4 Expert System

Transcript

Brief Introduction

Hi! Friends. In our last interaction, we discussed at length about the differences between the website and web portal. I am sure by now you must be thinking of, how can we develop a web portal for the purpose of providing extension services, specially the farming community, the input supply agencies, and various other service providers of agriculture and allied sciences. In this interaction we will be discussing about the concept of expert system and some of the examples of expert system that are being adopted in agriculture as well as agricultural extension services.

What is an Expert?

So now let us have a look at the concept of an expert. Basically we call an individual as an expert,

• Who solves the problem easily. Who provides us the solutions for existing problems.
• And who asks the appropriate questions. Because without asking questions he cannot solve this problem. And the number of characteristics that we are discussing here.
• And ultimately they use their knowledge very efficiently to provide the services to the end-users

Then the similar things are being simulated with the help of the computer, which we call it as an expert system. Basically expert system is a software.

• That emulates the human expert. Whatever the services that an expert provides to solve a particular problem. So now the same things are being adopted with the help of a machine, which we said that it emulates the human expert.
• It deals with the small, well defined domains of expertise, wherein that depends on the information that what we are feeding to this systems, and accordingly it is analyzing the issues and giving the solutions.
• Is able to solve the real world problems based on a set of questions that the receiver is asking.
• It is able to act as a cost effective consultant, because it is a machine that we are designing in such a way, that it is providing solutions on a regular basis, on a continuous basis without tiring, which a human being it is a bit difficult. That is how it is becoming cost-effective.
• Then it can explain reasoning behind the solution it finds as per the information that has been fed to the system.
• And it should be able to learn from the experiences also.

Basic Functions of an Expert System

So if you look into the functions of this expert system. So there is a user who is in search of a solution for a particular problem. And he has number of facts with him, so that he is feeding to the system. And there is a knowledge base in the system, which is being fed by an expert in the form of a digitized information. So when you start asking the question, it establishes the interface with the engine. And it tries to develop the logical solution for that and ultimately, as an expert it gives solution to the end user.

This particular storehouse as well as the logical analysis process that we are trying to identify it as the expert system.

Problem Domain vs Knowledge Domain

So if you look into the problem domain versus knowledge domain.

• Every individual has a limitation that he can be an expert of a particular subject matter. He cannot think of being a subject matter specialist of each and every subject. So that is how the problem domain is much larger than the knowledge domain.

You can have the knowledge domain of maybe agriculture, finance, science, engineering etc. that becomes your domain knowledge, but when you look into the problem domain, it is very huge in nature

• The expert’s knowledge about solving specific problems is called the knowledge domain, which is relatively small compared to the problem domain, which is relatively larger
• The problem domain is always superset of the knowledge domain, which is larger in nature.

Creating an Expert System

Then how can we create an expert system.

• With the help of extracting the knowledge and methods from an expert. That’s what is the knowledge acquisition process. It is a very lengthy process, wherein we need to collect a huge amount of data to develop an expert system. Because it is the first and foremost process is the knowledge acquisition process.
• And the second one is the knowledge representation process. How we are processing and how the outputs are delivered.

So these are the two basic steps in creating an expert system.

Paradigm of Rules

So this entire process works on a set of rules. How we frame these rules. So there is if-then relationship. What is that if-then relationship. If for example in case of agriculture. If the leaves are green or yellow or black or whatever it is. So the color that we will be denoting here and the condition of the roots that we are trying to put here. So then what might be the possible effect of the color of leaf is like this, and the color of roots are like this. Or the status of leaves are like this or/and the status of roots are like this. So then what are the possible effects of that. And accordingly the solution is provided.

Another set of rules. If your soil color is maybe red, then the texture is sandy, then water holding capacity is less. So then the suitable crops might be, the crops like maybe the bajra or maybe groundnut, whatever it is, so on and so forth. It depends on the if and then relationship. If is defining various conditions, wherein we are trying to fill in the blanks by providing appropriate conditions and accordingly the system is going to give us the results.

Desirable Features of an Expert System

Then desirable features of expert system include,

• Dealing with uncertainty. Because we don’t know what kind of questions that we are going to face. And at the same time what kind of answers that the system has to provide. So that is why it is trying to collect huge amount of information.
• It should provide the explanation, based on the if and then relationship, as we have seen in the previous slide.
• Then the ease of modification. There should be a space for modification of maybe the response that the system is giving.
• And there should be transportability
• As far as adaptive learning features in case of expert system.

So this is example of an expert system, which is very popular nowadays is the ‘Rice Doctor’. So we will be discussing these slides like the expert system that is developed by the Tamil-Nadu Agricultural University, in collaboration with the Indian Council of Agricultural Research.

Expert system for the paddy. So if I enter into this particular system. So what are the problems that I am facing. Maybe related to the insect management. So then I go for the crop protection page of this expert system. So based on my questions, so there are different pests that are mentioned there. So it asks which kind of pest problem that you are facing. So looking into the nature of damage, I have identified that pest as a stem borer. So for identification of that particular insect. So there are number of photographs that are provided on that particular page.

Looking into the photographs that are available on that page as a part of expert system and what are the symptoms that I observed in my field. So combining these two things, I am coming to the conclusion that yes it was the stem borer. So then it gives you the symptoms. What are the possible symptoms of this stem borer. So how it looks like. Then how exactly the crops looks like at the field, which is affected with this stem borer. So then based on these questions and the relevant answers from the receiver. So it is trying to provide you the possible solutions.

The management practices. Maybe the chemicals that you need to go for. Or some of the management practices that you can think of by using various alternative methods of control. So this is how the expert system is trying to provide you the relevant answers, based on the set of questions that the receiver is asking.

Advantages and Limitations

The advantages of this expert system include,

• Capture of scarce expertise. Because providing appropriate information itself is the biggest challenge, specially in case of the agriculture and allied sciences. Looking into the nature and the biological system, that what we are facing.
• Then superior problem solving. It is situation specific, location specific. The same pest and the same disease, which is prevalent here in the similar conditions, it may not be prevalent in some other parts of the country or other parts of the State itself. So that is how the problem solving mechanism becomes more complex in case of agriculture and allied sciences.
•  And the reliability is one of the important feature of this expert system, you can rely on because it has already collected huge amount of data, based on the data it is giving you the response, under the supervision of an expert who has designed this particular model.
• Then work with even the incomplete information. So even if you are not able to identify all the symptoms. But if you are able to provide some of the responses. Even then you will be getting the best possible results.
• Then you can transfer this particular knowledge, one to many.

The disadvantages or the limitations of this expert system include.

• The expertise is hard to extract from the experts. Maybe it is the human expert. Or maybe it is the system who is acting as an expert. Because you don’t know how and you don’t want to tell, and all do it very different manner as far as the human beings are concerned. And similarly, the similar things are being fed to the computer systems. And accordingly it becomes a bit difficult for the end-user to get the response.
• Then knowledge not always readily available, because of the lacunas of the information that is being provided to the expert system. If that information is not complete in nature. So accordingly the results are also being influenced.
• Then difficult to independently validate the expertise. We need a huge number of experts to validate the experiences. So that is how, that’s another limitation of this
• Then high development, initial costs are a bit higher for development of these particular systems.
• Then it only work well in the narrow domains. Because we cannot think of all the problems and every problem related to agriculture and allied sciences. We need to start with a specific problem, then we need to go ahead with.
• Then the machines cannot learn from their experiences. It is the individual who is learning. Then it has to be adopted as per the needs of the machine, and as per the logical conditions of the machine.
• And it is not suitable for all types of problems. Because there are number of location specific problems, region specific problems and the problems which need higher attention by the end-user. Such issues we cannot take it up, in case of the expert system.

Conclusion

To conclude we can say that the expert system

• Helps to overcome the shortage of experts in the specific knowledge domain. As we have already said the routine type of problems, that can be taken care by these expert systems as on date. So looking into the specific problems. So that can be taken care of by the specific individuals who are already available, because with that we are saving the time of the experts also. So they can spare their time for some other activities.
• Then it save the time of expert. Machine answer similar to the question by all receivers, because there are number of people who are facing the similar problems. And all of them if they come to maybe an extension professional, or a scientist. It takes lot of time of the expert as well as the extension professional to answer the same questions. But once we put these things into the machine. Machine can tirelessly take care of these situations, and they keep on providing the answers to the users.
• Then the learner should adopt the method of learning by doing. Until and unless he becomes part of this system, so he cannot understand the problems as well as the possible solutions. So he should be part of this system as well as his participation should be ensured. So that he can get the better solutions for this.

So this is in a nutshell about the concept of the expert systems. And the number of expert systems that are available, because Indian Council of Agricultural Research has also developed a number of expert systems, which are available on Indian Agricultural Statistical Research Institute website. IASRI website. So there are a number of expert systems in the name of Agridaksha, Maizedaksha, then Aqua portals are available, wherein you can find the solutions for the problems that are what the farming community is facing.

So with this we are coming to the end of this discussion. And in the next interaction we will be discussing about the artificial intelligence and its application in agriculture.

Thank You.