Expert System
Text Book
THE ESSENCE OF EXPERT SYSTEM
Introductions
The ideas of AI began to emerge as a
separate field of study in the 1940s. Some
50 years on much has happened to AI –
there have been many success stories and
many perceived failures.
2
The beginning of AI
Research in AI began during the 1940s; it
was during this time that the first
generation of computers appeared in
research institutes. The foundations of
machine logic were based on mathematics
pioneered by work of Kurt Godel, Alonzo
Church and Alun Turing and Whitehead and
Russell (1913) helped to produce formalized
methods in logical reasoning.
3
Turing (1950) developed the Turing
machine, which demonstrated that a
simple number processor could
manipulate symbols as well as numbers.
He later become better known for the do
called Turing Test, used for comparing
machine and human intelligence.
4
It was not until 1956 that John Mcarthy first
used the term “artificial intelligence” during
a conference held at Dartmouth College in
USA.
Twenty-five years later, as is now known, it
is clear that such predictions were over
optimistic. AI did not fulfill expectations. The
reason for this lay in the assumptions made
by these scientists at that time.
5
General problem-solving AI programs
AI scientists tried to emulate the “thinking
processes” underlying human reasoning by
finding general methods for solving broad
classes of problems. One such project,
developed by Newell and Simon (1976), was
known as the general problem solver
algorithm or GPS.
6
The birth of KBS
One of the shortcomings of the general problem-
solving approach was that as the size of the problem
increased, so the ensuing search space grew
exponentially. Only by making such programs less
general, and concentrating on the specific
knowledge pertaining to the problem, could such
search spaces be reduced.
From the first crop of KBS that emerged, there was
an emphasis on real-life domains such as diagnosing
infectious diseases or predicting mineral deposits in
various geographic regions of the world. This was an
era experimentation on real-life problems to see if
the ideas were achievable in practice.
7
Some experimental expert systems
8
The emergence of commercial expert
systems
Many of the expert systems that were built in the 1970s
were experimental and confined mainly to academic
research. It was not until the 1980s that they started to
make the transition from research laboratory to
commercial system.
Japanese and US interest in expert system
In Japan, a massive 10 years fifth-generation project in
1982. The USA also took a great deal of interest in the
emerging technology with the Advanced Research
Projects Agency (ARPA) funding two major programming
in AI.
9
UK and European interest in expert system
The UK also has a long tradition in
computer science and AI. The “Turing Test”
(Turing 1950) has been a litmus test for
machine intelligence for many years.
The economic integration that took place
in European during the 1980s also paved
the way for European initiatives in IT
projects.
10
AI into the twenty-first century
There are several projects attracting the attention of the
AI community at present. The most notable is the CYC
project (Lenat and Guha 1991) which started in 1984.
Expert systems have brought considerable
benefits to organizations over the last 15 years.
Some generic areas are listed below:
Diagnosis (e.g the MYCIN)
Design (e.g NSPP)
Planning (e.g the DART)
Configuration(e.g the XCON)
Schduling (e.g GPSS)
11
Human expertise compared with expert system
The general advantages and disadvantages
of expert system and human expertise can
be summarized as follows
Computer advantages:
Human expertise is perishable.
Human expertise is not always consistent.
Human expertise is difficult to transfer.
Human expertise is expensive
12
Human advantages:
Humans are creative, often
inspired ;computers are uninspired.
Humans are flexible ,and easily adapt .
Humans possess common sense.
Expert system programs have capabilities
for learning that transcend those available
in conventional program.
13
Benefits of expert system
Organization advantages
Knowledge retention.
Knowledge distribution.
Training .
Competitive edge .
Cost reduction .
Advantages of user
Knowledge accessibility .
Training .
Consistency .
14
Expert system as decision support system
The main characteristics of DSS are as
follows:
Provide management with rapid access to
information that can be used in the decision-
making process .
Integrate or techniques with information
processing software .
Improve the impact of the management
decision process .
Help management deal with unstructured
problems .
15
Differences between DSS and expert
:systems
Expert system solutions are not applied to
problems requiring the mathematical
optimization techniques that are applied to
traditional DSS .
Expert systems can be applied to problems
whose objectives and constraints are
difficult to specify in quantitative terms .
Expert systems are effective for eliciting
alternatives as part of the solution process .
16
The differences between DSS and expert system are
summarized in the table
17
