Professional Documents
Culture Documents
Week 2 - Introduction To Expert Systems
Week 2 - Introduction To Expert Systems
Intelligence
Module Code & Version #
• Expert system
• Categories of expert system
• Components of expert system
• Expert System shell
• Characteristics of expert system
• Expert System Life Cycle
• Expert System
• Domain expert
• Knowledge Engineer
• Domain
knowledge domain.
Identify the domain of expertise for which each of the following human
experts is well known
• Jamie Oliver
• Gary Lineker
• Patrick Moore
• Stephen Hawking
Application Description
Design Domain Camera lens design, automobile design.
Diagnosis Systems to deduce cause of disease
Medical Domain from observed data, conduction medical
operations on humans.
Comparing data continuously with observed
Monitoring Systems system or with prescribed behavior such as
leakage monitoring in long petroleum pipeline.
Medical Domain
CaDet, an expert system that can classified tumor at initial
phase.
PXDES is used to analyze degree of lung cancer
Design Domain
FSSE (Feng Shui Expert Evaluator). This system will identify
whether a location fulfill the feng shui criteria and suggest to
the architect on the best angle to design the building.
Note:
Expert system can be implemented in other domains such as tourism, commercial,
education, etc.
All expert systems produce advice as output but the way that
the output is used determines the category of the expert
system. There are several ways of categorizing the output
provided by an expert system.
For example:
• Companies with no legal department use expert systems to
advise on certain aspects of the law.
• Oil companies use expert systems like GeoQuest and GeoPlay
to advise on the most likely places to drill for oil.
• British Gas use expert systems to predict where corrosion is
most likely to occur in the gas supply system.
For example:
• Botanists use expert systems to classify rare plants
• Geologists use expert systems to categories minerals
• Thesys is an expert system that is used in some universities to
allow students to grade their own written reports before
submitting them
• Expert systems are used by many supermarket chains to
identify patterns of customer spending.
For example:
• Medical expert systems are used to diagnose illnesses and
diseases.
• Government agencies are using expert systems for
troubleshooting and intelligent alternative selection.
• Airlines use expert systems to help the pilot to diagnose
problems with the aircraft.
• GEC Marconi used an expert system called APEX to identify faults
in complex circuit boards – in fact, diagnostic expert systems are
widely used to help speed up the repair of complex equipment.
For example:
• The SUMit expert system was used by KLM airlines to plan
employee rotes.
• The expert system used by the Standard Communications
Company is used to plan the design of fibre optic cables
• The expert system used by Lufthansa aircraft is used to plan the
flight path for an aircraft’s journey
• Military expert systems help provide strategic advantage over
the enemy by assisting in battle assessment
Knowledge
Base
Human Inference
Knowledge
Expert Engine
Engineer
User
Interface
End User
(Maybe non-
expert)
Human Expert
Someone who is an expert where his/her knowledge is taken to
develop an expert system.
Knowledge Engineer
Knowledge engineer is a technical person who collects and
integrates knowledge into the knowledge base.
Forward Chaining
• It is a strategy of an expert system to answer the
question, “What can happen next?”
• Here, the inference engine follows the chain of conditions
and derivations and finally deduces the outcome. It considers
all the facts and rules, and sorts them before concluding to a
solution.
Forward Chaining
• This strategy is followed for working on conclusion, result, or
effect. For example, prediction of share market status as an
effect of changes in interest rates.
User Interface
• User interface provides interaction between user of the ES
and the ES itself. It is generally a natural language processing,
takes the user's query in a readable form and passes it to the
inference engine.
• After that, it displays the results to the user. In other words,
it's an interface that helps the user communicate with the
expert system.
Knowledge Knowledge
Acquisition Representation
System
Validation
Knowledge Acquisition
Unstructured Interview
Unstructured Interview
• There are THREE (3) types of question that can be use such
as probes, prompts and seed questions.
2. A probe question – encourage the expert to provide
further information without indicating what that
information should be.
3. A prompts question – more directed and can help
return the interview to a relevant topic that is
incomplete.
Structured Interview
• A framework for the interview is determined in advance.
• It involves the use of check-lists and questionnaires to
ensure focus is maintained.
• Strict structure interview, allow the engineer to compare
answers between experts, less strict (semi-structured
interview) combine a focus detail with some freedom to
explore areas of interest.
Focused Discussion
• It is based on a particular problem or scenario – case study,
critical incident or specific solution.
• Case analysis – consider a case study that might occur in the
domain. The expert explains how it would be or was solve
by verbally or demonstration.
• Critical incident analysis – look at the unusual and probably
serious incidents.
• Critiquing – expert is asked to comment on someone else’s
solution to a problem or design.
Role reversal
• It switches the role between an engineer and expert.
• There are TWO (2) types of role reversal:
1. Teach-back interview – the engineer “teaches” the
expert on a subject that has already been discussed
and allows the expert to amend the knowledge if
necessary.
2. Twenty questions – engineer chooses a topic from a
predetermined set and the expert asks questions
about the topic in order to determine which one has
been selected.
Think-aloud
• It is use to elicit information about specific tasks.
• The expert is asked to think-aloud while carrying out the
task.
Knowledge Representation
System Validation
Q&A