A conic section function network synapse and neuron implementation in VLSI hardware

Author

Yildirim, Tiilay ; Marsland, John S.

Author_Institution

Dept. of Electr. Eng. & Electron., Liverpool Univ., UK

Volume

2

fYear

1996

fDate

3-6 Jun 1996

Firstpage

974

Abstract

An analogue VLSI design which computes radial basis function (RBF) and multilayer perceptron (MLP) propagation rules on a single chip is proposed to form a conic section function network (CSFN) synapse and neuron. This novel VLSI circuit has been designed to compute both the dot product (weighted sum) for MLP and the Euclidean distance for RBF. These two propagation rules are then aggregated to use for a CSFN synapse and neuron design. This network allows the use of bounded and unbounded decision regions, depending on the data distribution of a given application. The cdsSpice simulation results are also presented, which verifies the function of the designed synapse and neuron

Keywords

CMOS analogue integrated circuits; VLSI; feedforward neural nets; multilayer perceptrons; neural chips; CMOS IC; Euclidean distance; VLSI; cdsSpice simulation; conic section function network; dot product; multilayer perceptron; propagation rules; radial basis function network; Analog computers; Circuits; Computer networks; Electronic mail; Equations; Hardware; Intelligent networks; Neural networks; Neurons; Very large scale integration;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1996., IEEE International Conference on

Conference_Location

Washington, DC

Print_ISBN

0-7803-3210-5

Type

conf

DOI

10.1109/ICNN.1996.549029

Filename

549029