Title :
Circuit design of an adjustable neuron activation function and its derivative
Author :
Lu, Chun ; Shi, Bingxue
Author_Institution :
Inst. of Microelectron., Tsinghua Univ., Beijing, China
fDate :
3/16/2000 12:00:00 AM
Abstract :
A novel simple neuron circuit is presented. It can realise both the sigmoidal activation function and its derivative. The neuron can be adjusted to suit a wide range of conditions by altering the threshold and gain factor of the activation function
Keywords :
analogue processing circuits; learning (artificial intelligence); neural chips; adjustable neuron activation function; gain factor; neuron circuit; sigmoidal activation function; threshold;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20000401
