Analog capabilities of the BSB model as applied to theanti-radiation homing missile problem

摘要

The binary versus analog issues for inputs and outputs using theBSB model for the antiradiation homing sensor application, i.e.relatively unstructured data, is discussed. Using a fine-grain analog tobinary code and weak positive feedback recall, the authors find that afully interconnected BSB model produces analog as well as thetraditional outputs. The analog outputs represent the shape and themagnitude of the binary feature distributions. The feature average(binary output) and the feature distribution can be used to moreaccurately classify the source of the signals than the averageinformation alone. The authors believe that this distributiondemonstration is a first for neural network technology. They have alsofound that, under unity feedback, the fully connected BSB model iscapable of learning analog input shapes. This capability is helpful forthe radar emitter classification goal, e.g. permitting storage offrequency spectra