CARAVAN 17

December 2002

Breeding Barley for Drought Resistance

By Samir El-Sebae Ahmed,
Ahmed Rafae and
Khalid Shaalan

An expert system is a computer program that simulates a human expert in solving a problem in a narrow domain area. Computer expert systems have been used in agriculture since the early 1980s. Several systems have been developed in the United States and Europe for plant-disorder diagnosis, to assess production practices, technologies, and other aspects of farming. In Egypt, the Central Laboratory for Agricultural Expert Systems (CLAES) of the Agricultural Research Center (ARC), Ministry of Agriculture and Land Reclamation, has the mandate to identify and develop expert systems and train researchers, extension specialists, and growers in their use, including maintenance and updating. CLAES has developed and deployed several plant and animal agricultural expert systems.

Expert systems differ from traditional computer software applications in that they deal with a symbolic knowledge base and not a database or mathematical models, but they can also be integrated with traditional software to form a complete system. They safeguard and continue to make good use of the expertise of highly qualified specialists, even after the specialists retire or leave an institute’s service. An expert system can also integrate the knowledge base of different specialties, such as irrigation, nutrition, plant pathology, and entomology in one package. The expert system’s advice takes into consideration the specific data of a certain plantation, whereas traditional extension documents leave part of the

NEPER strategic component main menu

reasoning up to the grower, as it is impossible to include all possible situations. Therefore, expert systems can play a valuable role in transferring agricultural research results to farmers’ fields efficiently, through the existing extension systems, as the number of extension and growers is large relative to the number of researchers who can respond to the needs of extensionists and growers.
Annually, since 1997, a two-week regional training course on “Utilization of Expert Systems in Agricultural Research and Production” has been conducted jointly by the International Center for Agricultural Research in the Dry Areas (ICARDA) and CLAES in Cairo, Egypt (Table 1). The course introduces expert systems technology to senior and mid-level management from the countries of Central and West Asia and North Africa. The course covers the basic components of an expert system and how they differ from other types of software programs, such as databases and simulation modeling. The course also covers briefly the expert system development life cycle, namely: knowledge acquisition, knowledge analysis and modeling, requirement specifications, design, implementation, verification and validation, evaluation, maintenance, and training. An expert system development tool is also introduced and participants develop experimental prototype expert systems in their specialties.

Table 1. Number and origin of national scientists who participated in the ICARDA/CLAES joint training course on "Utilization of Expert Systems in Agricultural Research and Production," 1997-2001.

Region	No. of participants	Country

West Asia	20	Iraq, Iran, Jordan, Lebanon, Syria, Turkey
Nile Valley and Red Sea	13	Egypt, Eritrea, Ethiopia, Sudan
Arabian Peninsula	10	Oman, Qatar, Saudi Arabia, UAE, Yemen
North Africa	5	Algeria, Libya, Morocco, Tunisia
Central Asia	3	Armenia, Kazakhstan, Tajikistan
Total	51

The Follow-up Study
To determine the impact of this new technology on the former participants and their national agricultural programs, a comprehensive questionnaire was prepared and forwarded in early 2002 to all the former training course participants. Twenty-two completed questionnaires were received and analyzed. Fifteen former participants were invited to a follow-up training workshop held in October 2002 at CLAES headquarters in Cairo.
The workshop was held to 1) assess the impact of the annual course, 2) identify the problems faced by former participants in

First follow-up workshop in Egypt to evaluate the impact of expert systems in agricultural research and production. Former trainees in the ICARDA/CLAES courses participated in the workshop.

developing expert systems in their countries, 3) update the former participants on the latest methodologies and tools for developing expert systems, and 4) help ICARDA make correct decisions on continuation, content, duration and further improvement in the course. Thirteen invitees took part from Egypt, Ethiopia, Jordan, Qatar, Lebanon, Morocco, Sudan, Syria, Turkey, and Yemen.
The workshop participants recommended that the course be continued at CLAES. It was also recommended that more emphasis be placed on issues related to the design and implementation of expert systems for specific agricultural operations. A set of expert system building tools was developed at CLAES for user-friendly utilization by agricultural researchers and others. These tools will be used to develop expert systems in future courses.

Workshop outcomes

Course content and material

More information on design and implementation issues related to expert systems, such as user-interfaces, control strategy, etc., needs to be included in the course.
The expert system building tools should be upgraded for smooth, user-friendly utilization by agricultural researchers and other users.
Practical training should address wider issues of interest to national programs.
Course materials and the expert system building tool and its user manual should be available on the Internet for easy downloading and up-dating.

Who should attend the course?

Research scientists and extension/technology transfer specialists should continue to attend the course, but they must have a good grasp of the English language and computers.
Computer specialists should be invited to the course to encourage the creation of teams of expert systems specialists at the national program level.

Course duration and location

Most of the participants said the present two-week course is too short. It was suggested that the course be extended by a week to allow more practical work on design and implementation issues. If cost prevents this, then the two-week course should be intensified so that more time per-day is allocated for practical training.
The course should continue to be conducted at CLAES headquarters in Cairo and/or at another country in the region.

Introducing new courses

Courses oriented toward design and implementation issues related to expert systems and building tools should be organized for computer specialists with programming skills.
Courses to ‘train-the-trainers’ in expert systems should be organized at CLAES headquarters, with follow-up, on-the-job training at the country, sub-regional, and regional level.
A highly specialized course was recommended for national programs progressing well in building their own expert systems.

Communication between CLAES and former participants

To enhance links between former participants, CLAES and ICARDA, it was recommended that a website be built that includes contact information of the former participants, as well as course material, tools, and solutions to common problems. This website was developed during the workshop: http://www.claes.sci.eg/meetings/icardaws
It was suggested that participants who develop their own expert systems might benefit from feedback from CLAES. It was requested that any scientist who develops an expert system send it to CLAES for further validation and evaluation.

Conclusion

Expert systems were recognized by most of the participants as a useful tool for transferring agricultural knowledge between researchers, extension officers and growers, as well as for assisting decision makers in solving agricultural problems. National cooperators appreciate the joint efforts of ICARDA and CLAES in developing and organizing this course.

Dr Samir El-Sebae Ahmed (S.Ahmed@cgiar.org) is Head of the Human Resources Development Unit at ICARDA, Dr Ahmed Rafae is Director of the Central Laboratory for Agricultural Expert Systems (CLAES), Cairo, Egypt, and Dr Khalid Shaalan is Senior Scientist, CLAES.