Analogue neural networks with distributed neurons

Authors:
Publisher website
Google Scholar
Abstract
A new technique is proposed for implementing neurons in analogue electronic neural networks. It features integrating a fraction of a neuron with each synapse. Each neuron is distributed over all associated synapses. This methodology operates with just one building block, the distributed-neuron synapse, in contrast to existing techniques which are based on two distinct blocks, the neuron and synapse. This approach can be applied to a variety of implementations, including transconductor and MOSFET-op-amp neural networks. Modularity, ease of interconnectivity, expandability and reconfigurability are the advantages of this technique.

This publication has 1 reference indexed in Scilit: