Learned classification of sonar targets using a massively parallel network

Abstract

We have applied massively parallel learning networks to the classification of sonar returns from two undersea targets and have studied the ability of networks to correctly classify both training and testing examples. Networks with an intermediate layer of hidden pro- cessing units achieved a classification accuracy as high as 100 percent on a training set of 104 returns. These networks correctly classified a test set of 104 returns not contained in the training set with an accuracy of up to 90.4 percent. Networks without an intermediate layer of pro- cessing units achieved only 73.1 percent correct on the same test set. Performance improved and the variability due to the initial conditions for training decreased with the number of hidden units. The effect of training set design on test set performance was also examined. The performance of a three-layered network was better than trained hu- man listeners and the network generalized better than a nearest neigh- bor classifier.