Emerged as a biology-inspired method, Spiking Neural Networks (SNNs) mimic the spiking nature of brain neurons and have received lots of research attention. SNNs deal with binary spikes as their activation and therefore derive extreme energy efficiency on hardware. However, it also leads to an intrinsic obstacle that training SNNs from scratch requires a re-definition of the firing function for computing gradient. Artificial Neural Networks (ANNs), however, are fully differentiable to be trained with gradient descent. In this paper, we propose a joint training framework of ANN and SNN, in which the ANN can guide the SNN's optimization. This joint framework contains two parts: First, the knowledge inside ANN is distilled to SNN by using multiple branches from the networks. Second, we restrict the parameters of ANN and SNN, where they share partial parameters and learn different singular weights. Extensive experiments over several widely used network structures show that our method consistently outperforms many other state-of-the-art training methods. For example, on the CIFAR100 classification task, the spiking ResNet-18 model trained by our method can reach to 77.39% top-1 accuracy with only 4 time steps.

翻译：[翻译摘要] 作为受生物学启发的方法，脉冲神经网络（SNNs）模拟了脑神经元的脉冲特性，并获得了大量研究关注。SNN以二进制脉冲作为激活值，因而在硬件上具有极高的能效。然而，这也带来了根本性障碍：从头训练SNN需要重新定义用于梯度计算的脉冲函数。而人工神经网络（ANN）则完全可微，可通过梯度下降法进行训练。本文提出了一种ANN与SNN的联合训练框架，其中ANN可引导SNN的优化过程。该联合框架包含两部分：首先，通过网络的多个分支将ANN内部知识蒸馏到SNN中；其次，我们对ANN和SNN的参数进行约束，使其共享部分参数而学习不同的奇异权重。在多种广泛使用的网络结构上进行的大量实验表明，我们的方法持续优于其他多种最先进的训练方法。例如，在CIFAR100分类任务中，使用我们的方法训练的脉冲ResNet-18模型仅需4个时间步即可达到77.39%的top-1准确率。