Compressed Meta-Optical Encoder for Image Classification
This work addresses the problem of high latency and power consumption in optical/hybrid CNNs for image classification, though it is incremental as it builds on existing knowledge distillation and meta-optics techniques.
The authors tackled the challenge of implementing optical nonlinearity in hybrid convolutional neural networks by using knowledge distillation to compress a modified AlexNet to a single linear convolutional layer with an electronic backend, achieving comparable performance to a purely electronic CNN. They experimentally demonstrated over two orders of magnitude reduction in multiply-accumulate operations (from 17M to 86K) and classification accuracy exceeding 93% on MNIST.
Optical and hybrid convolutional neural networks (CNNs) recently have become of increasing interest to achieve low-latency, low-power image classification and computer vision tasks. However, implementing optical nonlinearity is challenging, and omitting the nonlinear layers in a standard CNN comes at a significant reduction in accuracy. In this work, we use knowledge distillation to compress modified AlexNet to a single linear convolutional layer and an electronic backend (two fully connected layers). We obtain comparable performance to a purely electronic CNN with five convolutional layers and three fully connected layers. We implement the convolution optically via engineering the point spread function of an inverse-designed meta-optic. Using this hybrid approach, we estimate a reduction in multiply-accumulate operations from 17M in a conventional electronic modified AlexNet to only 86K in the hybrid compressed network enabled by the optical frontend. This constitutes over two orders of magnitude reduction in latency and power consumption. Furthermore, we experimentally demonstrate that the classification accuracy of the system exceeds 93% on the MNIST dataset.