A New Leasing Algorithm for Multi-Layered Neural Networks Based on Adaptive Algorithms With Implementation by Hopfield Networks

Abstract

This paper presents a new learning algorithm for multi-layered neural networks and its neural implementation. By formulating the learning problem in multi-layered neural networks as the set of least-squares equations with interconnected variables, a multi-stage adaptive algorithm is designed to solve the equations. Moreover, the Hopfield networks can be adopted to implement the proposed algorithm. The proposed algorithm may be seen as a 'Learning Algorithm for Neural Networks by Neural Networks' and provides an alternative for training neural networks in some applications.