High‐Throughput Multichannel Parallelized Diffraction Convolutional Neural Network Accelerator. Issue 12 (3rd October 2022)

• Record Type: Journal Article
• Authors: Hu, Zibo; Li, Shurui; Schwartz, Russell L. T.; Solyanik‐Gorgone, Maria; Miscuglio, Mario; Gupta, Puneet; Sorger, Volker J.
• Is Part Of: Laser & photonics reviews. Volume 16:Issue 12(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-10-03
• Subjects: accelerators -- convolution -- diffraction -- digital micromirror devices -- neural networks -- parallelization; Lasers -- Periodicals; Photonics -- Periodicals; Lasers -- Périodiques; Photonique -- Périodiques; 621.36
• DOI: 10.1002/lpor.202200213
• Languages: English
• ISSNs: 1863-8880

Abstract

Abstract: Convolutional neural networks are paramount in image and signal processing, and are responsible for the majority of image recognition power consumption today, concentrated mainly in convolution computations. With convolution operations being computationally intensive, next‐generation hardware accelerators need to offer parallelization and high efficiency. Diffractive optics offers the promise of low‐latency, highly parallel convolution operations. However, thus far parallelism is only partially harvested, thereby significantly underdelivering in comparison to its throughput potential. Here, a parallelized operation high‐throughput Fourier optic convolutional accelerator is demonstrated. For the first time, simultaneous processing of multiple kernels in Fourier domain enabled by optical diffraction orders is achieved alongside input parallelism. The proposed approach can offer ≈100× speedup over the previous generation optical diffraction‐based processor and 10× speedup over other state‐of‐the‐art optical Fourier classifiers. Abstract : A digital micromirror device (DMD)‐based efficient multichannel Fourier optics convolution accelerator is demonstrated. For the first time, parallelized operations are achieved by exploiting the higher diffraction orders. It provides a 100× speedup over the previous single‐channel system by utilizing the properties of light more efficiently, without the need for expensive optical components. This approach can benefit most free‐space systems.