POTENTLAL APPLICATIOX OF NEURAL NETWORKS TO HAZARDOUS \ i ' ASTE PROCESSING

The concept of neural networks has increasingly gained the attention of scientists and engineers from diverse disciplines in recent years. Conventionally, a neural network comprises an ensemble of interconnected processing units (artificial neurons) with modifiable connection strength between any two units. Through appropriate learning rules, the neural network can be trained to perform tasks too difficult or impossible for a digital computer. Examples of such tasks include adaptation, pattern recognition, pattern completion, and categorization. In this work, the history of neural network technology is reviewed, and the future of the technology is discussed, Some potential applications of the technology to hazardous waste detection, " h a t i o n and processing are presented. t * * Also with Odin Corporation, Manhattan, Kansas Author to whom correspondence should be addressed Department of Electrical and Computer Engineering, Kansas State University; also with Odin Corporation