Expert Systems

What is an expert system

An expert system is a software system with a large database (often called a knowledge base) and a built-in set of facts and rules which enable it to appear to be an expert in a certain area similar to a human expert, by analysing and solving complex problems. An expert system is a knowledge-based system that attempts to replace a human 'expert' in a particular field i.e. emulating a human in the decision making / problem solving process.

Every Expert system will consist of 3 main components

Knowledge base e.g. large database of knowledge (aka A knowledge base)
Inference engine - it is based on a set of rules for making deductions from the data
User Interface - facilities for searching the knowledge databaseAn expert system is a software system / type of artificial intelligence

Purpose of an Expert System

Using a knowledge-based system, a user can interrogate a vast knowledge base to find a solution.
The user enters the information he or she already has, and the knowledge-based system infers an answer.
If the system cannot reach a conclusion, it requests more information. This process is continued until either the system has a high degree of confidence that it knows what the solution is, or it is sure that, having exhausted every line of inquiry, it does not know the solution.

Artificial Intelligence (AI) uses software to try to give computers the capabilities of human intelligence (reasoning, speech recognition, learning etc).

A neural network is a way of connecting and programming processors that attempts to simulate the way a human brain works (Neurons are connected through synapses!). Still experimental but used in pattern recognition and learning. They use parallel processing techniques where a number of different processors are performing tasks at the same time, rather than sequential processing where a single processor performs a sequence of tasks one after another.

There are special high level programming languages used for Artificial Intelligence software such as PROLOG and ASPRIN.

Examples

• Software which diagnoses problems and gives advice on what the causes of those problems are.

• They can also give advice on solutions.

• A program that gives advice on (medical) matters.

There are many uses of expert systems in ICT such as

Medical diagnosis
Medical screening for cancer and brain tumours
Matching people to jobs or careers advice
Training on oil rigs
Diagnosing faults in car engines e.g. expert systems used by car mechanics
Legal advisory systems
Mineral prospecting
GUI helpline
Oil exploration
Determining an insurance premium or deciding whether to provide a bank load/mortgage

Data Capture

Expert Systems - Data Capture

• Keyboard

• Touch screen

• Voice

• Sensor

Example - Health Care

Expert Systems - MEDICINE

Expert systems are widely used in medicine and health-care applications. The purpose of an expert system is to help diagnose an illness/give medical advice. It allows the doctors a second opinion.

Specific Uses - Medical Expert Systems

Mycin - This expert system was designed to identify bacteria causing severe infections, such as bacteria and meningitis, and to recommend antibiotics. A medical diagnosis expert system (eg MYCIN) would have information about diseases and their symptoms, the drugs used in treatments etc.

A patient is asked by a doctor about symptoms and the replies are input to the expert system. The computer searches its database, uses its rules and makes suggestions about the disease and its treatments. Sometimes probabilities are assigned to diagnoses.

The computer does not take the place of the doctor but can be used to help the doctor make decisions.

A medical diagnosis expert system.

An ES might be used by the doctor to diagnose unusual or complex conditions from information gathered from the patient. The doctor / medical profession might be concerned that ES might lead to the loss of doctors' jobs or deskilling of the profession. The ES might save the doctor's time by covering simple conditions first and allowing the doctor to concentrate on more difficult medical situations. Doctors might also welcome the ES if it enables unusual medical conditions to be diagnosed rapidly and reliably.

Medical screening for cancer and brain tumours

An expert system is a software system with a large database (often called a knowledge base) and a built-in set of facts and rules which enable it to appear to be an expert in a certain area similar to a human expert, by analysing and solving complex problems. An ES might be used by the doctor to diagnose unusual or complex conditions from information gathered from the patient.

The ES might reduce the number of years spent training to be a doctor. It might save the doctor's time by covering simple conditions first and allowing the doctor to concentrate on more difficult medical situations. Doctors might also welcome the ES if it enables unusual medical conditions to be diagnosed rapidly and reliably. The doctor / medical profession might be concerned that ES might lead to the loss of status / esteem or deskilling of the profession.

How are expert systems used?

There is an input stage of the system (e.g. providing a sequences of questions on-screen)
There is a processing stage (e.g. the program asks different questions, depending on responses received, until sufficient information obtained to make a decision.) The ES processes facts according to a set of rules.
Next is the output stage (e.g. a diagnosis and treatment advice is presented on screen). The program comments on how the conclusion was reached, explaining its reasoning. This allows the user to evaluate the validity of the decision. The program indicates the certainty of the result produced (e.g. there is a 70% change of outcome A, but 30% chance of outcome B). This alerts the user to possible alternative solutions.

Advantages of using an expert system in medicine

Can store more knowledge/information than a human expert. It can draw on a wide variety of sources such as stored knowledge from books case studies to help in diagnosis and advice.
One expert system can be used across a network / the Internet
Data can be kept up to date and updating an expert system is cheaper than training a human expert
Doctor’s time can be spent more effectively if supported by an expert system
Helps to give a more accurate diagnosis (Not ‘accurate’ on its own) - less likely to make mistakes
Computer does not 'forget' like a doctor or make mistakes
The expert system is always available 24 hours a day and will never 'retire'. / 24/7
Some people prefer entering personal data into a computer rather than talking to a person
Some people prefer the privacy of talking to a computer.
Can be used to generate a second opinion / support suspicion
Allows self-diagnosis and avoids waiting times.
Allows access to an expert where not available locally / saves leaving home to go to doctor
Cheaper to use an expert system e.g. rather than training a human expert.
One expert system can be used across a network / the Internet. The system can be used at a distance over a network. So rural areas or even poorer third world countries have access to experts.
Provides accurate predictions with probabilities of all possible problems with more accurate advice.
Can be used as a training tool for doctors

Disadvantages of using an expert system in medicine

Over reliance upon computers, some doctors could be de-skilled by over dependence upon computer advice.
Fewer medical staff could be needed
Lacks the 'human touch' (Lack of common sense, human senses). Lack of common sense // cannot interpret human behaviour / factor in context
No human available if you do not know what to do.
Initial setup and development costs- ‘Expensive’ needs to be qualified as development costs or expensive to purchase initially.
Not 100% accurate e.g. dependent upon the correct information being given. If data or rules wrong the wrong advice could be given e.g. Correct diagnoses depends on correct data being initially entered (not just GIGO)
Over reliance on the software. People using an expert system without guidance could under/over/mis-diagnose
Impersonal // some people do not like communicating with a machine // lack of human contact

EXPERT SYSTEM - CAR MECHANICS

Advantages of using an expert system in Car Mechanics

Diagnose a car’s fault
Can remap an engine
Suggested remedial actions, e.g. replace clutch

Disadvantages of using an expert system in Car Mechanics

The computer can store far more information than a human
The computer does not 'forget' or make mistakes / less human error
Data can be kept up-to-date
The expert system is always available 24 hours a day and will never 'retire' NOT just '24/7'
The system can be used remotely, so rural areas or even poorer third world countries have access to experts
Provides accurate predictions with probabilities of all possible problems with more accurate advice
Can be used to train mechanics.

Other General Advantages

Advantages of the expert system to the user (e.g. faster and more consistent loan application decisions, more accurate medical diagnosis with less chance of important symptoms being overlooked.)

An on-line Example : An expert system for cataloguing coins.

Page updated

Report abuse