Expert System & Neural Network

Overview Expert systems and neural networks are truly amazing technologies. Both offer possibilities for developing more useful information systems, but expert systems technology is generally regarded as simpler and more widely used than neural network technology. An expert system uses sets of rules and data to produce a decision or recommendation. Neural networks, on the other hand, attempt to simulate the human brain by collecting and processing data for the purpose of “remembering” or “learning”. The primary difference between an expert system and a neural network is that a neural network can adapt its criteria to better match the data it analyzes, while an expert system produces results without adjusting for changes in the analyzed data. Many resources for both technologies can be found on the Internet, from simple explanations to elaborate demonstrations.

Expert Systems For a quick overview of expert systems, connect to the CLIPS site at clipsrules.sourceforge.net/  Acquired Intelligence, Inc. has Web-based demonstrations, including an expert system that will help you to identify whales (www.aiinc.ca/demos/whale.html).You’ll find excellent descriptions of medical expert systems at the Medical Expert Systems site (www.pubmedcentral.nih.gov/). The Knowledge-Base Projects and Groups site at www.cs.utexas.edu/users/mfkb/related.html provides links to resources on knowledge bases.

Neural Networks Neural network technology requires massive computer capacity. Therefore, practical applications using this technology have been limited. However, it is a fascinating technology to consider--imagine computers that function like human brains! “An Introduction to Neural Networks” by Leslie Smith (www.cs.stir.ac.uk/~lss/NNIntro/InvSlides.html) is a good place to begin exploring this technology. Also visit the WEBSOM: Self-Organizing Map for Internet Exploration site for a demonstration of an artificial neural network model called “The Self-Organizing Map” (SOM). The demonstration allows you to explore over one million documents collected from more than 80 Usenet newsgroups. Is this type of neural network useful? Consider the following: current estimates indicate that the Web contains over one billion pages. Web search engines index only about 55 percent of those pages at best. You will find that WEBSOM, with its intuitive graphical interface, makes exploring Usenet documents enjoyable.