An optically programmed neural network

Abstract
The author report on the design, construction and operation of a hybrid electrooptic computer intended for neural-network applications. They have configured this system to implement what they believe is the largest fully interconnected neural network built to date. The 14400 synapses in this network possess full analog depth and can be dynamically programmed over three decades of connection strength. The conductivity at each synapse is independently determined by the light flux it receives. A description is given of the use of amorphous silicon to implement a photoconductive array and of the building up from such an array of two neural-network configurations.