Understanding the functional roles of modelling components in spiking neural networks

Spiking neural networks (SNNs), inspired by the neural circuits of the brain, are promising in achieving high computational efficiency with biological fidelity. Nevertheless, it is quite difficult to optimize SNNs because the functional roles of their modelling components remain unclear. By designin...

Full description

Saved in:
Bibliographic Details
Main Authors: Huifeng Yin, Hanle Zheng, Jiayi Mao, Siyuan Ding, Xing Liu, Mingkun Xu, Yifan Hu, Jing Pei, Lei Deng
Format: Article
Language:English
Published: IOP Publishing 2024-01-01
Series:Neuromorphic Computing and Engineering
Subjects:
Online Access:https://doi.org/10.1088/2634-4386/ad6cef
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846149632043778048
author Huifeng Yin
Hanle Zheng
Jiayi Mao
Siyuan Ding
Xing Liu
Mingkun Xu
Yifan Hu
Jing Pei
Lei Deng
author_facet Huifeng Yin
Hanle Zheng
Jiayi Mao
Siyuan Ding
Xing Liu
Mingkun Xu
Yifan Hu
Jing Pei
Lei Deng
author_sort Huifeng Yin
collection DOAJ
description Spiking neural networks (SNNs), inspired by the neural circuits of the brain, are promising in achieving high computational efficiency with biological fidelity. Nevertheless, it is quite difficult to optimize SNNs because the functional roles of their modelling components remain unclear. By designing and evaluating several variants of the classic model, we systematically investigate the functional roles of key modelling components, leakage, reset, and recurrence, in leaky integrate-and-fire (LIF) based SNNs. Through extensive experiments, we demonstrate how these components influence the accuracy, generalization, and robustness of SNNs. Specifically, we find that the leakage plays a crucial role in balancing memory retention and robustness, the reset mechanism is essential for uninterrupted temporal processing and computational efficiency, and the recurrence enriches the capability to model complex dynamics at a cost of robustness degradation. With these interesting observations, we provide optimization suggestions for enhancing the performance of SNNs in different scenarios. This work deepens the understanding of how SNNs work, which offers valuable guidance for the development of more effective and robust neuromorphic models.
format Article
id doaj-art-eedc1f0716694835b37bb95be39befa8
institution Kabale University
issn 2634-4386
language English
publishDate 2024-01-01
publisher IOP Publishing
record_format Article
series Neuromorphic Computing and Engineering
spelling doaj-art-eedc1f0716694835b37bb95be39befa82024-11-29T14:19:33ZengIOP PublishingNeuromorphic Computing and Engineering2634-43862024-01-014303400910.1088/2634-4386/ad6cefUnderstanding the functional roles of modelling components in spiking neural networksHuifeng Yin0Hanle Zheng1Jiayi Mao2Siyuan Ding3https://orcid.org/0009-0004-4722-9053Xing Liu4Mingkun Xu5Yifan Hu6Jing Pei7Lei Deng8https://orcid.org/0000-0002-5172-9411Center for Brain Inspired Computing Research (CBICR), Department of Precision Instrument, Tsinghua University , Beijing, People’s Republic of ChinaCenter for Brain Inspired Computing Research (CBICR), Department of Precision Instrument, Tsinghua University , Beijing, People’s Republic of ChinaCenter for Brain Inspired Computing Research (CBICR), Department of Precision Instrument, Tsinghua University , Beijing, People’s Republic of ChinaWeiyang College, Tsinghua University , Beijing, People’s Republic of ChinaCollege of Electronic Information and Automation, Tianjin University of Science and Technology , Tianjin, People’s Republic of ChinaGuangdong Institute of Intelligence Science and Technology , Zhuhai, People’s Republic of ChinaCenter for Brain Inspired Computing Research (CBICR), Department of Precision Instrument, Tsinghua University , Beijing, People’s Republic of ChinaCenter for Brain Inspired Computing Research (CBICR), Department of Precision Instrument, Tsinghua University , Beijing, People’s Republic of ChinaCenter for Brain Inspired Computing Research (CBICR), Department of Precision Instrument, Tsinghua University , Beijing, People’s Republic of ChinaSpiking neural networks (SNNs), inspired by the neural circuits of the brain, are promising in achieving high computational efficiency with biological fidelity. Nevertheless, it is quite difficult to optimize SNNs because the functional roles of their modelling components remain unclear. By designing and evaluating several variants of the classic model, we systematically investigate the functional roles of key modelling components, leakage, reset, and recurrence, in leaky integrate-and-fire (LIF) based SNNs. Through extensive experiments, we demonstrate how these components influence the accuracy, generalization, and robustness of SNNs. Specifically, we find that the leakage plays a crucial role in balancing memory retention and robustness, the reset mechanism is essential for uninterrupted temporal processing and computational efficiency, and the recurrence enriches the capability to model complex dynamics at a cost of robustness degradation. With these interesting observations, we provide optimization suggestions for enhancing the performance of SNNs in different scenarios. This work deepens the understanding of how SNNs work, which offers valuable guidance for the development of more effective and robust neuromorphic models.https://doi.org/10.1088/2634-4386/ad6cefspiking neural networksneuromorphic computingfunctional rolesmodelling componentsrobustness
spellingShingle Huifeng Yin
Hanle Zheng
Jiayi Mao
Siyuan Ding
Xing Liu
Mingkun Xu
Yifan Hu
Jing Pei
Lei Deng
Understanding the functional roles of modelling components in spiking neural networks
Neuromorphic Computing and Engineering
spiking neural networks
neuromorphic computing
functional roles
modelling components
robustness
title Understanding the functional roles of modelling components in spiking neural networks
title_full Understanding the functional roles of modelling components in spiking neural networks
title_fullStr Understanding the functional roles of modelling components in spiking neural networks
title_full_unstemmed Understanding the functional roles of modelling components in spiking neural networks
title_short Understanding the functional roles of modelling components in spiking neural networks
title_sort understanding the functional roles of modelling components in spiking neural networks
topic spiking neural networks
neuromorphic computing
functional roles
modelling components
robustness
url https://doi.org/10.1088/2634-4386/ad6cef
work_keys_str_mv AT huifengyin understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks
AT hanlezheng understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks
AT jiayimao understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks
AT siyuanding understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks
AT xingliu understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks
AT mingkunxu understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks
AT yifanhu understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks
AT jingpei understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks
AT leideng understandingthefunctionalrolesofmodellingcomponentsinspikingneuralnetworks