Sigmoid generators for neural computing using piecewise approximations
- 1 January 1996
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 45 (9), 1045-1049
- https://doi.org/10.1109/12.537127
Abstract
A piecewise second order approximation scheme is proposed for computing the sigmoid function. The scheme provides high performance with low implementation cost; thus, it is suitable for hardwired cost effective neural emulators. It is shown that an implementation of the sigmoid generator outperforms, in both precision and speed, existing schemes using a bit serial pipelined implementation. The proposed generator requires one multiplication, no look-up table and no addition. It has been estimated that the sigmoid output is generated with a maximum computation delay of 21 bit serial machine cycles representing a speedup of 1.57 to 2.23 over other proposals.Keywords
This publication has 6 references indexed in Scilit:
- Finite precision error analysis of neural network electronic hardware implementationsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Sigmoid generators for neural computing using piecewise approximationsIEEE Transactions on Computers, 1996
- Implementation issues of sigmoid function and its derivative for VLSI digital neural networksIEE Proceedings E Computers and Digital Techniques, 1992
- Simple approximation of sigmoidal functions: realistic design of digital neural networks capable of learningPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1991
- Efficient implementation of piecewise linear activation function for digital VLSI neural networksElectronics Letters, 1989
- Parallel Distributed ProcessingPublished by MIT Press ,1986