ral Networks Expan orizons

tatistical signal processing covers an area where physics and mathematics meet and interact to solve a wide range of problems. Its origins may be traced back to the 1943 classified RCA report by North, republished in [l], the 1946 classic paper [Z] by Van Vleck and Middleton, and the pioneering work by Wiener [3] . In particular, the classical methods of statistical signal processing are founded on three basic assumptions: linearity, stationarity, and second-order statistics with particular emphasis on Gaussianity. These assumptions are invoked for the sake of mathematical tractability. Yet most, if not all, the physical signals that we have to deal with in real-life applications are generated by dynamic processes that are simultaneously nonlinear, nonstationary, and non-Gaussian. The end result of designing a signal-processing system along traditional lines is a suboptimal solution. One way in which the performance of the system can be improved is to consider the use of neural networks in combination with other suitable techniques (e.g., time-frequency analysis), depending on the task at hand. Interest in neural networks, or to be more precise, artificial neural networks, has always been motivated by the fact that the human brain functions in a manner entirely different from the conventional digital computer. The human brain is a gigantic, and yet highly efficient, information-processing machine that encompasses a wide variety of complex signal processing operations. To appreciate the enormous scale of these operations, we need only look at our visual and auditory systems and be amazed at the “seamless” nature of the way in which different forms of information gathered by our eyes and ears are individually processed and then finally fused together. Work on neural networks may be traced back to the pioneering paper [4] by McCulloch and Pitts in 1943, which was followed by Rosenblatt’s development of the perceptron [5] and Widrow’s development of the adaline [6] in the late 1950s. After going through a period of dormancy (in an engineering context) in the 1970s, neural networks reemerged in the 1980s with the publication of Hopfield’s