PP-NAS: Searching for Plug-and-Play Blocks on Convolutional Neural Networks
文献类型:期刊论文
作者 | Xiao, Anqi1,2; Shen, Biluo1,2; Tian, Jie1,2,3,4; Hu, Zhenhua1,2 |
刊名 | IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS |
出版日期 | 2023-04-26 |
页码 | 13 |
ISSN号 | 2162-237X |
关键词 | Convolution Search problems Optimization Computer architecture Object detection Molecular imaging Visualization Multiscale neural architecture search (NAS) plug-and-play representation learning |
DOI | 10.1109/TNNLS.2023.3264551 |
通讯作者 | Tian, Jie(tian@ieee.org) ; Hu, Zhenhua(zhenhua.hu@ia.ac.cn) |
英文摘要 | Multiscale features are of great importance in modern convolutional neural networks, showing consistent performance gains on numerous vision tasks. Therefore, many plug-and-play blocks are introduced to upgrade existing convolutional neural networks for stronger multiscale representation ability. However, the design of plug-and-play blocks is getting more and more complex, and these manually designed blocks are not optimal. In this work, we propose PP-NAS to develop plug-and-play blocks based on neural architecture search (NAS). Specifically, we design a new search space PPConv and develop a search algorithm consisting of one-level optimization, zero-one loss, and connection existence loss. PP-NAS minimizes the optimization gap between super-net and subarchitectures and can achieve good performance even without retraining. Extensive experiments on image classification, object detection, and semantic segmentation verify the superiority of PP-NAS over state-of-the-art CNNs (e.g., ResNet, ResNeXt, and Res2Net). Our code is available at https://github.com/ainieli/PP-NAS. |
资助项目 | National Natural Science Foundation of China (NSFC)[92059207] ; National Natural Science Foundation of China (NSFC)[62027901] ; National Natural Science Foundation of China (NSFC)[81930053] ; National Natural Science Foundation of China (NSFC)[81227901] ; Chinese Academy of Sciences (CAS) Youth Interdisciplinary Team[JCTD-2021-08] ; Zhuhai High-Level Health Personnel Team Project[HLHPTP201703] ; Google's TPU Research Cloud (TRC) |
WOS研究方向 | Computer Science ; Engineering |
语种 | 英语 |
出版者 | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC |
WOS记录号 | WOS:000980562400001 |
资助机构 | National Natural Science Foundation of China (NSFC) ; Chinese Academy of Sciences (CAS) Youth Interdisciplinary Team ; Zhuhai High-Level Health Personnel Team Project ; Google's TPU Research Cloud (TRC) |
源URL | [http://ir.ia.ac.cn/handle/173211/53236] |
专题 | 自动化研究所_中国科学院分子影像重点实验室 |
通讯作者 | Tian, Jie; Hu, Zhenhua |
作者单位 | 1.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 101408, Peoples R China 2.Chinese Acad Sci, Inst Automat, Beijing Key Lab Mol Imaging, CAS Key Lab Mol Imaging, Beijing 100190, Peoples R China 3.Xidian Univ, Engn Res Ctr Mol & Neuro Imaging, Sch Life Sci & Technol, Minist Educ, Xian 710071, Peoples R China 4.Beihang Univ, Sch Engn Med, Beijing Adv Innovat Ctr Big Data Based Precis Med, Beijing 100191, Peoples R China |
推荐引用方式 GB/T 7714 | Xiao, Anqi,Shen, Biluo,Tian, Jie,et al. PP-NAS: Searching for Plug-and-Play Blocks on Convolutional Neural Networks[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,2023:13. |
APA | Xiao, Anqi,Shen, Biluo,Tian, Jie,&Hu, Zhenhua.(2023).PP-NAS: Searching for Plug-and-Play Blocks on Convolutional Neural Networks.IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,13. |
MLA | Xiao, Anqi,et al."PP-NAS: Searching for Plug-and-Play Blocks on Convolutional Neural Networks".IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS (2023):13. |
入库方式: OAI收割
来源:自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。