A PROGRAMMABLE CMOS GAUSSIAN SYNAPSE FOR ANALOG VLSI NEURAL NETWORKS

Feedforward neural networks with a single hidden layer using a gaussia n function have been proven to have the capability of universal approx imation in a satisfactory sense. Back-propagation neural networks with gaussian function synapses have a better convergence property over th ose with linear multiplying synapses. A programmable gaussian synapse for analogue VLSI neural networks with hardware implementation and the programming techniques to program the cell are presented. The standar d deviation and the magnitude of the gaussian synapse can be programme d externally. The proposed gaussian synapse was designed with single-e nded inputs. To verify the programmability of the proposed gaussian sy napse, a prototype chip was fabricated using a 1.2 mu m CMOS process a nd the experimental results obtained are presented.