A novel approach to error function minimization for feedforward neural networks

TL;DR

This paper introduces a new method for minimizing error functions in feedforward neural networks that improves convergence monitoring and reduces ambiguities in parameter selection.

Contribution

It presents a novel approach to error minimization that enhances convergence detection and parameter robustness in neural network training.

Findings

Improved detection of convergence to the global minimum.

Reduced ambiguity in choosing minimization parameters.

Enhanced reliability of neural network training process.

Abstract

Feedforward neural networks with error backpropagation (FFBP) are widely applied to pattern recognition. One general problem encountered with this type of neural networks is the uncertainty, whether the minimization procedure has converged to a global minimum of the cost function. To overcome this problem a novel approach to minimize the error function is presented. It allows to monitor the approach to the global minimum and as an outcome several ambiguities related to the choice of free parameters of the minimization procedure are removed.