IMPLEMENTATION ISSUES OF SIGMOID FUNCTION AND ITS DERIVATIVE FOR VLSI DIGITAL NEURAL NETWORKS

被引：16

作者：

MURTAGH, P

TSOI, AC

机构：

来源：

IEE PROCEEDINGS-E COMPUTERS AND DIGITAL TECHNIQUES | 1992年 / 139卷 / 03期

关键词：

NEURAL NETWORKS; SIGMOID FUNCTION;

D O I：

10.1049/ip-e.1992.0033

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

This paper proposes a number of different implementations for the first derivative of the sigmoid function. The implementation of the sigmoid function employs a powers-of-two piecewise linear approximation. The best implementation scheme for the derivative is suggested based on overall speed performance (circuit speed and training time) and hardware requirements.

引用

页码：207 / 214

页数：8

共 18 条

[1]

ALIPPI C, 1991, P IEEE INT S CIRC SY, P1505

[2]

[Anonymous], 1991, INTRO THEORY NEURAL, DOI DOI 10.1201/9780429499661

[3]

[Anonymous], 1989, ANALOG VLSI NEURAL S

[4]

BAILEY J, 1988, P IEEE INT C NEURAL, V2, P173

[5]

BERGMANN NW, 1990, UNPUB INT J PARA SEP

[6]

DENYER P, 1985, VLSI SIGNAL PROCESSI

[7]

DEPRETTERE E, 1984, P INT C ACOUST SPEEC

[8]

KUNG SY, 1988, VLSI ARRAY PROCESSOR

[9] 2S COMPLEMENT PIPELINE MULTIPLIERS [J].

LYON, RF .

IEEE TRANSACTIONS ON COMMUNICATIONS, 1976, 24 (04) :418-425

[10]

McClelland J. L., 1988, EXPLORATIONS PARALLE, P121

← 1 2 →