A Computation Modification for Multi-layered Neural Network Using Extended Kalman Filter

Abstract

A lot of learning algorithms for deep layered network are sincerely suffered from complex computation and slow convergence because of a very large number of free parameters. We need to develop an efficient algorithm for deep neural network. The Kalman filter concept can be applied to parameter estimation of neural network to improve computation performance. The algorithms based extended Kalman filter has a serious drawback in its computational complexity. We discuss how a fast algorithm should be developed for reduction in computation time.