A Free Lunch From ANN: Towards Efficient, Accurate Spiking Neural Networks Calibration

Yuhang Li, Shikuang Deng, Xin Dong, Ruihao Gong, Shi Gu
Proceedings of the 38th International Conference on Machine Learning, PMLR 139:6316-6325, 2021.

Abstract

Spiking Neural Network (SNN) has been recognized as one of the next generation of neural networks. Conventionally, SNN can be converted from a pre-trained ANN by only replacing the ReLU activation to spike activation while keeping the parameters intact. Perhaps surprisingly, in this work we show that a proper way to calibrate the parameters during the conversion of ANN to SNN can bring significant improvements. We introduce SNN Calibration, a cheap but extraordinarily effective method by leveraging the knowledge within a pre-trained Artificial Neural Network (ANN). Starting by analyzing the conversion error and its propagation through layers theoretically, we propose the calibration algorithm that can correct the error layer-by-layer. The calibration only takes a handful number of training data and several minutes to finish. Moreover, our calibration algorithm can produce SNN with state-of-the-art architecture on the large-scale ImageNet dataset, including MobileNet and RegNet. Extensive experiments demonstrate the effectiveness and efficiency of our algorithm. For example, our advanced pipeline can increase up to 69% top-1 accuracy when converting MobileNet on ImageNet compared to baselines. Codes are released at https://github.com/yhhhli/SNN_Calibration.

Cite this Paper

BibTeX
@InProceedings{pmlr-v139-li21d, title = {A Free Lunch From ANN: Towards Efficient, Accurate Spiking Neural Networks Calibration}, author = {Li, Yuhang and Deng, Shikuang and Dong, Xin and Gong, Ruihao and Gu, Shi}, booktitle = {Proceedings of the 38th International Conference on Machine Learning}, pages = {6316--6325}, year = {2021}, editor = {Meila, Marina and Zhang, Tong}, volume = {139}, series = {Proceedings of Machine Learning Research}, month = {18--24 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v139/li21d/li21d.pdf}, url = {https://proceedings.mlr.press/v139/li21d.html}, abstract = {Spiking Neural Network (SNN) has been recognized as one of the next generation of neural networks. Conventionally, SNN can be converted from a pre-trained ANN by only replacing the ReLU activation to spike activation while keeping the parameters intact. Perhaps surprisingly, in this work we show that a proper way to calibrate the parameters during the conversion of ANN to SNN can bring significant improvements. We introduce SNN Calibration, a cheap but extraordinarily effective method by leveraging the knowledge within a pre-trained Artificial Neural Network (ANN). Starting by analyzing the conversion error and its propagation through layers theoretically, we propose the calibration algorithm that can correct the error layer-by-layer. The calibration only takes a handful number of training data and several minutes to finish. Moreover, our calibration algorithm can produce SNN with state-of-the-art architecture on the large-scale ImageNet dataset, including MobileNet and RegNet. Extensive experiments demonstrate the effectiveness and efficiency of our algorithm. For example, our advanced pipeline can increase up to 69% top-1 accuracy when converting MobileNet on ImageNet compared to baselines. Codes are released at https://github.com/yhhhli/SNN_Calibration.} }
Endnote
%0 Conference Paper %T A Free Lunch From ANN: Towards Efficient, Accurate Spiking Neural Networks Calibration %A Yuhang Li %A Shikuang Deng %A Xin Dong %A Ruihao Gong %A Shi Gu %B Proceedings of the 38th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Marina Meila %E Tong Zhang %F pmlr-v139-li21d %I PMLR %P 6316--6325 %U https://proceedings.mlr.press/v139/li21d.html %V 139 %X Spiking Neural Network (SNN) has been recognized as one of the next generation of neural networks. Conventionally, SNN can be converted from a pre-trained ANN by only replacing the ReLU activation to spike activation while keeping the parameters intact. Perhaps surprisingly, in this work we show that a proper way to calibrate the parameters during the conversion of ANN to SNN can bring significant improvements. We introduce SNN Calibration, a cheap but extraordinarily effective method by leveraging the knowledge within a pre-trained Artificial Neural Network (ANN). Starting by analyzing the conversion error and its propagation through layers theoretically, we propose the calibration algorithm that can correct the error layer-by-layer. The calibration only takes a handful number of training data and several minutes to finish. Moreover, our calibration algorithm can produce SNN with state-of-the-art architecture on the large-scale ImageNet dataset, including MobileNet and RegNet. Extensive experiments demonstrate the effectiveness and efficiency of our algorithm. For example, our advanced pipeline can increase up to 69% top-1 accuracy when converting MobileNet on ImageNet compared to baselines. Codes are released at https://github.com/yhhhli/SNN_Calibration.
APA
Li, Y., Deng, S., Dong, X., Gong, R. & Gu, S.. (2021). A Free Lunch From ANN: Towards Efficient, Accurate Spiking Neural Networks Calibration. Proceedings of the 38th International Conference on Machine Learning, in Proceedings of Machine Learning Research 139:6316-6325 Available from https://proceedings.mlr.press/v139/li21d.html.

Related Material