Monthly
288 pp. per issue
6 x 9, illustrated
ISSN
0899-7667
E-ISSN
1530-888X
2014 Impact factor:
2.21

Neural Computation

May 2013, Vol. 25, No. 5, Pages 1277-1301
(doi: 10.1162/NECO_a_00436)
© 2013 Massachusetts Institute of Technology
Semistability of Steepest Descent with Momentum for Quadratic Functions
Article PDF (320.24 KB)
Abstract

Two steepest-descent algorithms with momentum for quadratic functions are considered. For a given learning rate, the sufficient and necessary conditions for the semistability of the steepest-descent algorithms with momentum are established. Moreover, the optimal momentum factors that generally make for the fastest convergence are obtained.