Expert System is a class of software systems that emulates the knowledge and behavior of human experts to solve a problem. The expert system has a long history in the field of computing since 1970. It was hyped in a big way, in its initial days of evolution but could not keep its promises in the subsequent years. With successful developments in computational intelligence technologies, expert system and other AI systems have made a comeback to find a meaningful place in information systems, which are becoming more and more intelligent with every passing day. This article provides a refreshing introduction to expert system and discusses how to develop it using JESS.
Expert System is a branch of Artificial
Intelligence (AI). It was developed by
researchers during 1970s and applied
commercially throughout the 1980s. Expert
system is a computer program that emulates
the decision-making of a human expert in
narrow domain of expertise such as
diagnosing a heart disease, predicting a stock
in the stock market, etc. Expert system
reasons by heuristics or approximate
methods. It explains and justifies solutions
in user-friendly terms. One of the early
expert systems with a visible success story
was MYCIN, a program for diagnosing
bacterial infections in blood cells. Most of
the successful expert systems were built
around rules for medical diagnosis,
engineering, science, and business
applications. Expert systems had a number
of perceived advantages over human experts.
For instance, unlike people, they could
perform at the peak of efficiency, 24 hours a day, throughout the year without fatigue.
There are numerous dramatic examples in the computer science literature of these early
systems matching or exceeding the performance of their human counterparts in specific
and limited situations. Predictions were made that sophisticated expert systems would
be able to reproduce general human intelligence and problem-solving abilities. Over
time, the hype receded, and it was understood that researchers had vastly underestimated
the complexity of the common-sense knowledge that underpins general human reasoning.
Nevertheless, excellent applications for expert systems remain to see this day. |