QSFM: Model Pruning Based on Quantified Similarity Between Feature Maps for AI on Edge

doi:10.1109/JIOT.2022.3190873

CASIA OpenIR > 多模态人工智能系统全国重点实验室 > 生物识别与安全技术

	QSFM: Model Pruning Based on Quantified Similarity Between Feature Maps for AI on Edge
	Wang, Zidu 1,2; Liu, Xuexin 1,2; Huang, Long 1; Chen, Yunqing 1; Zhang, Yufei 1; Lin, Zhikang 1; Wang, Rui1
发表期刊	IEEE INTERNET OF THINGS JOURNAL
ISSN	2327-4662
	2022-12-01
卷号	9 期号:23 页码:24506-24515
通讯作者	Wang, Rui(wangrui@ustb.edu.cn)
摘要	Convolutional neural networks (CNNs) have been applied in numerous Internet of Things (IoT) devices for multifarious downstream tasks. However, with the increasing amount of data on edge devices, CNNs can hardly complete some tasks in time with limited computing and storage resources. Recently, filter pruning has been regarded as an effective technique to compress and accelerate CNNs, but existing methods rarely prune CNNs from the perspective of compressing high-dimensional tensors. In this article, we propose a novel theory to find redundant information in 3-D tensors, namely, quantified similarity between feature maps (QSFM), and utilize this theory to guide the filter pruning procedure. We perform QSFM on data sets (CIFAR-10, CIFAR-100, and ILSVRC-12) and edge devices and demonstrate that the proposed method can find the redundant information in the neural networks effectively with comparable compression and tolerable drop of accuracy. Without any fine-tuning operation, QSFM can compress ResNet-56 on CIFAR-10 significantly (48.7% FLOPs and 57.9% parameters are reduced) with only a loss of 0.54% in the top-1 accuracy. For the practical application of edge devices, QSFM can accelerate MobileNet-V2 inference speed by 1.53 times with only a loss of 1.23% in the ILSVRC-12 top-1 accuracy.
关键词	Tensors Internet of Things Convolution Three-dimensional displays Quantization (signal) Hardware Training Edge computing filter pruning Internet of Things (IoT) model compression neural networks
DOI	10.1109/JIOT.2022.3190873
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[62173158] ; National Natural Science Foundation of China[61379134]
项目资助者	National Natural Science Foundation of China
WOS研究方向	Computer Science ; Engineering ; Telecommunications
WOS类目	Computer Science, Information Systems ; Engineering, Electrical & Electronic ; Telecommunications
WOS记录号	WOS:000904931000086
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计	被引频次：5[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.ia.ac.cn/handle/173211/51094
专题	多模态人工智能系统全国重点实验室_生物识别与安全技术
通讯作者	Wang, Rui
作者单位	1.Univ Sci & Technol Beijing, Sch Comp & Commun Engn, Beijing 100083, Peoples R China 2.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China
第一作者单位	中国科学院自动化研究所
推荐引用方式 GB/T 7714	Wang, Zidu,Liu, Xuexin,Huang, Long,et al. QSFM: Model Pruning Based on Quantified Similarity Between Feature Maps for AI on Edge[J]. IEEE INTERNET OF THINGS JOURNAL,2022,9(23):24506-24515.
APA	Wang, Zidu.,Liu, Xuexin.,Huang, Long.,Chen, Yunqing.,Zhang, Yufei.,...&Wang, Rui.(2022).QSFM: Model Pruning Based on Quantified Similarity Between Feature Maps for AI on Edge.IEEE INTERNET OF THINGS JOURNAL,9(23),24506-24515.
MLA	Wang, Zidu,et al."QSFM: Model Pruning Based on Quantified Similarity Between Feature Maps for AI on Edge".IEEE INTERNET OF THINGS JOURNAL 9.23(2022):24506-24515.