Photonic Spiking Neural Networks with Highly Efficient Training Protocols for Ultrafast Neuromorphic Computing Systems

Abstract

Photonic technologies offer great prospects for novel, ultrafast, energy-efficient, and hardware-friendly neuromorphic (brain-like) computing platforms. Moreover, neuromorphic photonic approaches based on ubiquitous, technology-mature, and low-cost vertical-cavity surface-emitting lasers (VCSELs) (devices found in fiber-optic transmitters, mobile phones, and automotive sensors) are of particular interest. Given that VCSELs have shown the ability to realize neuronal optical spiking responses (at ultrafast GHz rates), their use in spike-based information-processing systems has been proposed. In this study, spiking neural network (SNN) operation, based on a hardware-friendly photonic system of just one VCSEL, is reported alongside a novel binary weight “significance” training scheme that fully capitalizes on the discrete nature of the optical spikes used by the SNN to process input information. The VCSEL-based photonic SNN was tested with a highly complex multivariate classification task (MADELON) before its performance was compared using a traditional least-squares training method and an alternative novel binary weighting scheme. Excellent classification accuracies of >94% were achieved by both training methods, exceeding the benchmark performance of the dataset in a fraction of the processing time. The newly reported training scheme also dramatically reduces the training set size requirements and the number of trained nodes (≤1% of the total network node count). This VCSEL-based photonic SNN, in combination with the reported “significance” weighting scheme, therefore grants ultrafast spike-based optical processing highly reduced training requirements and hardware complexity for potential application in future neuromorphic systems and artificial intelligence applications.