Scalability of All-Optical Neural Networks Based on Spatial Light Modulators

Optical implementation of artificial neural networks has been attracting great attention due to its potential in parallel computation at speed light. Although all-optical deep (AODNNs) with a few neurons have experimentally demonstrated acceptable errors recently, the feasibility large-scale AODNNs remains unknown because error might accumulate inevitably increasing number and connections. Here, we demonstrate scalable AODNN programmable linear operations tunable nonlinear activation functions. We verify scalability by measuring analyzing propagating from single neuron entire network. The is further confirmed recognizing handwritten digits fashions, respectively.