Title :
Hardware realisation of a neuron transfer function and its derivative
Author_Institution :
MESA Res. Inst., Twente Univ., Enschede
fDate :
3/31/1994 12:00:00 AM
Abstract :
In on-chip learning feedforward neural networks, both nonlinear sigmoid-like function and its derivative are required. A simple CMOS circuit, based on a CMOS differential pair in weak inversion, that realises both functions is presented
Keywords :
CMOS integrated circuits; feedforward neural nets; learning (artificial intelligence); CMOS circuit; CMOS differential pair; hardware realisation; neuron transfer function; nonlinear sigmoid-like function; on-chip learning feedforward neural networks; weak inversion;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19940375
