资源受限硬件设备的神经结构搜索:综述

IF 1.7 Q3 COMPUTER SCIENCE, INFORMATION SYSTEMS IET Cyber-Physical Systems: Theory and Applications Pub Date : 2023-07-03 DOI:10.1049/cps2.12058
Yongjia Yang, Jinyu Zhan, Wei Jiang, Yucheng Jiang, Antai Yu
{"title":"资源受限硬件设备的神经结构搜索:综述","authors":"Yongjia Yang,&nbsp;Jinyu Zhan,&nbsp;Wei Jiang,&nbsp;Yucheng Jiang,&nbsp;Antai Yu","doi":"10.1049/cps2.12058","DOIUrl":null,"url":null,"abstract":"<p>With the emergence of powerful and low-energy Internet of Things devices, deep learning computing is increasingly applied to resource-constrained edge devices. However, the mismatch between hardware devices with low computing capacity and the increasing complexity of Deep Neural Network models, as well as the growing real-time requirements, bring challenges to the design and deployment of deep learning models. For example, autonomous driving technologies rely on real-time object detection of the environment, which cannot tolerate the extra latency of sending data to the cloud, processing and then sending the results back to edge devices. Many studies aim to find innovative ways to reduce the size of deep learning models, the number of Floating-point Operations per Second, and the time overhead of inference. Neural Architecture Search (NAS) makes it possible to automatically generate efficient neural network models. The authors summarise the existing NAS methods on resource-constrained devices and categorise them according to single-objective or multi-objective optimisation. We review the search space, the search algorithm and the constraints of NAS on hardware devices. We also explore the challenges and open problems of hardware NAS.</p>","PeriodicalId":36881,"journal":{"name":"IET Cyber-Physical Systems: Theory and Applications","volume":null,"pages":null},"PeriodicalIF":1.7000,"publicationDate":"2023-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1049/cps2.12058","citationCount":"0","resultStr":"{\"title\":\"Neural architecture search for resource constrained hardware devices: A survey\",\"authors\":\"Yongjia Yang,&nbsp;Jinyu Zhan,&nbsp;Wei Jiang,&nbsp;Yucheng Jiang,&nbsp;Antai Yu\",\"doi\":\"10.1049/cps2.12058\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>With the emergence of powerful and low-energy Internet of Things devices, deep learning computing is increasingly applied to resource-constrained edge devices. However, the mismatch between hardware devices with low computing capacity and the increasing complexity of Deep Neural Network models, as well as the growing real-time requirements, bring challenges to the design and deployment of deep learning models. For example, autonomous driving technologies rely on real-time object detection of the environment, which cannot tolerate the extra latency of sending data to the cloud, processing and then sending the results back to edge devices. Many studies aim to find innovative ways to reduce the size of deep learning models, the number of Floating-point Operations per Second, and the time overhead of inference. Neural Architecture Search (NAS) makes it possible to automatically generate efficient neural network models. The authors summarise the existing NAS methods on resource-constrained devices and categorise them according to single-objective or multi-objective optimisation. We review the search space, the search algorithm and the constraints of NAS on hardware devices. We also explore the challenges and open problems of hardware NAS.</p>\",\"PeriodicalId\":36881,\"journal\":{\"name\":\"IET Cyber-Physical Systems: Theory and Applications\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":1.7000,\"publicationDate\":\"2023-07-03\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://onlinelibrary.wiley.com/doi/epdf/10.1049/cps2.12058\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IET Cyber-Physical Systems: Theory and Applications\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://onlinelibrary.wiley.com/doi/10.1049/cps2.12058\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"COMPUTER SCIENCE, INFORMATION SYSTEMS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IET Cyber-Physical Systems: Theory and Applications","FirstCategoryId":"1085","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1049/cps2.12058","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"COMPUTER SCIENCE, INFORMATION SYSTEMS","Score":null,"Total":0}
引用次数: 0

摘要

随着功能强大、能耗低的物联网设备的出现,深度学习计算越来越多地应用于资源受限的边缘设备。然而,计算能力低的硬件设备与深度神经网络模型日益复杂的不匹配,以及日益增长的实时性要求,给深度学习模型的设计和部署带来了挑战。例如,自动驾驶技术依赖于环境的实时物体检测,无法容忍将数据发送到云、处理然后将结果发送回边缘设备的额外延迟。许多研究旨在寻找创新的方法来减少深度学习模型的大小、每秒浮点运算的数量和推理的时间开销。神经结构搜索(NAS)使自动生成高效的神经网络模型成为可能。作者总结了资源受限设备上现有的NAS方法,并根据单目标或多目标优化对其进行了分类。我们回顾了搜索空间、搜索算法以及NAS对硬件设备的限制。我们还探讨了硬件NAS的挑战和悬而未决的问题。
本文章由计算机程序翻译,如有差异,请以英文原文为准。

摘要图片

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
Neural architecture search for resource constrained hardware devices: A survey

With the emergence of powerful and low-energy Internet of Things devices, deep learning computing is increasingly applied to resource-constrained edge devices. However, the mismatch between hardware devices with low computing capacity and the increasing complexity of Deep Neural Network models, as well as the growing real-time requirements, bring challenges to the design and deployment of deep learning models. For example, autonomous driving technologies rely on real-time object detection of the environment, which cannot tolerate the extra latency of sending data to the cloud, processing and then sending the results back to edge devices. Many studies aim to find innovative ways to reduce the size of deep learning models, the number of Floating-point Operations per Second, and the time overhead of inference. Neural Architecture Search (NAS) makes it possible to automatically generate efficient neural network models. The authors summarise the existing NAS methods on resource-constrained devices and categorise them according to single-objective or multi-objective optimisation. We review the search space, the search algorithm and the constraints of NAS on hardware devices. We also explore the challenges and open problems of hardware NAS.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
IET Cyber-Physical Systems: Theory and Applications
IET Cyber-Physical Systems: Theory and Applications Computer Science-Computer Networks and Communications
CiteScore
5.40
自引率
6.70%
发文量
17
审稿时长
19 weeks
期刊最新文献
Guest Editorial: IoT-based secure health monitoring and tracking through estimated computing SEIR-driven semantic integration framework: Internet of Things-enhanced epidemiological surveillance in COVID-19 outbreaks using recurrent neural networks A machine learning model for Alzheimer's disease prediction Securing the Internet of Medical Things with ECG-based PUF encryption Status, challenges, and promises of data-driven battery lifetime prediction under cyber-physical system context
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1