Si microring resonator crossbar array for on-chip inference and training of optical neural network

TL;DR

This paper introduces a scalable, power-efficient silicon photonic neural network using a microring resonator crossbar array, capable of on-chip inference and training, addressing previous limitations in scalability and power consumption.

Contribution

The paper presents a fully integrated 4x4 MRR crossbar array prototype and demonstrates on-chip backpropagation for training, advancing silicon photonic ONNs.

Findings

Successful demonstration of a 4x4 MRR crossbar array for matrix-vector multiplication.

On-chip backpropagation enabled for training the optical neural network.

Potential for scalable, power-efficient optical deep learning accelerators.

Abstract

Deep learning is one of the most advancing technologies in various fields. Facing the limits of the current electronics platform, optical neural networks (ONNs) based on Si programmable photonic integrated circuits (PICs) have attracted considerable attention as a novel deep learning scheme with optical-domain matrix-vector multiplication (MVM). However, most of the proposed Si programmable PICs for ONNs have several drawbacks such as low scalability, high power consumption, and lack of frameworks for training. To address these issues, we have proposed a microring resonator (MRR) crossbar array as a Si programmable PIC for an ONN. In this article, we present a prototype of a fully integrated 4 ${\rm \times}$ 4 MRR crossbar array and demonstrated a simple MVM and classification task. Moreover, we propose on-chip backpropagation using the transpose matrix operation of the MRR crossbar array, enabling the on-chip training of the ONN. The proposed ONN scheme can establish a scalable, power-efficient deep learning accelerator for applications in both inference and training tasks.