Monolithically Integrated Optical Convolutional Processors on Thin Film Lithium Niobate

Abstract

Photonic neural networks (PNNs) of sufficiently large physical dimensions and high operation accuracies are envisaged as an ideal candidate for breaking the major bottlenecks in the current artificial intelligence architectures in terms of latency, energy efficiency and computational power. To achieve this vision, it is of vital importance to scale up the PNNs and in the meantime reduce the high demand on the dimensions required by the PNNs. The underlying cause of this strategy is the enormous gap between the scales of photonic and electronic integrated circuits. Here, we demonstrate monolithically integrated optical convolutional processors on thin film lithium niobate (TFLN) to enable large-scale programmable convolution kernels and in turn greatly reduce the dimensions required by the subsequent fully connected layers. Experimental validation achieves high classification accuracies of 96%/86% on the MNIST/Fashion-MNIST datasets and 84.6% on the AG News dataset, while dramatically reducing the required subsequent fully connected layer dimensions to 196x10 (from 784x10) and 175x4 (from 800x4), respectively. Furthermore, our devices can be driven by commercial field-programmable gate array (FPGA) systems, a unique advantage in addition to their scalable channel number and kernel size, our architecture provides a solution to build practical machine learning photonic devices.