Exact and Approximate Methods for Computing the Hessian of a Feedforward Artificial Neural Network
Year: 1993
Type: article
Abstract: We present two optimization techniques based on cubic curve fitting; one based on function values and derivatives a t two previous points, the other based on derivatives a t three previous points.The latter approach is viewed from a derivative space perspective, obviating the need t o compute the vertical translation of the cubic, thus simplifying the fitting problem.We dieinonstrate the effectiveness of the second method in training neural networks on parity problems of various sizes, and compare our results to a modified Quickprop algorithm and t o gradient descent.
Authors: Craig W. Codrington, M.F. Tenorio
Institution: Purdue University West Lafayette
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1599277955
