Spiking Neural Networks (SNNs) have recently become more popular as a
biologically plausible substitute for traditional Artificial Neural Networks
(ANNs). SNNs are cost-efficient and deployment-friendly because they process
input in both spatial and temporal manners using binary spikes. However, we
observe that the information capacity in SNNs is affected by the number of
timesteps, leading to an accuracy-efficiency tradeoff. In this work, we study a
fine-grained adjustment of the number of timesteps in SNNs. Specifically, we
treat the number of timesteps as a variable conditioned on different input
samples to reduce redundant timesteps for certain data. We call our method
Spiking Early-Exit Neural Networks (SEENNs). To determine the appropriate
number of timesteps, we propose SEENN-I which uses a confidence score
thresholding to filter out the uncertain predictions, and SEENN-II which
determines the number of timesteps by reinforcement learning. Moreover, we
demonstrate that SEENN is compatible with both the directly trained SNN and the
ANN-SNN conversion. By dynamically adjusting the number of timesteps, our SEENN
achieves a remarkable reduction in the average number of timesteps during
inference. For example, our SEENN-II ResNet-19 can achieve 96.1% accuracy with
an average of 1.08 timesteps on the CIFAR-10 test dataset.

本研究探讨了一种调整 Spiking Neural Networks 中时间步数的方法 ——Spiking Early-Exit Neural Networks (SEENN)，并提出了两种确定适当时间步数的策略，SEENN-I 和 SEENN-II。我们证明 SEENN 对于直接训练 SNN 和 ANN-SNN 的转换两种情况均兼容，并通过动态调整时间步数，在推理中实现了显著减少冗余时间步数的效果。例如，在 CIFAR-10 测试数据集上，我们的 SEENN-II ResNet-19 可实现 96.1％的准确度，平均时间步数为 1.08。