Implementing a multi-layer neural network using a single physical layer of analog neurons

Abstract

A system including a multi-layer analog neural network that has a single layer of physical analog neurons that is re-usable for implementing a plurality of layers of the multi-layer analog neural network. Each of the physical analog neurons is configured to receive a neuron input and to process the neuron input to generate a neuron output that is fed as input to all physical analog neurons of the single layer, and each of the physical analog neurons includes a respective weight memory. The system controller is operable to obtain, for each physical analog neuron, a respective set of neuron weight vectors with each neuron weight vector corresponding to a respective layer of the plurality of layers of the multi-layer analog neural network; store, for each physical analog neuron, the respective set of neuron weights in the respective weight memory of the physical analog neuron.