Shrinking Your TimeStep: Towards Low-Latency Neuromorphic Object Recognition with Spiking Neural Network

TL;DR

This paper introduces the Shrinking SNN (SSNN), a novel approach that reduces inference latency in neuromorphic object recognition by dividing SNNs into stages with shrinking timesteps, maintaining high accuracy at low latency.

Contribution

The paper proposes a new multi-stage SNN architecture with temporal shrinking and early classifiers to achieve low-latency recognition without performance loss.

Findings

Improves baseline accuracy by up to 21.41% on neuromorphic datasets.

Achieves 73.63% accuracy on CIFAR10-DVS with only 5 timesteps.

Effectively reduces latency while maintaining high recognition performance.

Abstract

Neuromorphic object recognition with spiking neural networks (SNNs) is the cornerstone of low-power neuromorphic computing. However, existing SNNs suffer from significant latency, utilizing 10 to 40 timesteps or more, to recognize neuromorphic objects. At low latencies, the performance of existing SNNs is drastically degraded. In this work, we propose the Shrinking SNN (SSNN) to achieve low-latency neuromorphic object recognition without reducing performance. Concretely, we alleviate the temporal redundancy in SNNs by dividing SNNs into multiple stages with progressively shrinking timesteps, which significantly reduces the inference latency. During timestep shrinkage, the temporal transformer smoothly transforms the temporal scale and preserves the information maximally. Moreover, we add multiple early classifiers to the SNN during training to mitigate the mismatch between the surrogate gradient and the true gradient, as well as the gradient vanishing/exploding, thus eliminating the performance degradation at low latency. Extensive experiments on neuromorphic datasets, CIFAR10-DVS, N-Caltech101, and DVS-Gesture have revealed that SSNN is able to improve the baseline accuracy by 6.55% ~ 21.41%. With only 5 average timesteps and without any data augmentation, SSNN is able to achieve an accuracy of 73.63% on CIFAR10-DVS. This work presents a heterogeneous temporal scale SNN and provides valuable insights into the development of high-performance, low-latency SNNs.