Optical neural networks (ONN) based on micro-ring resonators (MRR) have emerged as a promising alternative to significantly accelerating the massive matrix-vector multiplication (MVM) operations in artificial intelligence (AI) applications. However, the limited scale of MRR arrays presents a challenge for AI acceleration. The disparity between the small MRR arrays and the large weight matrices in AI necessitates extensive MRR writings, including reprogramming and calibration, resulting in considerable latency and energy overheads. To address this problem, we propose a novel design methodology to lessen the need for frequent weight reloading. Specifically, we propose a reuse and blend (R&B) architecture to support efficient layer-wise and block-wise weight sharing, which allows weights to be reused several times between layers/blocks. Experimental results demonstrate the R&B system can maintain comparable accuracy with 69% energy savings and 57% latency improvement. These results highlight the promise of the R&B to enable the efficient deployment of advanced deep learning models on photonic accelerators.

翻译：基于微环谐振器（MRR）的光学神经网络（ONN）已成为显著加速人工智能（AI）应用中大规模矩阵向量乘法（MVM）运算的一种有前景的替代方案。然而，MRR阵列的有限规模对AI加速构成了挑战。小型MRR阵列与AI中大型权重矩阵之间的差距，需要大量的MRR写入操作（包括重新编程和校准），从而导致显著的延迟和能量开销。为解决此问题，我们提出了一种新颖的设计方法，以减少频繁的权重重载需求。具体而言，我们提出了一种复用与融合（R&B）架构，以支持高效的层间和块间权重共享，使得权重可以在层/块之间被多次复用。实验结果表明，R&B系统能够在保持相当精度的同时，实现69%的能耗节省和57%的延迟改善。这些结果凸显了R&B在光子加速器上高效部署先进深度学习模型方面的潜力。