Optical implementation of a second-order neural network

Abstract
An optical implementation of a single-layer, second-order neural network is presented. The quadratic products are obtained by passing the optical beam twice through the input ferroelectric liquid-crystal (FLC) spatial light modulator (SLM), with the interconnection weights being implemented by a further two-dimensional 128 × 128 FLC SLM. The machine successfully associates eight randomly chosen pattern–target pairs (dimensions 16 and 4, respectively) and can learn the parity association. Translation invariance is also demonstrated. Results from a computer model indicate that input SLM contrast ratios of 4:1 and electronic noise of 10% of the maximum output can be tolerated.