ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator
Spiking neural network (SNN) is a brain-inspired model with more spatio-temporal information processing capacity and computational energy efficiency. However, with the increasing depth of SNNs, the memory problem caused by the weights of SNNs has gradually attracted attention. In this study, we prop...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Frontiers Media S.A.
2023-09-01
|
Series: | Frontiers in Neuroscience |
Subjects: | |
Online Access: | https://www.frontiersin.org/articles/10.3389/fnins.2023.1225871/full |
_version_ | 1797687222567174144 |
---|---|
author | Yijian Pei Changqing Xu Changqing Xu Zili Wu Yi Liu Yintang Yang |
author_facet | Yijian Pei Changqing Xu Changqing Xu Zili Wu Yi Liu Yintang Yang |
author_sort | Yijian Pei |
collection | DOAJ |
description | Spiking neural network (SNN) is a brain-inspired model with more spatio-temporal information processing capacity and computational energy efficiency. However, with the increasing depth of SNNs, the memory problem caused by the weights of SNNs has gradually attracted attention. In this study, we propose an ultra-low latency adaptive local binary spiking neural network (ALBSNN) with accuracy loss estimators, which dynamically selects the network layers to be binarized to ensure a balance between quantization degree and classification accuracy by evaluating the error caused by the binarized weights during the network learning process. At the same time, to accelerate the training speed of the network, the global average pooling (GAP) layer is introduced to replace the fully connected layers by combining convolution and pooling. Finally, to further reduce the error caused by the binary weight, we propose binary weight optimization (BWO), which updates the overall weight by directly adjusting the binary weight. This method further reduces the loss of the network that reaches the training bottleneck. The combination of the above methods balances the network's quantization and recognition ability, enabling the network to maintain the recognition capability equivalent to the full precision network and reduce the storage space by more than 20%. So, SNNs can use a small number of time steps to obtain better recognition accuracy. In the extreme case of using only a one-time step, we still can achieve 93.39, 92.12, and 69.55% testing accuracy on three traditional static datasets, Fashion- MNIST, CIFAR-10, and CIFAR-100, respectively. At the same time, we evaluate our method on neuromorphic N-MNIST, CIFAR10-DVS, and IBM DVS128 Gesture datasets and achieve advanced accuracy in SNN with binary weights. Our network has greater advantages in terms of storage resources and training time. |
first_indexed | 2024-03-12T01:15:09Z |
format | Article |
id | doaj.art-ddc47e6be3d34ff193dfe9c3961e7524 |
institution | Directory Open Access Journal |
issn | 1662-453X |
language | English |
last_indexed | 2024-03-12T01:15:09Z |
publishDate | 2023-09-01 |
publisher | Frontiers Media S.A. |
record_format | Article |
series | Frontiers in Neuroscience |
spelling | doaj.art-ddc47e6be3d34ff193dfe9c3961e75242023-09-13T15:05:30ZengFrontiers Media S.A.Frontiers in Neuroscience1662-453X2023-09-011710.3389/fnins.2023.12258711225871ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimatorYijian Pei0Changqing Xu1Changqing Xu2Zili Wu3Yi Liu4Yintang Yang5Guangzhou Institute of Technology, Xidian University, Xi'an, ChinaGuangzhou Institute of Technology, Xidian University, Xi'an, ChinaSchool of Microelectronics, Xidian University, Xi'an, ChinaSchool of Computer Science and Technology, Xidian University, Xi'an, ChinaSchool of Microelectronics, Xidian University, Xi'an, ChinaSchool of Microelectronics, Xidian University, Xi'an, ChinaSpiking neural network (SNN) is a brain-inspired model with more spatio-temporal information processing capacity and computational energy efficiency. However, with the increasing depth of SNNs, the memory problem caused by the weights of SNNs has gradually attracted attention. In this study, we propose an ultra-low latency adaptive local binary spiking neural network (ALBSNN) with accuracy loss estimators, which dynamically selects the network layers to be binarized to ensure a balance between quantization degree and classification accuracy by evaluating the error caused by the binarized weights during the network learning process. At the same time, to accelerate the training speed of the network, the global average pooling (GAP) layer is introduced to replace the fully connected layers by combining convolution and pooling. Finally, to further reduce the error caused by the binary weight, we propose binary weight optimization (BWO), which updates the overall weight by directly adjusting the binary weight. This method further reduces the loss of the network that reaches the training bottleneck. The combination of the above methods balances the network's quantization and recognition ability, enabling the network to maintain the recognition capability equivalent to the full precision network and reduce the storage space by more than 20%. So, SNNs can use a small number of time steps to obtain better recognition accuracy. In the extreme case of using only a one-time step, we still can achieve 93.39, 92.12, and 69.55% testing accuracy on three traditional static datasets, Fashion- MNIST, CIFAR-10, and CIFAR-100, respectively. At the same time, we evaluate our method on neuromorphic N-MNIST, CIFAR10-DVS, and IBM DVS128 Gesture datasets and achieve advanced accuracy in SNN with binary weights. Our network has greater advantages in terms of storage resources and training time.https://www.frontiersin.org/articles/10.3389/fnins.2023.1225871/fullspiking neural networksbinary neural networksneuromorphic computingsparsityvisual recognition |
spellingShingle | Yijian Pei Changqing Xu Changqing Xu Zili Wu Yi Liu Yintang Yang ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator Frontiers in Neuroscience spiking neural networks binary neural networks neuromorphic computing sparsity visual recognition |
title | ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator |
title_full | ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator |
title_fullStr | ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator |
title_full_unstemmed | ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator |
title_short | ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator |
title_sort | albsnn ultra low latency adaptive local binary spiking neural network with accuracy loss estimator |
topic | spiking neural networks binary neural networks neuromorphic computing sparsity visual recognition |
url | https://www.frontiersin.org/articles/10.3389/fnins.2023.1225871/full |
work_keys_str_mv | AT yijianpei albsnnultralowlatencyadaptivelocalbinaryspikingneuralnetworkwithaccuracylossestimator AT changqingxu albsnnultralowlatencyadaptivelocalbinaryspikingneuralnetworkwithaccuracylossestimator AT changqingxu albsnnultralowlatencyadaptivelocalbinaryspikingneuralnetworkwithaccuracylossestimator AT ziliwu albsnnultralowlatencyadaptivelocalbinaryspikingneuralnetworkwithaccuracylossestimator AT yiliu albsnnultralowlatencyadaptivelocalbinaryspikingneuralnetworkwithaccuracylossestimator AT yintangyang albsnnultralowlatencyadaptivelocalbinaryspikingneuralnetworkwithaccuracylossestimator |