, it will present the "rules-as-you-go" approach to an efficient transformation of AGROVOC relationships into a more differentiated set of concept relationships. This approach relies on the ontology editor to discover - from the transformation of example concept pairs - rules that allow for the automatic conversion of other concept pairs exhibiting the same pattern, thus drastically reducing the ontology development effort.

Title: FROM AGRICULTURAL THESAURUS TO ONTOLOGY

Creator: Chang Chun & Lu Wenlin, Chinese Academy of Agricultural Sciences, Beijing, China
Format: PDF, 4 pp.

Abstract

We describe a methodology for converting a subset of terms and relations in the Chinese Agricultural Thesaurus from a database to the version of RDFS specific to the Kaon Suite of tools.

top of page

Title: THE DEVELOPMENT OF PHRASE-BASED T2E ACTIVE READING VIA WEB

Creator: Asanee Kawtrakul, Raffin Maneechayangkoon, Mukda Suktarachan, & Patchaya Boonkwan, Kasetsart University, Bangkok, Thailand
Format: PDF, 19 pp.

Abstract

This paper presents the multilingual active reading system for information exchange to assist local farmers to follow up on agricultural information through the Internet. Our system consists of three modules - language configuration, active reading assistant, and table translator. The language configuration module enables the end-users to customize the source language and the target language.

Title: THE IMPACT OF STANDARDIZED TERMINOLOGIES AND DOMAIN ONTOLOGIES IN MULTILINGUAL INFORMATION PROCESSING

Creator: Maruf Hasan, Thai Computational Linguistics Laboratory, Thailand
Format: PDF, 21 pp.

Abstract

Efficient bootstrapping of statistically derived and ontology-driven knowledge is solving practical problems in domains such as Bio- and Agro- informatics. Multilingual information processing benefits greatly from such bootstrapping, as well as from terminology standardization and ontology mapping. In this talk, I will review the latest initiatives in standardization of terminology, and introduce some general-purpose and domain-specific ontologies. Finally, I will draw conclusions by pointing out their impacts in natural language processing, machine translation and multilingual information processing research.

Title: COMBINING TERMINOLOGIES AND ONTOLOGIES TO INTEGRATE BIOMEDICAL INFORMATION

Creator: Anita Burgun, EA MCCB, Laboratoire d'Informatique Médicale, Rennes, France
Format: PDF, 5 pp.

Abstract

The post genomics era is characterized by huge amounts of biomedical information, distributed in multiple databanks (e.g. SWISS-PROT, OMIM, LocusLink, GenBank, as well as many others). Despite recent efforts to provide standard ontologies such as the Gene Ontology, semantic heterogeneity is a major obstacle to information integration. Each databank has its own identifiers for genes and gene products; the names of biological entities are associated with synonymy and ambiguity; there are numerous biomedical terminologies in use (e.g. MeSH for indexing biomedical literature). Therefore, besides the need for ontologies, there are needs for various mappings (e.g. between terminologies) and cross-references between databanks. This paper presents an ongoing project, BioMeKE, which aims at developing an information integration system providing a unified access to biomedical resources. Semantic integration in BioMeKE is based on the combination of existing terminological and ontological resources, including the Unified Medical Language System (UMLS), which integrates sixty families of biomedical vocabularies in a repository of around 900,000 concepts organized according to a set of 135 Semantic Types. Other resources provide synonyms and cross-references, such as the Genew database.

Title: ONTOLOGY-BASED METADATA SCHEMA FOR CHINESE DIGITAL LIBRARIES

Creator: Mao-sheng Lai & Xiu-dan Yang, Department of Information Management, University of Peking, China
Format: PDF, 7 pp.

Abstract

There will be three parts in this paper, which mainly focuses on ontology-based metadata for Chinese digital library projects. Firstly, we discuss today's principal Chinese metadata schemes, their common methodology; common elements and application problems from the point of view of one static metadata schema. Secondly, we will discuss the semantics of definitions of elements for Chinese metadata schemes in digital library projects, focus on mapping and cross-walking among different metadata schemes based the ontology methodology. Lastly, we will give a formal ontology for the Chinese digital library, and the ontology concepts description will start from the IFLA proposal for FRBR, but we will rename some concepts of IFLA according to Chinese specialization. We anticipate that the ontology will be used to describe metadata for digital content, enabling computational inference to support powerful user queries. The ontology will include at least two modules: one generic to many digital libraries, and another specific to, for instance, the Rare Book Digital Library of Peking University.

Title: PEDRO: A TOOL TO DEVELOP AN APPLICATION THAT CREATES DATA ENTRY FORMS BASED ON A DATA MODEL WRITTEN IN A PARTICULAR STYLE OF XML SCHEMA

Creator: Kevin Garwood, E-Science North West Centre, Manchester University, U.K.
Format: PDF, 9 pp.

Abstract

PEDRO is an application that creates data entry forms based on a data model written in a particular style of XML Schema. Users can enter data through forms, and create data files that conform to the schema. They can use controlled vocabularies to mark-up text fields and have the application perform basic validation on field data. Once the user has finished writing a data file, PEDRO will tell them if they have left out any required records.

Title: CONSTRUCTION OF KNOWLEDGE DATABASE FOR AUTOMATIC INDEXING AND CLASSIFYING BASED ON CHINESE LIBRARY CLASSIFICATION

Creator: Hou Hanqing & Chun-xiang Xue, School of Information Science & Technology, Nanjing Agricultural University, China
Format: PDF, 7 pp.

Abstract

Class number, descriptors and keywords are three kinds of subject identifiers, and among them exist some concept mapping relationships. Furthermore, there are many manually indexed records, including the class number of the Chinese Library Classification(CLC) scheme, and keywords or descriptors, in the Chinese bibliographic database. Looking at these manually indexed data, we found that the CLC could be applied to organize the concepts and terms in every field, to create a knowledge database which reflects the mapping relationships between class number and keywords or descriptors. This knowledge database is an full-fledged specialist system, which includes the CLC and its index, the Chinese Thesaurus, Concordance between class numbers and keywords, keywords list, stop-words list, a dictionary of synonyms, a place name list, etc. After extracting the terms from all the available literature, such as title, abstract, keywords provided by Creators even full-text, weighted indexing, then measuring the semantic similarity between the indexing terms and the terms or phrases in the knowledge database, it is possible to carry out automatic indexing and classifying. .

Title: AOS AND ONLINE AGRICULTURAL INFORMATION SERVICE IN GUANGDONG, CHINA

Creator: Zhong Wang, Liang Huang and Jie Chen, Institute of Sci-Tech Information, Guangdong Academy of Agricultural Sciences, China
Format: PDF, 3 pp.

Abstract

In recent years, the Institute of Sci-Tech Information of the Guangdong Academy of Agricultural Sciences has developed many websites to provide online service of agricultural information management for local governments and enterprises. The content of these websites is mainly generated by information organized in a group of databases of related agricultural fields. Over time we have observed that the development and progress in the construction and relevant applications of the AOS could enhance and leverage our current services. This paper provides a rough discussion and the preliminary output of adopting the corresponding technologies to leverage the original information in those databases to the knowledge level, as well as the potential effects on the current and future related services.

top of page

Title: AN INTELLIGENT RETRIEVAL SYSTEM FOR CHINESE AGRICULTURAL SCIENTIFIC LITERATURE

Creator: Ping Qian, Xiaolu Su, Scientech Documentation and Information Center,Chinese Academy of Agricultural Sciences
Format: PDF, 3 pp.

Abstract

An intelligent retrieval system for Chinese agricultural scientific literature will be introduced in this paper; the system is based on the ontology theory and native XML database. The Chinese Agricultural Scientific Literature Database, containing more than 560,000 records, was adopted as the data source and the Chinese Library Classification Method was used as the standard. The system has an actual classification tree which shows the distribution of the classes for agricultural science literature there is also a class-keyword database containing more than 320,000 records. The XML and TAMINO were used for the development of the hierarchical data structure; the main functions of the system include user registration, login and management, as well as browse search and intelligent search. The system is easier to use and improves the speed and accuracy of retrieval.

Title: DESCRIPTION LOGICS AS DATABASES FOR STRUCTURING AGRICULTURAL INFORMATION SYSTEMS

Creator: Howard Beck & Soonho Kim, Institute of Food & Agricultural Sciences, University of Florida
Format: PDF, 7 pp.

Abstract

A database management system that uses a description logic as a data modelling language is used as a basis for building agricultural databases. This system supports ontologies as well as semantic query processing, conceptual clustering, and natural language processing. On-line graph-based tools are used for creation of applications. Application examples include an interface for image classification and retrieval in the crop-pest domain, a case-based reasoning system for rootstock selection in citrus, an environment for crop modelling and simulation, and tools for the development of agricultural educational resource.

top of page

Title: OntoEdu - ONTOLOGY-BASED EDUCATION GRID SYSTEM FOR E-LEARNING

Creator: Cui Guangzuo, Modern Education Technology Center, Peking University, China
Format: PDF, 5 pp.

Abstract

Based on several new technologies, such as ubiquitous computing, ontology engineering, semantic web and grid computing, this paper proposes a flexible educational architecture for e-learning, which is called OntoEdu. OntoEdu's core is educational ontology. It is divided into four parts: User adaptation, Service composition, Education ontology, Semantic Educational service Grid. With educational ontology and semantic grid, OntoEdu realizes the reusability concept, device and user adaptability, automatic composition, function and performance scalability. The simple OntoEdu1.0 implementation shows that OntoEdu architecture is viable and flexible..

Title: USING THE THAI AGROVOC FOR RETRIEVAL OF THAI AGRICULTURAL INFORMATION

Creator: Aree Thunkijjanukij, Thai National AGRIS Centre, Kasetsart University, Bangkok, Thailand
Format: PDF, 17 pp.

Abstract

Data and information produced locally are normally displayed in the native language. To be useful, therefore, they also have to be recorded and indexed in a local language. Indexing and computer processing for the Thai language is extremely difficult, since Thai uses no spaces between words. This makes it difficult to use a word separator for uncontrolled vocabulary indexing; thus, a controlled vocabulary becomes a key player in the creation of an efficient information retrieval system. AGROVOC is a multilingual, structured and controlled vocabulary/thesaurus for indexing data in agricultural information systems. AGROVOC has been greatly improved and is now the best general agricultural thesaurus available, and the only one with an international updating mechanism which will ensure its continued evolution. The Thai AGROVOC was developed using FAO's AGROVOC as a prototype. More than 16 thousand descriptors were translated and linked with 7 language descriptors as a multilingual thesaurus. An intelligent retrieval system was designed to improve the efficiency of the retrieval process, and provide query expansion using a thesaurus-derived ontology. The query system can expand for multiple languages, searching synonyms, and boarder, narrower and related terms, from local databases and internet resources.

top of page

Title: NEXT-GENERATION KNOWLEDGE MANAGEMENT FOR MULTILINGUAL AGRICULTURAL INFORMATION

Creator: Asanee Kawtrakul, Department of Computer Engineering, Kasetsart University, Bangkok
Format: PDF, 50 pp.

Abstract

At present, an increasingly dispersive flood of unstructured electronic articles and reports has adversely affected the information perception of readers. This problem is especially evident in large organizations. Besides these problems concerning information sources, there is, furthermore, a "knowledge pile-up" from which previous particular experiences are amalgamated. Many linguistic phenomena come into play in original full-text retrieval. Synonyms can seriously affect the retrieval process to the extent of producing deficient recall, or a fall-off in precision, since the documents retrieved do not match the need of users. In addition, the state-of-the-art retrieval systems render only full-form documents; they cannot yet summarize themes and knowledge.

Title: BUILDING AND TESTING FLORICULTURAL DOCUMENTATION ONTOLOGY-BASED DEMO RETRIEVAL SYSTEM

Creator: Li Jing, Library of Chinese Academy of Sciences
Format: PDF, 4 pp.

Abstract

After the experimentation of modelling the Floriculture ontology (FO) and developing a retrieval system based-on FO, the following conclusions were drawn:

1) Domain ontology (DO) is not built and used independently. DO must be based on the Upper ontology (UO) with correct Logic structure, and the Inference Engine (IE) of UO must be used in DO.

2) The Numbers of core concepts in DO (FO) < The Numbers of concepts about Horticulture << The Numbers of concepts in AOS <<< The Numbers of concepts in Cyc ® KB (Knowledge Base). The Rule is the Numbers of concepts distribution among DO, the more upper domain and General ontology (also called "Upper ontology") resembles the image of an upside-down pyramid.

3) The retrieval effect based on the ontology is better than the effect of full text retrieval theoretically, but the usage of the IE slows down retrieval speed.

4) Concept marking-up in documentation DB, concepts marked-up by CycL and transported in Cyc ® KB, performed manually, were almost complete. This is the main bottleneck to developing an ontology-based system.

top of page