Training Energy-Efficient Deep Spiking Neural Networks with Single-Spike Hybrid Input Encoding

TL;DR

This paper introduces a hybrid input encoding and a novel training method for deep spiking neural networks, significantly reducing inference latency and energy consumption while maintaining high accuracy on image classification tasks.

Contribution

It proposes a hybrid input encoding scheme and a spike timing based training method that improve efficiency and reduce latency in deep SNNs.

Findings

Achieved 66.46% top-1 accuracy on CIFAR-100 with 5 timesteps.

Reduced energy consumption by approximately 125 times compared to standard ANNs.

Inference speed improved by 5 to 300 times over existing SNN models.

Abstract

Spiking Neural Networks (SNNs) have emerged as an attractive alternative to traditional deep learning frameworks, since they provide higher computational efficiency in event driven neuromorphic hardware. However, the state-of-the-art (SOTA) SNNs suffer from high inference latency, resulting from inefficient input encoding and training techniques. The most widely used input coding schemes, such as Poisson based rate-coding, do not leverage the temporal learning capabilities of SNNs. This paper presents a training framework for low-latency energy-efficient SNNs that uses a hybrid encoding scheme at the input layer in which the analog pixel values of an image are directly applied during the first timestep and a novel variant of spike temporal coding is used during subsequent timesteps. In particular, neurons in every hidden layer are restricted to fire at most once per image which increases activation sparsity. To train these hybrid-encoded SNNs, we propose a variant of the gradient descent based spike timing dependent back propagation (STDB) mechanism using a novel cross entropy loss function based on both the output neurons' spike time and membrane potential. The resulting SNNs have reduced latency and high activation sparsity, yielding significant improvements in computational efficiency. In particular, we evaluate our proposed training scheme on image classification tasks from CIFAR-10 and CIFAR-100 datasets on several VGG architectures. We achieve top-1 accuracy of $66.46$\% with $5$ timesteps on the CIFAR-100 dataset with ${\sim}125\times$ less compute energy than an equivalent standard ANN. Additionally, our proposed SNN performs $5$-$300\times$ faster inference compared to other state-of-the-art rate or temporally coded SNN models.