An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features

Tingyu Chen; Xin Hu; Yiren Zhou; Zhuo Zou; Longfei Liang; Wen Chi Yang

doi:10.1109/ARACE56528.2022.00012

An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features

Tingyu Chen
, Xin Hu
, Yiren Zhou
, Zhuo Zou
, Longfei Liang
, Wen Chi Yang^*

^*Corresponding author for this work

Research output: Chapter in Book or Report/Conference proceeding › Conference Proceeding › peer-review

Abstract

We proposed a new Hebbian learning rule that Neglects Historical data and only Compares Voltages (referred to NHCV in the paper). Unlike the traditional Hebbian learning rules that rely on comparing the spike timing, NHCV is designed to adjust the weight of the synapse based on the voltage of the neuron as soon as it fires. NHCV is computationally efficient and have advantages in processing informative features. Compared to traditional STDP learning rules, it accelerated training process (0.5 to 2 seconds improvement on each sample) and achieved better accuracy on Wine dataset (5.7% absolute improvement) and Diabetes dataset (12% absolute improvement). We reveal that the information amount inside the features of a dataset considerably affects the performance of SNNs.

Original language	English
Title of host publication	Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	18-23
Number of pages	6
ISBN (Electronic)	9781665451536
DOIs	https://doi.org/10.1109/ARACE56528.2022.00012
Publication status	Published - 2022
Externally published	Yes
Event	2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022 - Qingdao, China Duration: 26 Aug 2022 → 28 Aug 2022

Publication series

Name	Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022

Conference

Conference	2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022
Country/Territory	China
City	Qingdao
Period	26/08/22 → 28/08/22

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

Neural Network Theory and Architectures
Performance analysis of Machine Learning Algorithms
Spiking Neural Network
Unsupervised and Supervised Learning

Access to Document

10.1109/ARACE56528.2022.00012

Cite this

Chen, T., Hu, X., Zhou, Y., Zou, Z., Liang, L., & Yang, W. C. (2022). An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features. In Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022 (pp. 18-23). (Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ARACE56528.2022.00012

Chen, Tingyu ; Hu, Xin ; Zhou, Yiren et al. / An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features. Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022. Institute of Electrical and Electronics Engineers Inc., 2022. pp. 18-23 (Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022).

@inproceedings{be4244d2246e44d28a5d71eebc93ebab,

title = "An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features",

abstract = "We proposed a new Hebbian learning rule that Neglects Historical data and only Compares Voltages (referred to NHCV in the paper). Unlike the traditional Hebbian learning rules that rely on comparing the spike timing, NHCV is designed to adjust the weight of the synapse based on the voltage of the neuron as soon as it fires. NHCV is computationally efficient and have advantages in processing informative features. Compared to traditional STDP learning rules, it accelerated training process (0.5 to 2 seconds improvement on each sample) and achieved better accuracy on Wine dataset (5.7\% absolute improvement) and Diabetes dataset (12\% absolute improvement). We reveal that the information amount inside the features of a dataset considerably affects the performance of SNNs.",

keywords = "Neural Network Theory and Architectures, Performance analysis of Machine Learning Algorithms, Spiking Neural Network, Unsupervised and Supervised Learning",

author = "Tingyu Chen and Xin Hu and Yiren Zhou and Zhuo Zou and Longfei Liang and Yang, \{Wen Chi\}",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022 ; Conference date: 26-08-2022 Through 28-08-2022",

year = "2022",

doi = "10.1109/ARACE56528.2022.00012",

language = "English",

series = "Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "18--23",

booktitle = "Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022",

}

Chen, T, Hu, X, Zhou, Y, Zou, Z, Liang, L & Yang, WC 2022, An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features. in Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022. Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022, Institute of Electrical and Electronics Engineers Inc., pp. 18-23, 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022, Qingdao, China, 26/08/22. https://doi.org/10.1109/ARACE56528.2022.00012

An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features. / Chen, Tingyu; Hu, Xin; Zhou, Yiren et al.
Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022. Institute of Electrical and Electronics Engineers Inc., 2022. p. 18-23 (Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022).

Research output: Chapter in Book or Report/Conference proceeding › Conference Proceeding › peer-review

TY - GEN

T1 - An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features

AU - Chen, Tingyu

AU - Hu, Xin

AU - Zhou, Yiren

AU - Zou, Zhuo

AU - Liang, Longfei

AU - Yang, Wen Chi

PY - 2022

Y1 - 2022

N2 - We proposed a new Hebbian learning rule that Neglects Historical data and only Compares Voltages (referred to NHCV in the paper). Unlike the traditional Hebbian learning rules that rely on comparing the spike timing, NHCV is designed to adjust the weight of the synapse based on the voltage of the neuron as soon as it fires. NHCV is computationally efficient and have advantages in processing informative features. Compared to traditional STDP learning rules, it accelerated training process (0.5 to 2 seconds improvement on each sample) and achieved better accuracy on Wine dataset (5.7% absolute improvement) and Diabetes dataset (12% absolute improvement). We reveal that the information amount inside the features of a dataset considerably affects the performance of SNNs.

AB - We proposed a new Hebbian learning rule that Neglects Historical data and only Compares Voltages (referred to NHCV in the paper). Unlike the traditional Hebbian learning rules that rely on comparing the spike timing, NHCV is designed to adjust the weight of the synapse based on the voltage of the neuron as soon as it fires. NHCV is computationally efficient and have advantages in processing informative features. Compared to traditional STDP learning rules, it accelerated training process (0.5 to 2 seconds improvement on each sample) and achieved better accuracy on Wine dataset (5.7% absolute improvement) and Diabetes dataset (12% absolute improvement). We reveal that the information amount inside the features of a dataset considerably affects the performance of SNNs.

KW - Neural Network Theory and Architectures

KW - Performance analysis of Machine Learning Algorithms

KW - Spiking Neural Network

KW - Unsupervised and Supervised Learning

UR - https://www.scopus.com/pages/publications/85143435835

U2 - 10.1109/ARACE56528.2022.00012

DO - 10.1109/ARACE56528.2022.00012

M3 - Conference Proceeding

AN - SCOPUS:85143435835

T3 - Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022

SP - 18

EP - 23

BT - Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022

Y2 - 26 August 2022 through 28 August 2022

ER -

Chen T, Hu X, Zhou Y, Zou Z, Liang L, Yang WC. An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features. In Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022. Institute of Electrical and Electronics Engineers Inc. 2022. p. 18-23. (Proceedings - 2022 Asia Conference on Advanced Robotics, Automation, and Control Engineering, ARACE 2022). doi: 10.1109/ARACE56528.2022.00012

An optimized Hebbian Learning Rule for Spiking Neural Networks on the Classification Problems with Informative Data Features

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this