All-optical Neural Network Research Articles

AN ALL-OPTICAL HOPFIELD NETWORK: THEORY AND EXPERIMENT

The implementation of an all-optical Hopfield-type neural network is made possible by eliminating the need for subtracting light intensities. This implementation is achieved without significantly degrading the network's performance, when only inhibitory connections (i.e., Jij <0) are used. The theoretical analysis of such a network, and its experimental implementation, exploiting a liquid crystal light valve for the neurons and an array of subholograms for the interconnections, are presented.

All-optical neural network with inhibitory neurons

A Hopfield-type neural network with only inhibitory interconnections is shown to perform similarly to networks with both inhibitory and excitatory interconnects. With the use of only inhibitory neurons the need to subtract light intensities is eliminated, so an all-optical neural network can be realized. This realization is achieved with a liquid-crystal light valve for implementing a two-dimensional array of 16 inhibitory neurons and an array of subholograms for the interconnections. Two stable states that can be retrieved from noisy optical inputs are demonstrated.