Lab 06 Sol

Questions 6
Dr. Roman Belavkin
BIS4435
Question 1
What is the main purpose of Expert Systems?
Answer: The main purpose of ES is to replicate knowledge and skills of

human experts in a particular area, and then to use this knowledge to solve
similar problems without human experts participation (computationally).
Question 2
What is the main difference between conventional computer programs and

production systems (rule–based systems)?
Answer: Conventional programs use pre–programmed algorithms usually

with many control flow statements (loops, iterations) that clearly define
the order in which the components of a program should work. Production
systems use only IF–THEN rules, and there are no special instructions about
the order in which they should be used. The system decides by itself which
rules are used based on the situation (problem state, current goal, etc).
Question 3
Why a production system model was used to implement the first rule–based
expert systems?
Answer: Production systems (rule–based systems) became a convenient

platform for models of human cognition. Because the aim of ES is to repli-
cate human expertise, production rules (IF–THEN rules) can be used to
represent and encode the knowledge of a human expert. Then the produc-
tion system can apply this knowledge to new problems.
Question 4
What types of knowledge are used by neural networks and by rule–based sys-
tems? What kinds of systems are they with respect to the type of knowledge
they use?
1
BIS4435 2
Answer: Neural networks use sub–symbolic knowledge stored in a form

of weights of many neurons. Rule–based systems use symbolic knowledge
stored in a form of rules and facts. Thus, neural networks are sub–symbolic
while rule–based are symbolic systems.
Question 5
What are the main components of a rule–based system?
Answer: The main components are:
• Knowledge base
• Working memory
• Inference engine
• Explanation system
• User interface
• Knowledge base editor
Working Memory
(current state)

-Inference
Engine I@
UI R
@
User 6 Knowledge Base
? (rules, facts)
KE
-Explanation
System 6
Question 6
Describe the phases of designing an expert system? What term is used to

call the whole process?
Answer: There are three stages in designing ES:
Knowledge acquisition : the process of getting the knowledge from ex-

perts (by interviewing and/or observing human experts, reading spe-
cific books, etc).
Knowledge representation : selecting the most appropriate structures

to represent the knowledge.
BIS4435 3
Knowledge validation : testing that the knowledge of ES is correct and

complete.
The whole process is called knowledge engineering.
Question 7
What are the main areas of application of ES according to Waterman (1986)?
Answer: The main areas of ES application are:

Interpretation — drawing high–level conclusions based on data.
Prediction — projecting probable outcomes.
Diagnosis — determining the cause of malfunctions, disease, etc.
Design — finding best configuration based on criteria.
Planning — proposing a series of actions to achieve a goal.
Monitoring — comparing observed behaviour to the expected behaviour.
Debugging and Repair — prescribing and implementing remedies.
Instruction — assisting students in learning.
Control — governing the behaviour of a system.
Question 8
Why is knowledge acquisition often referred to as the ES ‘bottleneck’ ?
Answer: Human experts are rare, expensive to train and limited in how
much they can do. While humans may be expert at problem solving they
may not understand much about how their perform their decision making.
Therefore, while there may be many problem domains suitable for ES de-
velopment, the cost and time and effort to acquire knowledge from human
experts often means projects take a long time, are expensive or may never
to started since they would be uneconomic.
However, with ever improving methods of automated processing (and
’machine learning’) from data in electronic form, in problem domains where
there are large amounts of data available, it may soon be possible to auto-
mate all or part of the knowledge acquisition process. Such developments
may be one way to reduce the reliance on human experts and begin to
‘unblock’ the bottleneck of knowledge acquisition at present.
Question 9
Why do many people say they will not trust a ‘robo-doc’ medical diagnosis
expert system?
BIS4435 4
Answer: For ES to be useful users have to be able to trust their conclu-

sions. When the conclusions have very important implications (such as life
and death, large amounts of money, significant implications of the quality of
people’s lives) it is only natural that people might prefer to trust in human
decision makers. While it would be accepted that human decision makers
make mistakes, they do not get ‘corrupt programs’ or behave erratically
(unless mentally unstable) - human expertise exhibits graceful degradation
in that as situations become less like those the expert has previously expe-
rienced, humans are still able to make reasonable decisions in most cases.
Much work in modern artificial intelligence involves improving the ’grace’ of
the responses of computer systems to unexpected situations.
For many the best solution is to augment human expertise with ES. So,
humans have assistance with their decision making (perhaps having sugges-
tions made they had not thought of), and computers have their performance
validated for unreasonable behaviour in the light of extraordinary combina-
tions of circumstances.
Question 10
Describe at least 3 advantages that expert systems offer organisations that

would otherwise have to employ human experts.
Answer: There are many different possible advantages, depending on prob-

lem domain, organisational structure and design of the expert system. Some
advantages include:
• Increased output / productivity - expert systems can often make de-

cisions much faster than humans, or deal with larger problems than
humans. An example would be the XCON system that resulted in
sped up order processing by DEC.
• Availability of scarce / expensive expertise - while human experts are

expensive and can only be in one place at one time, once developed
expert systems are cheap to duplicate and can therefore offer their
expertise at many locations simultaneously.
• Also the expertise of an ES may be used to provide training without

the need for so many human teachers.
• Reduced need for human work in dangerous situations - a self-sufficient

robot powered by an expert system can be used on Mars or underwater
or in buildings that have collapsed, reducing human expose to danger
and speeding up rescue or exploration work.

Lab 06 Sol

Uploaded by

Copyright:

Available Formats

Lab 06 Sol

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lab 06 Sol

Uploaded by

Copyright:

Available Formats

Questions 6

Dr. Roman Belavkin

What is the main purpose of Expert Systems?

Answer: The main purpose of ES is to replicate knowledge and skills of

What is the main difference between conventional computer programs and

Answer: Conventional programs use pre–programmed algorithms usually

Answer: Production systems (rule–based systems) became a convenient

Answer: Neural networks use sub–symbolic knowledge stored in a form

What are the main components of a rule–based system?

Answer: The main components are:

• Knowledge base editor

Describe the phases of designing an expert system? What term is used to

Answer: There are three stages in designing ES:

Knowledge acquisition : the process of getting the knowledge from ex-

Knowledge representation : selecting the most appropriate structures

Knowledge validation : testing that the knowledge of ES is correct and

Answer: The main areas of ES application are:

Answer: For ES to be useful users have to be able to trust their conclu-

Describe at least 3 advantages that expert systems offer organisations that

Answer: There are many different possible advantages, depending on prob-

• Increased output / productivity - expert systems can often make de-

• Availability of scarce / expensive expertise - while human experts are

• Also the expertise of an ES may be used to provide training without

• Reduced need for human work in dangerous situations - a self-sufficient

You might also like