Implementation of rule-based expert systems for time-critical applications using neural networks

The process of constructing a neural-network-based expert system from a given AND/OR inference net is examined. The major purpose of such an approach is to make use of the parallel-processing capabilities of neural networks in implementing expert systems for classifications. In order to provide a clear view of the design process of the proposed neural network expert systems an example of such a classification expert system and its neural network version are given. It is clear from the results that neural networks can be used as an implementational vehicle for expert systems for time-critical applications.<<ETX>>