Analysis of Deep learning Quantization Technology for Micro-sized IoT devices (초소형 IoT 장치에 구현 가능한 딥러닝 양자화 기술 분석)

KIM Young min (김영민); KyungHyun Han (한경현); Seong Oun Hwang (황성운)

doi:10.20465/KIOTS.2023.9.1.009

Analysis of Deep learning Quantization Technology for Micro-sized IoT devices

Journal of Internet of Things and Convergence
Abbr : JKIOTS
2023, 9(1), pp.9-17
DOI : 10.20465/KIOTS.2023.9.1.009
Publisher : The Korea Internet of Things Society
Research Area : Engineering > Computer Science > Internet Information Processing
Received : November 3, 2022
Accepted : December 14, 2022
Published : February 28, 2023

KIM Young min ¹, KyungHyun Han ², Seong Oun Hwang ¹

¹가천대학교
²홍익대학교(세종캠퍼스)

Accredited

ABSTRACT

Deep learning with large amount of computations is difficult to implement on micro-sized IoT devices or moblie devices. Recently, lightweight deep learning technologies have been introduced to make sure that deep learning can be implemented even on small devices by reducing the amount of computation of the model. Quantization is one of lightweight techniques that can be efficiently used to reduce the memory and size of the model by expressing parameter values with continuous distribution as discrete values of fixed bits. However, the accuracy of the model is reduced due to discrete value representation in quantization. In this paper, we introduce various quantization techniques to correct the accuracy. We selected APoT and EWGS from existing quantization techniques, and comparatively analyzed the results through experimentations The selected techniques were trained and tested with CIFAR-10 or CIFAR-100 datasets in the ResNet model. We found out problems with them through experimental results analysis and presented directions for future research.

KEYWORDS

Internet of Things, Deep Learning, Quantization, Model Training, Experimental Configuration

Citation status

* References for papers published after 2023 are currently being built.

[other] Howard, Andrew G. / 2017 / Mobilenets : Efficient convolutional neural networks for mobile vision applications / arXiv preprint arXiv : 1704. 04861

[journal] Blalock, Davis / 2020 / What is the state of neural network pruning? / Proceedings of machine learning and systems 2 : 129~146

[other] Hinton, Geoffrey / 2015 / Distilling the knowledge in a neural network / arXiv preprint arXiv : 1503. 02531

[confproc] Itay Hubara / 2016 / Binarized neural networks / Advances in neural information processing systems 29

[other] Raghuraman Krishnamoorthi / 2018 / Quantizing deep convolutional networks for efficient inference : A whitepaper / arXiv preprint arXiv : 1806. 08342

[confproc] Benoit Jacob / 2018 / Quantization and training of neural networks for efficient integer-arithmetic-only inference / Proceedings of the IEEE conference on computer vision and pattern recognition : 2704~2713

[other] Hao Wu / 2020 / Integer quantization for deep learning inference : Principles and empirical evaluation / arXiv preprint arXiv : 2004. 09602

[other] Song Han / 2015 / Deep compression : Compressing deep neural networks with pruning, trained quantization and Huffman coding / arXiv preprint arXiv : 1510. 00149

[confproc] Zhaohui Yang / 2020 / Searching for low-bit weights in quantized neural networks / Advances in neural information processing systems 33 : 4091~4102

[confproc] Kohei Yamamoto / 2021 / Learnable companding quantization for accurate low-bit neural networks / Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition : 5029~5038

[other] Yunchao Gong / 2014 / Compressing deep convolutional networks using vector quantization / arXiv preprint arXiv : 1412. 6115

[confproc] Yang, Jiwei / 2019 / Quantization networks / Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition : 7308~7316

[confproc] Gong, Ruihao / 2019 / Differentiable soft quantization : Bridging full-precision and low-bit neural networks / Proceedings of the IEEE/CVF International Conference on Computer Vision : 4852~4861

[confproc] Kim, Dohyung / 2021 / Distance-aware quantization / Proceedings of the IEEE/CVF International Conference on Computer Vision : 5271~5280

[other] Aojun Zhou / 2017 / Incremental network quantization : Towards lossless cnns with low-precision weights / arXiv preprint arXiv : 1702. 03044

[confproc] Yuhang Li / 2020 / Additive powers-of-two quantization: An efficient non-uniform discretization for neural networks / International Conference on Learning Representations

[confproc] Lee, Junghyup / 2021 / Network quantization with element-wise gradient scaling / Proceedings of the IEEE/CVF conference on computer vision and pattern recognition : 6448~6457

[other] Yoshua Bengio / 2013 / Estimating or propagating gradients through stochastic neurons for conditional computation / arXiv preprint arXiv : 1308. 3432

[journal] Avron, Haim / 2011 / Randomized algorithms for estimating the trace of an implicit symmetric positive semi-definite matrix / Journal of the ACM(JACM) 58(2) : 1~34

[other] Itay Hubara / 2020 / Improving post training neural quantization : Layer-wise calibration and integer programming / arXiv preprint arXiv : 2006. 10518

[confproc] Markus Nagel / 2019 / Data-free quantization through weight equalization and bias correction / Proceedings of the IEEE/CVF International Conference on Computer Vision : 1325~1334

[other] Li, Yuhang / 2021 / Brecq : Pushing the limit of post-training quantization by block reconstruction / arXiv preprint arXiv : 2102. 05426

[confproc] Kaiming He / 2016 / Deep residual learning for image recognition / Proceedings of the IEEE conference on computer vision and pattern recognition : 770~778

[other] Alex Krizhevsky / 2009 / Learning multiple layers of features from tiny images

[confproc] Lee, Junghyup / 2019 / Sfnet : Learning object-aware semantic correspondence / Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition : 2278~2287

KJCKorea
Journal Central

Journal of Internet of Things and Convergence 2023 KCI Impact Factor : 0.89

Analysis of Deep learning Quantization Technology for Micro-sized IoT devices

ABSTRACT

KEYWORDS

Citation status

* References for papers published after 2023 are currently being built.

Journal of Internet of Things and Convergence 2023 KCI Impact Factor : 0.89

Analysis of Deep learning Quantization Technology for Micro-sized IoT devices

ABSTRACT

KEYWORDS

Statistics

Tools

Issue List

Citation status

KCI Citation Counts (0)

REFERENCES (25) * References for papers published after 2023 are currently being built.

Search PDF

Citation

* References for papers published after 2023 are currently being built.