Recurrent backpropagation and the dynamical approach to adaptive neural computation

Creators: Pineda, Fernando J.

Abstract

Error backpropagation in feedforward neural network models is a popular learning algorithm that has its roots in nonlinear estimation and optimization. It is being used routinely to calculate error gradients in nonlinear systems with hundreds of thousands of parameters. However, the classical architecture for backpropagation has severe restrictions. The extension of backpropagation to networks with recurrent connections will be reviewed. It is now possible to efficiently compute the error gradients for networks that have temporal dynamics, which opens applications to a host of problems in systems identification and control.

Additional Information

© 1989 Massachusetts Institute of Technology. Received 14 March 1989; accepted 17 April 1989. Posted Online March 13, 2008. The author wishes to acknowledge very helpful discussions with Pierre Baldi, Richard Durbin, and Terrence Sejnowski. The work described in this paper was performed at the Applied Physics Laboratory, The Johns Hopkins University, sponsored by the Air Force Office of Scientific Research (AFOSR-87-354). The writing and publication of this paper was supported by the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not constitute or imply any endorsement by the United States Government or the Jet Propulsion Laboratory, California Institute of Technology.

Attached Files

Published - PINnc89.pdf

Files

PINnc89.pdf

Files (686.2 kB)

Name	Size	Download all
PINnc89.pdf md5:af8f70686adf0f7028d4fd6b34153d0b	686.2 kB	Preview Download

Additional details

	All versions	This version
Views	0	0
Downloads	0	0
Data volume	0 Bytes	0 Bytes