Scientific Team

Knowledge Management and Transfer Using Information Technology and Expert Systems

Ahmed Rafea

Central Laboratory for Agricultural Expert System

Agriculture Research Center, 6, El Nour Street

P.O. Box  438 Dokki, 

Giza, Egypt

The transfer of knowledge from consultants & scientists to extension workers and farmers represents a bottleneck for the development of agriculture in any country. The current era is witnessing a vast development in all fields of Agriculture.  Therefore there is a need for an unconventional method to transfer the knowledge of experts in certain domain to the general public of farmers, especially that the number of experts in new technologies is lesser than their demand in a certain domain. 

The information technology played an important role in information and knowledge dissemination in the last decade. The usage of IT to transfer information and knowledge in the agriculture domain is one of the areas investigated by many institutions. Most of the Schools of Agriculture in well-known universities have built sites on the Web to disseminate agricultural information to the extension services and growers. The Central Laboratory for Agricultural Expert Systems (CLAES) has been established in 1991 within the Agriculture Research Center in the Egyptian Ministry of Agriculture and Land Reclamation, to conduct research in the area of transferring the knowledge and expertise accumulated in agricultural research to extension workers and growers, using information technology in general and expert systems, in particular.

Expert systems are simply computer software programs that mimic the behavior of human experts. They are one of the successful applications of the Artificial Intelligence field, a branch in Computer Science that investigates how to make the machine think like human or do tasks that humans do. Agricultural Expert Systems developed so far covers most of the knowledge areas for crop management including strategic decision such as variety selection, land preparation, planting, irrigation, fertilization, and harvesting, and tactic decision such as disorder diagnosis, control, and treatment. A typical agricultural expert system will ask the user about his/her plantation data for soil, water, weather, and any other abnormal observation and/or requirements, and produce a specific recommendation. It is like a plant doctor, which gives a specific advice for a certain plantation. It is not like a book or a web site where a user can find a lot of information and it is up to him/her to decide what to do. CLAES has developed in the last 10 years a dozen of expert systems for different crops and for animal health care. Experiments were conducted to measure the economic and environmental impact of using expert system in the field. The experiments showed that the net production has increased by approximately 25%. The impact on environment conservation was assessed using two measures water saving and chemicals usage reduction. It was found that fields managed by expert systems used less water by approximately 35 % and less fertilizers by approximately 16%.  The impact on enhancing the performance of the extension workers when using the expert system was also measured. A tangible enhancement was observed which ranges from 80% to 157% in different expert systems.

The objectives of this presentation are two fold first, to show how expert systems integrated with other information technologies can be used to strengthening the link between research and extension and second to report on two regional projects which are being implemented by CLAES in collaboration with ICARDA to build regional expert systems for tomatoes and cucumber under plastic tunnel for Arabian Peninsula region and wheat and faba bean for the WANA region. These projects can be used as  models for regional collaboration in gathering knowledge related to a specific commodity at the regional level, building up an electronic repository of this knowledge, and  availing this repository on the Web.

The Virtual Extension and Research Communication Network (VERCON) will be presented as a successful example that demonstrates how different information and knowledge systems are integrated to serve researchers, extension workers and growers. Other stakeholders could also find the web site very useful (http//www.vercon.sci.eg). The site contains two expert systems fro rice and wheat, extension bulletins produced by research institutes and central administration for extension, statistical data produced by the economic sector, and growers problems system that enables extension workers to interact with researchers at different levels, and keep a repository of all problems raised and its solution, and unsolved problems, if any, to be transferred to researchers to find solutions for them through their research programs. The site also provides other services as news and forums.

The experience in porting the cucumber expert system, developed under Egyptian condition, to the Arabian Peninsula will be demonstrated. The prototype expert system for cucumber can be accessed through the Arabian Peninsula Research Program web site http//www.icarda.cgiar.org/APRP/IT.htm. A workshop was held in CLAES for one week for researchers from Arabian Peninsula countries and developers from CLAES to fix the interfaces and knowledge extracted from the Egyptian version and included in the APRP expert system site. It was intended that regular reviews be done for updating information related to disorders and pesticides. More efforts are needed to fulfill this intension. The other project for wheat and faba bean are being implemented early results will be presented. In this project CLAES designed forms to acquire verities, agricultural practices, and disorders and their control in different sub regions in WANA. Most of the knowledge have been acquired and currently the system are being implemented.

Keypoints of the presentation

Objective

The objectives of this presentation are two fold

• First, to show how expert systems integrated with other information technologies can be used to strengthening the link between research and extension and

• Second to report on two regional projects which are being implemented by CLAES in collaboration with ICARDA to build regional expert systems for tomatoes and cucumber under plastic tunnel for Arabian Peninsula region and wheat and faba bean for the WANA region.

Data, Information and Knowledge

·                   Data are raw fact that need to be processed to be useful. Statistical information is a good example of information generated from storage of raw data

·                   Information is useful piece of text, table,image, .. that can directly used to reach a certain goal

·                   Knowledge is a general rule discovered as a result of research, expertise, observation,.. that can generate a piece of information. An advice generated by an expert system is a good example of information generated from knowledge.

·                   Information and Knowledge Transfer Problem

·                   The transfer of information and knowledge from consultants & scientists to extension workers and farmers represents a bottleneck for the agriculture development and hence poverty alleviation and food security.

What Information?

·                   The needed information is information that helps poor farmers to increase their production and hence increase their incomes from one hand and increase also the national production from the other hand.

·                   This information can be classified mainly into three broad categories technical, financial and marketing.

·                   The technical information includes mainly sustainable agricultural practices.

·                   The financial information includes mainly agricultural inputs prices.

·                   The marketing information is mainly historical production statistics and timely crop prices.

Information Dissemination

It is very important to

·                   access the requested information timely,

·                   update this information regularly

·                   disseminate the information instantaneously

Proposed Information Technologies

·                   Database and Web search technology are to be used for storing the technical information existing in textual form, financial, and marketing information

·                   The expert system technology can be used to store the knowledge and expertise of human experts such that the user can customize his/her consultation dynamically according to his/her field data. Expert system can also provide him with explanation for the advice provided.

·                   Computer Networks are the medium through which information can be disseminated quickly to huge number of users.

Expert Systems

·                   Expert systems are simply computer software programs that mimic the behavior of human experts.

·                   Agricultural Expert Systems developed so far covers most of the knowledge areas for crop management including strategic decision such as variety selection, land preparation, planting, irrigation, fertilization, and harvesting, and tactic decision such as disorder diagnosis, control, and treatment.

·                   A typical agricultural expert system will ask the user about his/her plantation data for soil, water, weather, and any other abnormal observation and/or requirements, and produce a specific recommendation.

·                   It is like a plant doctor, which gives a specific advice for a certain plantation.

·                   CLAES has developed in the last 10 years a dozen of expert systems for different crops and for animal health care.

Expert Systems Impact

Experiments were conducted to measure the economic and environmental impact of using expert system in the field.

·                   The experiments showed that the net production has increased by approximately 25%.

·                   It was found that fields managed by expert systems used less water by approximately 35 % and less fertilizers by approximately 16% which means that their usages have positive impact on the environment

·                   The impact on enhancing the performance of the extension workers when using the expert system was also measured. A tangible enhancement was observed which ranges from 80% to 157% in different expert systems.

VERCON/Egypt

·                   Overall objective is to improve, through strengthened research-extension linkages, the agricultural advisory services provided to Egyptian farmers (especially resource poor farmers) in order to increase production in food and agriculture

·                   Specific objective is to establish a Virtual Extension and Research Communication Network in Egypt so as to strengthen and enable linkages among the research and extension components of the national agricultural system.

Overall Structure of the Network

·                   CLAES is the central node

·                   Two sites in Cairo are connected through dial-up connection AERDRI, CAAS.

·                   ES has already Internet connection through a leased line.

·                   About 50 sites in Kafr El-Sheikh, Behira, Fayoum, Ismaillia Assiut, and Noubaria have dial up connectivity through free Internet dial up Service.

System components

·                   A search facility for extension brochures

·                   A grower’s problem solving component

·                   A search front end for an agricultural statistical DB

·                   An Expert system consultation facility

·                   A discussion forum where participating parties in the VERCON system can exchange ideas

·                   News and Event Board

Regional Projects

·                   The cucumber expert system, developed under Egyptian condition, was ported to the Arabian Peninsula.

·                   The prototype expert system for cucumber can be accessed through the Arabian Peninsula Research Program web site http//www.icarda.cgiar.org/APRP/IT.htm.

·                   A workshop was held in CLAES for one week for researchers from Arabian Peninsula countries and developers from CLAES to fix the interfaces and knowledge extracted from the Egyptian version

·                   It was intended that regular reviews be done for updating information related to disorders and pesticides. More efforts are needed to fulfill this intention

·                   Two projects for wheat and faba bean are being implemented.

·                   In this project CLAES designed forms to acquire verities, agricultural practices, and disorders and their control in different sub regions in WANA. Most of the knowledge have been acquired and currently the system are being implemented.

Sample of Knowledge collected

Variety Selection

Inputs

·         Soil characteristics, altitude, irrigation water source, meteorological data, diseases data, growth duration.

Outputs

·         Variety name, seed rate, optimum yield, seed size, susceptibility to Orbanche and fungal diseases.

Land Preparation and Planting method

Inputs

·         Soil characteristics,  selected variety, irrigation water source, irrigation type.

Outputs

·         The tillage system including the tools used, applied method in details and the optimum date for applying tillage.

·         Soil leveling (in the case of surface irrigation)

·         Ridging (in the case of surface irrigation)

·         Dividing (in the case of surface irrigation)

·         Sowing including the tools used, applied method and the optimum date for applying sowing.

·         Soil diseases control including the tools used, applied method and the optimum date for applying soil diseases control.

Fertilization 

Input

·         Soil characteristics, irrigation water characteristics, presence of rhizobia, and previous crop.

Output

·         Fertilizer type, rate and schedule at Pre-planting, Germination, Seedling, Flowering,  Pod setting and Pod filling stages.

Irrigation (for irrigated faba bean)

Input

·         Soil characteristics, irrigation water characteristics, climate characteristics.

Output

·         Irrigation schedule at Pre-planting, Germination, Seedling, Flowering,  Pod setting, Pod filling stages.

Plant Protection

Input

·         Symptoms on the plant (leaves, stem, root, flowers, fruits), soil characteristics, climate characteristics.

Output

·         Correct diagnosis of diseases, insects and weeds.

·         Treatment of diseases, insects and weeds and whenever available integrated pest management based on correct diagnosis as well as on economic threshold.

Conclusion

·                   Information Technology can play a vital role in storing, retrieving, and disseminating agricultural information

·                   The field experiments proved that ESs are very useful in reducing cost, and increasing yield.

·                   The Field experiments showed that fields managed by the ES used less chemical fertilizers and pesticides and also less irrigation water, and hence better environment conservation

·                   The researchers and extension workers used the Expert System efficiently after two days of training

·                   National Expert Systems can transferred to the Region

·                   Regional Expert Systems Development will build a repository of Knowledge of research results accumulated in the NARS’s.