An expert system is a type of artificial intelligent where the decision-making ability of a human expert is emulated by a computer software system (Jackson 1999). A plant pest model is a simplification of reality, thus, it cannot be expected to be completely accurate. Human experience should always be relied upon when deciding upon pest management practices rather than place blind faith in any forecasting model. Expert systems strive to simulate complex human knowledge based reasoning and provide expertise where it otherwise would not be readily available. Such pest management decisions may involve data from multiple plant pest models, plant growth models, weather forecasts, pest identification photographs, scouting reports, pesticide application records, programmed identification keys, and other sources of information that are retrieved automatically or manually input. It may not be a perfect substitute for direct human expert experience, but an agricultural expert system can be a powerful tool in a grower’s toolbox. Nevertheless, it is only as good as the sum of its parts and reliable plant pest models must first be developed, tested, and refined.
An expert system utilizes a fixed inference engine, a continuously changing knowledge base, and a conversational user interface (Leondes 2002). The inference engine provides the logic reasoning that invents solutions to problems and is often developed independently by third party mathematicians and computer scientists. A conversational user interface is developed in conjunction with an inference engine when it is not possible to provide all the necessary data upfront. A knowledge base is built from a large compilation of 'IF-THEN-ELSE' statements and is reasoned upon by the inference engine. An expert system shell provides the framework of an expert system by providing an inference engine, conversational user interface, and an empty knowledge base. An expert system developer who uses a 3rd party expert system shell builds the knowledge base and often will integrate it into a larger core computer system that provides data to the inference engine, a front end to the conversational user interface, and processes the results from the inference engine.
DemiAg is our core project that has been in development for the last 2 decades and still has a long way to go. DemiAg is an open source project with a twist! DemiAg will eventually be released to hopefully hundreds of other computer programmers and scientists who will contribute to them using data provided by hopefully thousands of users like yourself. Other open source projects consist of computer software with its source code made available and licensed in which the copyright holder provides the rights to study, change and distribute the software to anyone and for any purpose. The problem is that DemiAg will rely upon a huge knowledge base of data that will be too big to simply be copied and run in isolation. The plan for DemiAg is to disseminate the source code in the typical open source manner as extensible modules while the knowledge base of data is distributed in mirrored nodes hosted by network servers across the internet. DeMilia Research LLC is currently creating the base module open source code release, laying the foundation for the DemiAg knowledge base nodes, and creating a front end user interface available to the public. DemiAg will be gradually supplemented with tools for modifying existing plant disease models and creating new ones. It is hoped that enough interest will be generated in the scientific community such that it will serve as a virtual workspace for plant disease model developers. In addition to DemiAg being released as a web based software application, smartphone and tablet computer apps will provide an alternate user interfaces and provide a platform for smaller specific purpose software applications such as DemiAg Tag. Future code modules, knowledge base nodes, and front end user interfaces hosted by others may be either public or private for use in both non-profit and commercial purposes that may either share or hoard their resources...we'll see what happens!
- Jackson, P. (1999). Introduction to expert systems. Harlow, England ; Reading, Mass., Addison-Wesley.
- Leondes, C. T. (2002). Expert systems : the technology of knowledge management and decision making for the 21st century. San Diego, Academic Press.