Enabling Spike-based Backpropagation for Training Deep Neural Network Architectures
This work addresses a key bottleneck in neuromorphic computing by enabling efficient training of deep SNNs, which could lead to more energy-efficient AI systems, though it is incremental as it builds on existing SNN methods.
The authors tackled the problem of training deep spiking neural networks (SNNs) directly with input spikes by proposing an approximate derivative method for leaky integrate-and-fire neurons, enabling spike-based backpropagation and achieving state-of-the-art classification accuracies on MNIST, SVHN, and CIFAR-10 datasets.
Spiking Neural Networks (SNNs) have recently emerged as a prominent neural computing paradigm. However, the typical shallow SNN architectures have limited capacity for expressing complex representations while training deep SNNs using input spikes has not been successful so far. Diverse methods have been proposed to get around this issue such as converting off-the-shelf trained deep Artificial Neural Networks (ANNs) to SNNs. However, the ANN-SNN conversion scheme fails to capture the temporal dynamics of a spiking system. On the other hand, it is still a difficult problem to directly train deep SNNs using input spike events due to the discontinuous, non-differentiable nature of the spike generation function. To overcome this problem, we propose an approximate derivative method that accounts for the leaky behavior of LIF neurons. This method enables training deep convolutional SNNs directly (with input spike events) using spike-based backpropagation. Our experiments show the effectiveness of the proposed spike-based learning on deep networks (VGG and Residual architectures) by achieving the best classification accuracies in MNIST, SVHN and CIFAR-10 datasets compared to other SNNs trained with a spike-based learning. Moreover, we analyze sparse event-based computations to demonstrate the efficacy of the proposed SNN training method for inference operation in the spiking domain.