High Throughput Multi-Channel Parallelized Diffraction Convolutional Neural Network Accelerator
This work addresses the computational bottleneck in convolutional neural networks for image and signal processing by enabling faster hardware acceleration, though it is incremental as it builds on existing optical diffraction methods.
The paper tackled the slow processing speed of optical diffraction-based convolutional neural network accelerators by developing a high-throughput Fourier optic system that processes multiple kernels simultaneously. The result is a system that is about 100 times faster than existing optical processors and rivals modern electronic solutions, capable of processing large-scale matrices about 10 times faster than state-of-the-art electronic systems.
Convolutional neural networks are paramount in image and signal processing including the relevant classification and training tasks alike and constitute for the majority of machine learning compute demand today. With convolution operations being computationally intensive, next generation hardware accelerators need to offer parallelization and algorithmic-hardware homomorphism. Fortunately, diffractive display optics is capable of million-channel parallel data processing at low latency, however, thus far only showed tens of Hertz slow single image and kernel capability, thereby significantly underdelivering from its performance potential. Here, we demonstrate an operation-parallelized high-throughput Fourier optic convolutional neural network accelerator. For the first time simultaneously processing of multiple kernels in Fourier domain enabled by optical diffraction has been achieved alongside with already conventional in the field input parallelism. Additionally, we show an about one hundred times system speed up over existing optical diffraction-based processors and this demonstration rivals performance of modern electronic solutions. Therefore, this system is capable of processing large-scale matrices about ten times faster than state of art electronic systems.