Efficient digital implementation of the sigmoid function for reprogrammable logic
Author:
- 1 January 2003
- journal article
- Published by Institution of Engineering and Technology (IET) in IEE Proceedings - Computers and Digital Techniques
- Vol. 150 (6) , 403-411
Abstract
Special attention must be paid to an efficient approximation of the sigmoid function in implementing FPGA-based reprogrammable hardware-based artificial neural networks. Four previously published piecewise linear and one piecewise second-order approximation of the sigmoid function are compared with SIG-sigmoid, a purely combinational approximation. The approximations are compared in terms of speed, required area resources and accuracy measured by average and maximum error. It is concluded that the best performance is achieved by SIG-sigmoid.Keywords
This publication has 10 references indexed in Scilit:
- Neural network adaptations to hardware implementations2004
- Neural networks in FPGAs2004
- Digital design of sigmoid approximator for artificial neural networksElectronics Letters, 2002
- Digital Signal Processing with Field Programmable Gate Arrays2001
- Sigmoid generators for neural computing using piecewise approximationsIEEE Transactions on Computers, 1996
- Finite precision error analysis of neural network hardware implementationsIEEE Transactions on Computers, 1993
- Simple sigmoid-like activation function suitable for digital hardware implementationElectronics Letters, 1992
- Implementing nonlinear activation functions in neural network emulatorsElectronics Letters, 1991
- Efficient implementation of piecewise linear activation function for digital VLSI neural networksElectronics Letters, 1989
- McBOOLE: A New Procedure for Exact Logic MinimizationIEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 1986