Neural network architecture for concurrent learning with antidromic spikes

Abstract

A neural network processing system having multiple layers is provided. Each layer includes a bidirectional Synaptic Network Channel (SNC) for concurrently transmitting weighted sums, y(t)'s and x(t)'s, as an elastic wave superposition of inputs, x(t)'s and y(t)'s, respectively. Each input is multiplied and added with corresponding weights w's encoded in variable splitters and combiners in forward and backward directions, respectively. Each layer includes unidirectional Signal Reshaping (SR) units, I's and L's for inference and learning, respectively, by generating inputs for a following layer in forward and backward directions from a current layer's weighted sums y(t)'s and x(t)'s, respectively. Each layer includes a Hybrid Coupler (HC) to connect the bidirectional SNC and the unidirectional SR units. Each layer includes a weight update unit to calculate each weight difference Δwusing an input y(t) or a weighted sum y(t) and an input x(t) to update a weight wfor a current layer.