Second order derivatives for network pruning: Optimal Brain Surgeon

19921,491 citationsJournal Articlegreen Open Access

Authors

Abstract

We investigate the use of information from all second order derivatives of the error function to perform network pruning (i.e., removing unimportant weights from a trained network) in order to improve generalization and increase the speed of further training.

Topics & Keywords

Advanced Memory and Neural Computing Neural Networks and Applications Ferroelectric and Negative Capacitance Devices

Publication Details

Published in: CaltechAUTHORS (California Institute of Technology)

Volume 5, pp. 164-171

Field-Weighted Citation Impact: 21.61

Command Palette

Second order derivatives for network pruning: Optimal Brain Surgeon

Authors

Abstract

Topics & Keywords

Publication Details