Evaluation of gradient descent learning algorithms with an adaptive local rate technique for hierarchical feed forward architectures

Gradient descent learning algorithms (namely Back Propagation and Weight Perturbation) can significantly increase their classification performances adopting a local and adaptive learning rate management approach. In this paper, we present the results of the comparison of the classification performance of the two algorithms in a tough application: quality control analysis in steel industry. The feed forward network is hierarchically organized (i.e. tree of Multi Layer Perceptrons). The comparison has been performed starting from the same operating conditions (i.e. network topology, stopping criterion, etc): the results show that the probability of correct classification is significantly better for the Weight Perturbation algorithm.