A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network from Convolutional Neural Network

TL;DR

This paper introduces a novel CNN-to-SNN conversion method called ECC that achieves near-zero accuracy loss with significantly less energy consumption by using short spike trains, improving efficiency in processing spatial-temporal data.

Contribution

The paper proposes the explicit current control (ECC) method for CNN-to-SNN conversion, enabling high accuracy with reduced energy use using short spike trains, and provides a practical tool for implementation.

Findings

ECC achieves near-zero accuracy loss with 256 timesteps on CIFAR datasets.

ECC reduces energy consumption compared to state-of-the-art methods.

Experimental results validate ECC's effectiveness across multiple datasets and models.

Abstract

Spiking neural networks (SNNs) offer an inherent ability to process spatial-temporal data, or in other words, realworld sensory data, but suffer from the difficulty of training high accuracy models. A major thread of research on SNNs is on converting a pre-trained convolutional neural network (CNN) to an SNN of the same structure. State-of-the-art conversion methods are approaching the accuracy limit, i.e., the near-zero accuracy loss of SNN against the original CNN. However, we note that this is made possible only when significantly more energy is consumed to process an input. In this paper, we argue that this trend of "energy for accuracy" is not necessary -- a little energy can go a long way to achieve the near-zero accuracy loss. Specifically, we propose a novel CNN-to-SNN conversion method that is able to use a reasonably short spike train (e.g., 256 timesteps for CIFAR10 images) to achieve the near-zero accuracy loss. The new conversion method, named as explicit current control (ECC), contains three techniques (current normalisation, thresholding for residual elimination, and consistency maintenance for batch-normalisation), in order to explicitly control the currents flowing through the SNN when processing inputs. We implement ECC into a tool nicknamed SpKeras, which can conveniently import Keras CNN models and convert them into SNNs. We conduct an extensive set of experiments with the tool -- working with VGG16 and various datasets such as CIFAR10 and CIFAR100 -- and compare with state-of-the-art conversion methods. Results show that ECC is a promising method that can optimise over energy consumption and accuracy loss simultaneously.