Introduction to Expert System

Overview
Machine Learning
Speech Understanding
Automatic
Programming Robotic
Natural Language Game Playing
Processing
Neural Network
Expert System
Fuzzy Logic
Intelligent Tutor Genetic Algorithm
AI
Computer Vision
Tree

Linguistics Computer Science

Psychology Management &
Philosophy Management Science
Electrical Engineering
What is Expert System (ES)?
“An intelligent computer program that uses knowledge and inference
procedures to solve problems that are difficult enough to require
significant human expertise for their solutions”
Prof. Edward Feigenbaum
Stanford University
an early pioneer of ES technology 1982
What is Expert System (ES)?
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
What is Expert System (ES)?
• “ES is a computer-based system that uses knowledge, facts, and reasoning
techniques to solve problems that normally require the abilities of human
experts.”
• (Martin & Oxman, 1988)

• “A computer program designed to model the problem solving ability of a

human expert”
• (Durkin, 1994)
What is Expert System (ES)?
• “A Computer program that represents and reasons with knowledge of some
specialist subject with a view to solving problem or giving advice”
• (Jackson, 1999)

• “A system that uses human knowledge captured in a computer to solve

problems that ordinarily require human expertise”
• (Turban & Aronson, 2001)
Overview of ES
• ES makes extensive use of specialized knowledge to solve
problems at the level of human expert.

• An expert is a person who has expertise in a certain area.

• ES is often applied to any system that uses ES technology -

Include special ES languages, programs, and hardware
designed to aid in the development and execution of ES.
 Basic Elements of an ES
TheUser
The User
interfacemay
interface may Knowledge
employthe
the Knowledge
employ Baseeditor
Base editor
following
following
styles:
styles:
General
General
••Question-and-
Question-and-
Inference
knowledgebase
knowledge base
User answer
answer
Inference
User engine
engine Case-specific
Case-specific
••Menu-driven
Menu-driven data
data
••Natural
Natural
language
language Explanation
Explanation
subsystem
subsystem
••Graphic
Graphic
interface
interface

ES Organization
Source: Padhy (2005)
Characteristics of Expert System
 Separates knowledge from control
 Processes expert knowledge
 Focuses expertise
 Reasons with symbols
 Reasons heuristically
 Permits inexact reasoning
 Is limited to solvable problems
 Thrives on reasonable complexity
 Makes mistakes
Characteristics of Expert System
• Basic characteristics required for an expert system:
• High performance
• Expertise
• Adequate response time – perform within a reasonable
amount of time
• Good reliability
• Self-knowledge – ability to examine their reasoning process
and explain their operation
• Understandable – be able to explain the steps of reasoning
while executing.
• Justification – justify the answers or advice.
• Flexibility – mechanism for adding, changing and deleting
knowledge.
Source: Padhy (2005) based on
Giarratano & Riley (2002)
ES vs. Conventional Programs (CP)
 ES simulate human reasoning about a problem
domain while CP simulate the domain itself.

 Building a computer model with the aim of realizing

problem-solving capabilities comparable to a domain
expert.

 Not intended to create a cognitive adequate model, i.e,

to simulate the cognitive processes of an expert in
general, but create a model which offers similar results
in problem-solving for problems in the area of concern.
ES vs. Conventional Programs (CP)
 ES perform reasoning over representations of human knowledge, numerical
calculations and data retrieval while CP perform numerical calculations and
data retrieval.

 ES solve problems using heuristic of approximation methods that do not

guarantee success while CP provide an algorithmic solutions.
ES vs other AI program
 ES deal with problems that require human expertise while most
AI programs focus on abstract mathematical problems of
simplified version of the real problems.

 ES must be fast and reliable while most AI programs do not run

very fast.

 ES must be able to explain and justify their solutions while

other AI programs do not have to meet this requirement.
ES Problem Categories
Mundane tasks:
Perception, natural
language, commonsense
reasoning, planning, etc.

Formal tasks:
Game playing, solving
mathematical
problems etc.

Expert tasks:
Scientific analysis, medical
diagnosis, etc.
ES Applications
Category
Categories
Problem addressed
Interpretation Inferring situation descriptions from sensor data

Diagnosis Inferring system malfunctions from observations

Design Configuring objects under constrains
Planning Designing actions
Monitoring Comparing observations to expected outcomes
Debugging Prescribing remedies for malfunctions
Repair Executing plans to administer prescribed remedies

Instruction Diagnosing, debugging and repairing

Control Governing overall system behavior
ES Problem Domain
 ES is develop based on expert’s knowledge.

 Expert’s knowledge is specific to one problem domain.

 Problem domain is the special problem area
such as medicine, finance, science, and etc.
ES Problem Domain
 Expert’s knowledge about solving specific problems is called the knowledge
domain of the expert.

Problem
Domain

Knowledge
Domain
Why Build ES?
 Reasons:

 Time availability (always available)

 Geographic (available anywhere)
 Safety (replaceable)
 Not perishable
 Performance (consistent)
 Speed (consistent)
 Cost (affordable)
Why Build ES?
 Human Expert vs ES
Factor Human Expert Expert System
Time availability Workday Always
Geographic Local Anywhere
Safety Irreplaceable Replaceable
Perishable Yes No
Performance Variable Consistent
Speed Variable Consistent (usually
faster)
Cost High Affordable
Why Build ES?
 To replace a human expert?
 Replacing \= eliminating
 Reasons underlying:
To make expertise available after hours
To make expertise available in other locations
To automate a routine tasks that requires an expert
To retain expertise after retirement of an expert
To reduce consultation fees
To avoid experts from danger
Why Build ES?
 To assist an expert?
 The system aids the expert in a routine or mundane tasks.
 Reasons underlying:
To improve productivity
To effectively manage the complexity of a task
To make available the information that is difficult to recall.
Advantages of Expert System
 Increased availability – Expertise available on any suitable
computer hardware.
 Reduced cost – The cost providing expertise per user is greatly
lowered
 Reduced danger – ES can be used in environments that might be
hazardous for a human
 Permanence – The expertise is permanent. Unlike human expert
who may retire, quit or die.
 Explanation – The ES can explain in detail the reasoning that led
to a conclusion
 Fast response – Fast and real time response may be necessary for
some application.
 Steady, unemotional and complete response at all times – This
may be very important in real time and emergency situations
when human expert may not operate at peak efficiency because
of stress or fatigue.
Example
To introduce terms like expert and expertise as they are
relevant to expert systems, let's suppose you have been
unable to start your car to go to work and have returned to
the house to call our favorite mechanic. The dialog might
continue something like this.

Hey Ace, this is Luis. My

Good car wouldn't start this
morning! this morning and I need some
is Ace Auto help...
Repair
Example
Here is the the beginning of the diagnostic telephone
"interview" with the mechanic...

What happens
when you turn It turns over OK, but it
the key in the just won't start.
ignition to try
to start the car?
Example
Based on your input that the starter operates, your
mechanic can abandon a number of hypotheses related to
electrical problems. Now the expert is evaluating another
possible explanation...

Well, now that you

mention it - I'm not
Hmmm...are
certain if the tank is
your sure that
empty, but it probably is.
you aren't out
of gas?
Example
At this point, your mechanic is attempting to confirm the
new hypothesis

As you crank No, I turned it over for a

the starter, do long time, but didn't
you smell gas? smell anything.
Example
Your mechanic now has enough evidence to diagnose the
problem. Once you've heard the recommendation, you
might want an explanation of how the conclusion was
obtained...

Thanks for the advice.

Based on what Mind telling me how you
you've told me, reached your conclusion?
I'm almost
certain your
car is out of
gas.
EXPERT SYSTEM APPLICATIONS
THE APPLICATION OF EXPERT SYSTEMS
 Its applications spread in a wide range i.e. in industrial and commercial
problems etc.
 Diagnosis and troubleshooting of devices and system of all kinds
 Planning and scheduling
 Configuration of manufactured objects
 Financial decision making
 Process monitoring and control
Examples of Expert System
 In medical
 In agricultural
 In education
Medical Applications
PXDES
 It is example of medical expert system.
 It is a lung disease, X-ray diagnosis.
 It takes our lungs picture from upper side of body which looks like a
shadow.
 The shadow is used to determine the type and degree of harmness.
 These systems include three modes:
(1) KNOWLEDGE BASE:-
It contains the data of X-ray representations of various stages of the disease.
(2) EXPLANATION INTERFACE:-
It details the conclusion.
(3) KNOWLEDGE ACQUISITION:-
It allow medical experts to add or change information in the system.
CaDet
 It is for early cancer detection.
 Clinical data related to early cancer detection and to cancer risk factors was
collected and incorporated in database, together with heuristic rules for
evaluating this data.
MYCIN

 One of the earliest expert system application

 It performs a task normally done by a human expert.
 It attempts to recommend appropriate therapies for patient with
bacterial infections.
 It uses LISP structures for writing internally rules.
 It uses these rules to reason backward to the clinical data available from
its goal of finding disease-causing organism.
AGRICULTURAL EXPERT
SYSTEMS

 WHY?
 It is same as other knowledge based systems.
 It uses to give answer about pest control, the need to spray, selection
of a chemical to spray, weather damage recovery such as freeze etc.
 ADVANTAGES:-
 It has ability to imitate human thought and reasoning.
 It makes modification of knowledge very convenient.
 It helps increases the production of crops
 It has ability to handle uncertain information
 It helps the farmers to take single point decision.
RICE-CROP DOCTOR
 This ES is developed by NATIONAL INSTITUTE OF AGRICULTURAL EXTENSION
MANAGEMENT.
 Its main work is to diagnose pests and diseases for rice crop and suggest preventive
measures.
 It has knowledge about diseases and pests for identification and suggesting preventive
measures.
 (1) DISEASES:
 Rice blast
 Brown spots
 Rice tungro virus
 Bacterial leaf blight etc
 (2) PESTS:
 Stem borers
 Brown plant hopper
 Rice leaf folder
 Green leaf hopper etc
AGREX
 It gives correct advice to farmers.
 Topics of advice are fertilizer application, crop protection, irrigation
scheduling and diagnosis of diseases in paddy and post harvest technology
of fruits and vegetables
EXPERT SYSTEMS IN EDUCATION
 WHY?
 Because it allow users to ask question on some education problems.
 This figure shows the architecture of ITS for teaching engineering student
which has embedded expert system inside.
 For each student expert system will create performance of student and
change levels (like easy to difficult).