Inference, inquiry and explanation in expert systems by means offuzzy neural networks

摘要

Current research interests in hybrid architectures for intelligentsystems focus on integrating the computational paradigms of expertsystems and neural networks in a manner that exploits the strengths ofboth systems and expands the applications to which either system couldbe applied individually. Such systems, which are called connectionistexpert systems, are discussed. The authors show how basic functions ofexpert systems, such as inference, inquiry and explanation, can beimplemented by means of fuzzy neural networks. They introduce somemeasures of uncertainty to be used by the connectionist inferencemachine of a classification expert system when performing thesefunctions. The emphasis is on the uncertainty processing by thedifferent elements of the network