Cargando…
A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network From Convolutional Neural Network
This article conforms to a recent trend of developing an energy-efficient Spiking Neural Network (SNN), which takes advantage of the sophisticated training regime of Convolutional Neural Network (CNN) and converts a well-trained CNN to an SNN. We observe that the existing CNN-to-SNN conversion algor...
Autores principales: | , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Frontiers Media S.A.
2022
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9179229/ https://www.ncbi.nlm.nih.gov/pubmed/35692427 http://dx.doi.org/10.3389/fnins.2022.759900 |
_version_ | 1784723223842127872 |
---|---|
author | Wu, Dengyu Yi, Xinping Huang, Xiaowei |
author_facet | Wu, Dengyu Yi, Xinping Huang, Xiaowei |
author_sort | Wu, Dengyu |
collection | PubMed |
description | This article conforms to a recent trend of developing an energy-efficient Spiking Neural Network (SNN), which takes advantage of the sophisticated training regime of Convolutional Neural Network (CNN) and converts a well-trained CNN to an SNN. We observe that the existing CNN-to-SNN conversion algorithms may keep a certain amount of residual current in the spiking neurons in SNN, and the residual current may cause significant accuracy loss when inference time is short. To deal with this, we propose a unified framework to equalize the output of the convolutional or dense layer in CNN and the accumulated current in SNN, and maximally align the spiking rate of a neuron with its corresponding charge. This framework enables us to design a novel explicit current control (ECC) method for the CNN-to-SNN conversion which considers multiple objectives at the same time during the conversion, including accuracy, latency, and energy efficiency. We conduct an extensive set of experiments on different neural network architectures, e.g., VGG, ResNet, and DenseNet, to evaluate the resulting SNNs. The benchmark datasets include not only the image datasets such as CIFAR-10/100 and ImageNet but also the Dynamic Vision Sensor (DVS) image datasets such as DVS-CIFAR-10. The experimental results show the superior performance of our ECC method over the state-of-the-art. |
format | Online Article Text |
id | pubmed-9179229 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2022 |
publisher | Frontiers Media S.A. |
record_format | MEDLINE/PubMed |
spelling | pubmed-91792292022-06-10 A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network From Convolutional Neural Network Wu, Dengyu Yi, Xinping Huang, Xiaowei Front Neurosci Neuroscience This article conforms to a recent trend of developing an energy-efficient Spiking Neural Network (SNN), which takes advantage of the sophisticated training regime of Convolutional Neural Network (CNN) and converts a well-trained CNN to an SNN. We observe that the existing CNN-to-SNN conversion algorithms may keep a certain amount of residual current in the spiking neurons in SNN, and the residual current may cause significant accuracy loss when inference time is short. To deal with this, we propose a unified framework to equalize the output of the convolutional or dense layer in CNN and the accumulated current in SNN, and maximally align the spiking rate of a neuron with its corresponding charge. This framework enables us to design a novel explicit current control (ECC) method for the CNN-to-SNN conversion which considers multiple objectives at the same time during the conversion, including accuracy, latency, and energy efficiency. We conduct an extensive set of experiments on different neural network architectures, e.g., VGG, ResNet, and DenseNet, to evaluate the resulting SNNs. The benchmark datasets include not only the image datasets such as CIFAR-10/100 and ImageNet but also the Dynamic Vision Sensor (DVS) image datasets such as DVS-CIFAR-10. The experimental results show the superior performance of our ECC method over the state-of-the-art. Frontiers Media S.A. 2022-05-26 /pmc/articles/PMC9179229/ /pubmed/35692427 http://dx.doi.org/10.3389/fnins.2022.759900 Text en Copyright © 2022 Wu, Yi and Huang. https://creativecommons.org/licenses/by/4.0/This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. |
spellingShingle | Neuroscience Wu, Dengyu Yi, Xinping Huang, Xiaowei A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network From Convolutional Neural Network |
title | A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network From Convolutional Neural Network |
title_full | A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network From Convolutional Neural Network |
title_fullStr | A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network From Convolutional Neural Network |
title_full_unstemmed | A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network From Convolutional Neural Network |
title_short | A Little Energy Goes a Long Way: Build an Energy-Efficient, Accurate Spiking Neural Network From Convolutional Neural Network |
title_sort | little energy goes a long way: build an energy-efficient, accurate spiking neural network from convolutional neural network |
topic | Neuroscience |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9179229/ https://www.ncbi.nlm.nih.gov/pubmed/35692427 http://dx.doi.org/10.3389/fnins.2022.759900 |
work_keys_str_mv | AT wudengyu alittleenergygoesalongwaybuildanenergyefficientaccuratespikingneuralnetworkfromconvolutionalneuralnetwork AT yixinping alittleenergygoesalongwaybuildanenergyefficientaccuratespikingneuralnetworkfromconvolutionalneuralnetwork AT huangxiaowei alittleenergygoesalongwaybuildanenergyefficientaccuratespikingneuralnetworkfromconvolutionalneuralnetwork AT wudengyu littleenergygoesalongwaybuildanenergyefficientaccuratespikingneuralnetworkfromconvolutionalneuralnetwork AT yixinping littleenergygoesalongwaybuildanenergyefficientaccuratespikingneuralnetworkfromconvolutionalneuralnetwork AT huangxiaowei littleenergygoesalongwaybuildanenergyefficientaccuratespikingneuralnetworkfromconvolutionalneuralnetwork |