A Silicon Photonic Accelerator for Convolutional Neural Networks with Heterogeneous Quantization

TL;DR

This paper introduces HQNNA, a silicon photonic accelerator that efficiently processes both homogeneous and heterogeneous quantized CNNs, significantly improving energy and throughput efficiency over existing photonic accelerators.

Contribution

The paper presents HQNNA, a novel silicon photonic CNN accelerator capable of handling heterogeneous quantization, enhancing accuracy and efficiency compared to prior homogeneous approaches.

Findings

Achieves up to 73.8x better energy-per-bit.

Achieves up to 159.5x better throughput-energy efficiency.

Supports both homogeneous and heterogeneous quantized CNN models.

Abstract

Parameter quantization in convolutional neural networks (CNNs) can help generate efficient models with lower memory footprint and computational complexity. But, homogeneous quantization can result in significant degradation of CNN model accuracy. In contrast, heterogeneous quantization represents a promising approach to realize compact, quantized models with higher inference accuracies. In this paper, we propose HQNNA, a CNN accelerator based on non-coherent silicon photonics that can accelerate both homogeneously quantized and heterogeneously quantized CNN models. Our analyses show that HQNNA achieves up to 73.8x better energy-per-bit and 159.5x better throughput-energy efficiency than state-of-the-art photonic CNN accelerators.