Photonic Integrated Neuro-Synaptic Core for Convolutional Spiking Neural Network

TL;DR

This paper introduces a photonic neuro-synaptic chip that integrates linear weighting and nonlinear spiking activation, enabling scalable neuromorphic photonic neural networks with high recognition accuracy.

Contribution

It presents a novel integrated photonic chip that combines key neural functions using a DFB laser with a saturable absorber, advancing large-scale photonic neural network implementation.

Findings

Achieved 87% accuracy on MNIST dataset.

Demonstrated parallel weighted function and spike activation.

Fabricated a four-channel DFB-SA array for matrix convolution.

Abstract

Neuromorphic photonic computing has emerged as a competitive computing paradigm to overcome the bottlenecks of the von-Neumann architecture. Linear weighting and nonlinear spiking activation are two fundamental functions of a photonic spiking neural network (PSNN). However, they are separately implemented with different photonic materials and devices, hindering the large-scale integration of PSNN. Here, we propose, fabricate and experimentally demonstrate a photonic neuro-synaptic chip enabling the simultaneous implementation of linear weighting and nonlinear spiking activation based on a distributed feedback (DFB) laser with a saturable absorber (DFB-SA). A prototypical system is experimentally constructed to demonstrate the parallel weighted function and nonlinear spike activation. Furthermore, a four-channel DFB-SA array is fabricated for realizing matrix convolution of a spiking convolutional neural network, achieving a recognition accuracy of 87% for the MNIST dataset. The fabricated neuro-synaptic chip offers a fundamental building block to construct the large-scale integrated PSNN chip.