Monolithically integrated optoelectronic thresholding devices for neural network applications

The realization of an optical neural network critically depends on the availability of two dimensional (2-D) arrays of neurons. These neurons are nonlinear processing elements that, in their simplest form, produce outputs which are the thresholded version of the inputs. Liquid crystal spatial light modulators are candidates for such 2-D array of neurons. They are not, however, flexible in their use. Optoelectronic integrated circuits (OEICs), either hybrid—such as liquid crystal on silicon, Si-PLZT, and flip-chip devices—or monolithic integration in III-V compounds, is another solution.