Abstract

              Artificial Intelligence has given new meaning and importance to knowledge. Now, it is possible to package specialized knowledge and sell it with a system that can use it to reason and draw conclusions. This paper studies how the concepts of AI can be useful to India.

       An Expert System is a computer program intended to embody the knowledge and ability of an expert in a certain domain. The spectrum of applications of expert systems technology is so wide as to defy easy characterization. We seek to use Expert Systems to improve communication, understanding and management of medical information. The  objective of  is  the coalescing of data, knowledge, and the tools necessary to apply  that data and knowledge in the decision-making process, at the time and place that a decision needs to be made.

       We present a System that can diagnose plant deficiency diseases. It can also diagnose some of the diseases that affect the wheat plant. It identifies the disease based on a knowledge system provided by experts. The System may be enhanced so as to extend its use to other agricultural domains, thus revolutionizing the face of Rural India. The ultimate aim of the system is to enable less qualified medical staff to make a diagnosis which would emulate that of the expert.

(i)Artificial Intelligence

     Artificial Intelligence is a branch of computer science concerned with the study and creation of computer systems that exhibit some form of intelligence: systems that learn new concepts and tasks, systems that can reason and draw useful conclusions about the world around us, systems that can understand a natural language or perceive and comprehend a visual scene, and systems that can perform feats that require human type of intelligence.

AI programming is characterized by declarative programming. The programmer does not have to fix all the details of execution of the program. Instead, he puts down explicitly all the basic facts and laws which appear necessary and useful for the solution to his problem.

(ii)Expert Systems

An expert system is a computer program that represents and reasons with knowledge of some specialist subject with a view to solving problems or giving advice. An Expert System differs from other AI applications in the fact that it deals with subject matter of realistic complexity that normally requires considerable amount of human Expertise. An Expert System exhibits high performance in terms of speed and reliability. It is also capable of explaining and justifying solutions in order to convince the user that its reasoning is correct.

(iii) AI in medicine

The Artificial Intelligence in Medicine field emerged in the early 1970’s in response to several simultaneous needs, opportunities and interests. An increased demand for high-quality medical services coupled with the explosive growth of medical knowledge has led to the suggestion that computer programs could be used to assist physicians and other health-care providers.

Owing to the advancements in research more and more data becomes available and most practitioners can no longer embody a broad and deep expertise in their fields. This problem is more prominent in developing countries like India where the number of primary physicians is meager compared to the population. This problem costs many lives owing to the non-availability of expert personnel.

The problem of non availability of experts is more marked in the agricultural field as India is an agriculture based economy and the majority of the populace makes their living directly through farming. Age-old farming techniques are followed which lead to less productivity levels. This coupled with the fact that most countryside farmers are unaware of the various diseases that affect their crops leads to a sorry state of affairs and huge losses are suffered due the lack of communication between the Agricultural Research stations and the Indian farmer. Such situations are not uncommon and they seriously limit the production. Effective control is based on accurate diagnosis of these diseases.

(iv)PLADS (PLAnt Diagnostic System)

Introduction

PLADS is an expert system that can diagnose diseases in wheat plants based on user inputs and help him enter accurate symptoms

Problem Domain

Initially we programmed a diagnostic system that identifies the nutrition deficiency which the plant suffers and one that diagnoses some types of diseases in the wheat plant.

The area under cultivation, production and yield per hectare of wheat  indicate its importance. This paper does not attempt at providing all the features of a plant diagnostic system, but seeks to direct attention towards the possibility of such an endeavour.

Target Users: Farmers, Students

Feasibility

   Writing an expert system generally involves a great deal of time and money. To avoid costly and embarrassing failures, we have tested the system for feasibility.

1. Need for a Solution

With the government launching Kisan call centers and shifting the focus to modernization of agriculture its time to make some revolutionary changes in the existing agricultural setup. Even if we introduce Kisan call centers the major problem still remains. We obviously do not have enough experts to work at the block level. Our farmers have been troubled for ages by lack of a proper and timely plant disease diagnosis system.

2. Costs Involved

The total cost of development is meager compared to the yearly crop losses that run into crores of rupees. Much labor is spared, as the related facts are already available with our experts.

3. Non-Availability of Human Expertise

Around 60% of our population is into farming. We do not have enough experts to cater to their needs. As a result the Indian farmer is still a poor picture.

4. Physical Skills or Common Sense

PLADS is highly user friendly and can generate excellent results with the help of a little knowledgeable person. PLADS works using symbolic reasoning techniques backed by a knowledge base.

5. Job Threat

The jobs of our experts are not being threatened; rather this is to use their expertise and make it widely available.

PLADS is a knowledge based system which performs a task by applying rules of thumb instead of applying Statistical methods. PLADS encodes the domain dependent knowledge of everyday practitioners in the field of Agriculture and use this knowledge to solve problems.

(v)System Architecture

 PLADS comprises of the following prominent modules: -

User Interface

The end user (farmers, students) interacts with the system through a user interface (GUI), which may use menus, natural language or any other style of interaction. This is determined by the front end in use (e.g. visual basic).

Inference Engine

Inference engine reasons with both the expert knowledge database and data specific to the disease to determine the degree of probability assured.

Expert Knowledge base

It is the database that stores the facts (e.g. symptoms) related to the wheat plant and fuzzy logics required to deal with disease diagnosis. The expert knowledge will typically be in the form of a set of IF-THEN rules.

Case Specific Data

The case specific data includes both data provided by the user and partial conclusions (along with certainty measures) based on this data. For a backward chaining system like PLADS the user data is kept in working memory to draw more accurate conclusions.

Explanation System

This allows PLADS to explain its reasoning to the user based on its resources.

Knowledge Base Editor

PLADS knowledge base editor will assist the expert or knowledge engineer to easily update and check the knowledge base, which ensures that the system remains future compatible.

      The expert system shell (represented by the dotted box) is normally a general purpose one, that with minor changes can be combined with any slightly different knowledge base required to solve a different problem. Using a expert system shell helps decrease the developing time and cost to a large extent.

We can use a suitable shell and link it to PLADS’s knowledge base to generate the final product.

(vi)Application

Like most expert Systems, PLADS is a backward chaining rule interpreter. PLADS replaces the 2 Boolean values, true or false with a range of values called certainty factors. It enhances the diagnostic system and guarantees better results. PLADS caches all the facts it derives in a database. PLADS performs the computation the first time only. When a similar computation is called for again, it simply fetches the stored result. PLADS is goal driven, once it identifies the goal, the rules that are appropriate to the goal are applied. The Expert System Shell will have 2 classes of users. The experts will use the shell while developing the system and the end users who will use the completed Expert System.

(vii)The Algorithm

It is a typical case of representing knowledge by using rules.

PLADS’s algorithm is made up of a set of basic IF-THEN rules written using Common LISP programming language.

;;;Context:

(defun pds( )

           "Determine what disease infects the plant"

(

 ( list (defcontext leaf( ))

         (defcontext deadspots( ))

         (defcontext stalkslender( ))

         (defcontext disease( )     (identity)))))

;;;Parameters:

(defparm leaf(member drying wilted chlorotic)  "leaf is :"t)

(defparm deadspots(member rapidlyenlarging scattered atmargins)  "deadspots :"t)

(defparm stalkslender(member true false)  "stalkslender:"t)

(defparm identity disease(member NitrogenDeficiency Magnesium deficiency Potassiumdeficiency Zincdeficiency Copperdeficiency Manganesedeficiency Irondeficiency))

;;;Rules

(clear~rules)

(defrule 1

if(leaf=drying)

  (stalkslender=true)

then 0.5

 (identify disease as Nitrogendeficiency))

(defrule 2

if(leaf=chlorotic)

  (deadspots=scattered)

  (stalkslender=true)

then 0.8

 (identify disease as Magnesiumdeficiency))

(defrule 3

if(leaf=chlorotic)

  (deadspots=atmargins)

  (stalkslender=true)

then 0.8

 (identify disease as Potassiumdeficiency))

(defrule 4

if(deadspots=enlarging)

then 0.8

 (identify disease as Zincdeficiency))

(defrule 5

if(leaf=wilted)

  (stalkslender=true)

then 0.8

 (identify disease as Copperdeficiency))

(defrule 6

if(leaf=chlorotic)

  (deadspots=scattered)

then 0.5

 (identify disease as Manganesedeficiency))

(defrule 7

if(leaf=chlorotic)

  (stalkslender=true)

then 0.5

 (identify disease as Irondeficiency))

;;;Context:

(defun pds( )

      "Determine what organism infects

wheat plant"

(

 (list(defcontext leaf( ))

       (defcontext plant( ))

       (defcontext pathogen( )             identity)))))

;;;Parameters:

(defparm leaf(member powderypatches chlorotic blight spots Brownspots)"leaf is:"t)

(defparm plant(member brown dwarf dying bleachedspikelets black)"plant is:"t)

(defparm identity pathogen(member Erysiphegraminis SeptoriaNodorum Fusarium Tritici Tilletia Fusariumgraminearum pucciniarecodita swmvirus))

;;;Rules

(clear~rules)

(defrule 51

if(leaf=powderypatches)

then 0.8

 (identify pathogen as Erysiphegraminis ))

(defrule 52

if(leaf=chlorotic)

  (leaf=brownspots)

then 0.8

 (identify pathogen as SeptoriaNodorum))

(defrule 53

if(leaf=blight)

  (plant=brown)

then 0.8

 (identify pathogen as Fusarium))

(defrule 54

if(leaf=chlorotic)

  (plant=dwarf)

then 0.5

 (identify pathogen as Tritici))

(defrule 55

if(leaf=lightspots)

  (plant=dwarf)

then 0.8

 (identify pathogen as Tilletia))

(defrule 56

if(plant=bleachedspikelets)

then 0.8

 (identify pathogen as Fusariumgraminearum))

(defrule 57

if(leaf=brownspots)

  (plant=black)

then 0.8

 (identify pathogen as pucciniarecodita))

(defrule 58

if(leaf=chlorotic)

  (plant=dwarf)

then 0.5

 (identify pathogen as swmvirus))

Degrees of Probability:

0.2 Low Probability

0.5 Moderate Probability

0.8 High Probability

How it works

IT conducts a question and answer session. The basic operation of IT is simple. The main function pds performs a consultation from the context name. It then asks the user for the values of various parameters which were previously declared. It then calls another function to find the result. The system does a pre-scan before checking all the clauses whether it is already known to be false or of a probability less that the threshold value.

(viii)Conclusion:

Expert System Shells are widely in use, but they are used to solve fairly simple problems.

One of the first Medical Diagnostic System, built in 1976 had a correct diagnosis rate which was better than most physicians. Despite immense improvements in technology, the role of expert systems in the applied medical field is very limited.

It is obvious that PLADS can boost the efficiency of any person at a Kisan call center with a little knowledge in the field. The knowledge base that we provide is to an extent of an expert level. We propose the development of this idea by experts in the fields of Agriculture and Plant Pathology. Thus, PLADS will aid the agricultural community and help improve efficiency in the field of plant disease diagnosis.

References:

1. “Introduction to Expert Systems”, Peter Jackson.

2. “Introduction to Artificial Intelligence and Expert Systems”, Dan W. Patterson.

3. “Paradigms of Artificial Intelligence Programming: Case studies in Common Lisp”,  

     Peter Norvig.

4. “Intelligence and Artificial Intelligence: An interdisciplinary Debate”,

     U.Ratsch,M.M.Richter,I.O.Stamatescu

5. “Plant Pathology”, R.S.Mehrotra.

6. “Plant Physiology”, Salisbury and Ross.

7. “Biomedical Information Technology: Global Social             Responsibilities  for the Democratic age”, Harold Sackman.

8. http://www.cee.hw.ac.uk/

9. http://www.surrey.ac.uk/

10. http://www.scism.sbu.ac.uk/

11. http://www.wikipedia.org/

12 .http://www.wtec.org/loyola

13 .www.faqs.org

14. http://www.nationalpak.com/

15. http://edis.ifas.ufl.edu