An analog neural network circuit with a learning rule via simultaneous perturbation

Abstract

This paper proposes a learning rule of neural networks and describes an analog feedforward neural network circuit using the learning rule. The learning rule used is a stochastic gradient-like algorithm via a simultaneous perturbation. The learning rule requires only forward operations of the neural network. Therefore, it is suitable for hardware implementation. We describe details of the fabricated neural network circuit. The exclusive-OR problem and the TCLX problem are considered. In a fabricated analog neural network circuit, the input, output and weights are realized by voltages.