Deep Subdomain Adaptation Network for Image Classification

doi:10.1109/TNNLS.2020.2988928

	Deep Subdomain Adaptation Network for Image Classification
	Zhu, Yongchun 1,2; Zhuang, Fuzhen 1,2; Wang, Jindong 3; Ke, Guolin 3; Chen, Jingwu ; Bian, Jiang 3; Xiong, Hui 4; He, Qing 1,2
	2021-04-01
发表期刊	IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
ISSN	2162-237X
卷号	32 期号:4 页码:1713-1722
摘要	For a target task where the labeled data are unavailable, domain adaptation can transfer a learner from a different source domain. Previous deep domain adaptation methods mainly learn a global domain shift, i.e., align the global source and target distributions without considering the relationships between two subdomains within the same category of different domains, leading to unsatisfying transfer learning performance without capturing the fine-grained information. Recently, more and more researchers pay attention to subdomain adaptation that focuses on accurately aligning the distributions of the relevant subdomains. However, most of them are adversarial methods that contain several loss functions and converge slowly. Based on this, we present a deep subdomain adaptation network (DSAN) that learns a transfer network by aligning the relevant subdomain distributions of domain-specific layer activations across different domains based on a local maximum mean discrepancy (LMMD). Our DSAN is very simple but effective, which does not need adversarial training and converges fast. The adaptation can be achieved easily with most feedforward network models by extending them with LMMD loss, which can be trained efficiently via backpropagation. Experiments demonstrate that DSAN can achieve remarkable results on both object recognition tasks and digit classification tasks. Our code will be available at https://github.com/easezyc/deep-transfer-learning.
关键词	Task analysis Adaptation models Kernel Feature extraction Learning systems Semantics Training Domain adaptation fine grained subdomain
DOI	10.1109/TNNLS.2020.2988928
收录类别	SCI
语种	英语
资助项目	National Key Research and Development Program of China[2018YFB1004300] ; National Natural Science Foundation of China[U1836206] ; National Natural Science Foundation of China[U1811461] ; National Natural Science Foundation of China[61773361] ; Project of Youth Innovation Promotion Association CAS[2017146]
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Computer Science, Hardware & Architecture ; Computer Science, Theory & Methods ; Engineering, Electrical & Electronic
WOS记录号	WOS:000637534200025
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计	被引频次：504[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/16673
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Zhuang, Fuzhen
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 3.Microsoft Res, Beijing, Peoples R China 4.Rutgers State Univ, New Brunswick, NJ USA
推荐引用方式 GB/T 7714	Zhu, Yongchun,Zhuang, Fuzhen,Wang, Jindong,et al. Deep Subdomain Adaptation Network for Image Classification[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,2021,32(4):1713-1722.
APA	Zhu, Yongchun.,Zhuang, Fuzhen.,Wang, Jindong.,Ke, Guolin.,Chen, Jingwu.,...&He, Qing.(2021).Deep Subdomain Adaptation Network for Image Classification.IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,32(4),1713-1722.
MLA	Zhu, Yongchun,et al."Deep Subdomain Adaptation Network for Image Classification".IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 32.4(2021):1713-1722.