[ Expert Systems ]
n
psychology, the paradox of the expert states that the more one knows, the harder it is for
one to recall any specific information. One would think that the amount of knowledge
an old man has acquired would take him a very long time to sort through his life-time of
wisdom to figure out why his car won't start one morning. In reality, this is not
the case. Assuming that he has some knowledge of cars, that wise man would probably
be able to diagnose his car's problem in a couple of minutes and be on his way. How
does he figure out what to check or try out to see what was wrong with it when such a
complicated thing as a car could have a multitude of things that could go wrong?
When asked the same question, he said, "I just systematically checked certain things
in a certain order. First, I looked to see if there's any gas in the fuel
tank. If there is, then maybe there's a problem with the spark plug so I check that
next..."
The wise man's car example illustrated one of the ways to make decisions on what
knowledge to actually access. The man had certain rules on what to check that
prevented him from being overwhelmed by all the possible reasons that won't prevent his
car from starting. Controlling decision-making processes is another concern in AI
and has been successful in employing this type of method in what is known as expert
systems to prevent a computer from being lost in a sea of information.
As the name implies, expert systems are designed to replicate the
decision-making processes of human experts in a particular field. As a combination
of a knowledge base, decision rules, and an inference engine, an expert system uses an
elaborate set of rules that it can infer knowledge from one to another. For the
example of the non-starting car, an expert system from Ford's Service Bay Diagnostic
System would go through a set of rules:
