ADJOINT-OPERATORS AND NONADIABATIC LEARNING ALGORITHMS IN NEURAL NETWORKS

被引：7

作者：

TOOMARIAN, N

BARHEN, J

机构：

[1] CALTECH,JET PROP LAB,PASADENA,CA 91109

[2] CALTECH,DIV ENGN & APPL SCI,PASADENA,CA 91109

来源：

APPLIED MATHEMATICS LETTERS | 1991年 / 4卷 / 02期

基金：

美国国家航空航天局;

关键词：

D O I：

10.1016/0893-9659(91)90172-R

中图分类号：

O29 [应用数学];

学科分类号：

070104 ;

摘要：

Adjoint sensitivity equations are presented, which can be solved simultaneously (i.e., forward in time) with the dynamics of a nonlinear neural network. These equations provide the foundations for a new methodology which enables the implementation of temporal learning algorithms in a highly efficient manner.

引用

页码：69 / 73

页数：5

共 15 条

[1]

BARHEN J, 1990, IN PRESS NONLIPSCHIT

[2]

BARHEN J, 1989, IEEE COMPUT, V22, P67

[3]

BARHEN J, 1990, IN PRESS APPL MATH L

[4]

Barhen J., 1990, ADV NEURAL INFORMATI, V2, P498