Cross-attention-map-based regularization for adversarial domain adaptation

doi:10.1016/j.neunet.2021.10.013

CORC > 自动化研究所 > 中国科学院自动化研究所 > 综合信息系统研究中心 > 工业智能技术与系统

	Cross-attention-map-based regularization for adversarial domain adaptation
	Jingwei, Li 1,2; Huanjie, Wang 1,2; Ke, Wu 1,2; Chengbao, Liu 2; Jie, Tan 1,2
刊名	NEURAL NETWORKS
	2022
卷号	145 页码:128-138
关键词	Domain adaptation Few-shot learning Attention mechanism Contrastive learning
ISSN号	0893-6080
DOI	10.1016/j.neunet.2021.10.013
通讯作者	Jie, Tan(jie.tan@ia.ac.cn)
英文摘要	In unsupervised domain adaptation (UDA), many efforts are taken to pull the source domain and the target domain closer by adversarial training. Most methods focus on aligning distributions or features between the source domain and the target domain. However, little attention is paid to the interaction between finer-grained levels, such as classes or samples of the two domains. In contrast to UDA, another transfer learning task, i.e., few-shot learning (FSL), takes full advantage of the finer-grained-level alignment. Many FSL methods implement the interaction between samples of support sets and query sets, leading to significant improvements. We wonder whether we can get some inspiration from these methods and bring such ideas of FSL to UDA. To this end, we first take a closer look at the differences between FSL and UDA and bridge the gap between them by high-confidence sample selection (HCSS). Then we propose cross-attention map generation module (CAMGM) to interact samples selected by HCSS. Moreover, we propose a simple but efficient method called cross-attention-map-based regularization (CAMR) to regularize the feature maps generated by the feature extractor. Experiments on three challenging datasets demonstrate that CAMR can bring solid improvements when added to the original objective. More specifically, the proposed CAMR can outperform original methods by 1% to 2% in most tasks without bells and whistles. (C) 2021 Elsevier Ltd. All rights reserved.
资助项目	National Natural Science Foundation of China under People's Republic of China[U1801263] ; National Natural Science Foundation of China under People's Republic of China[U1701262]
WOS研究方向	Computer Science ; Neurosciences & Neurology
语种	英语
出版者	PERGAMON-ELSEVIER SCIENCE LTD
WOS记录号	WOS:000717665500007
资助机构	National Natural Science Foundation of China under People's Republic of China
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/46512]
专题	综合信息系统研究中心_工业智能技术与系统
通讯作者	Jie, Tan
作者单位	1.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 2.Chinese Acad Sci, Inst Automat, 95 East Zhongguancun Rd, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Jingwei, Li,Huanjie, Wang,Ke, Wu,et al. Cross-attention-map-based regularization for adversarial domain adaptation[J]. NEURAL NETWORKS,2022,145:128-138.
APA	Jingwei, Li,Huanjie, Wang,Ke, Wu,Chengbao, Liu,&Jie, Tan.(2022).Cross-attention-map-based regularization for adversarial domain adaptation.NEURAL NETWORKS,145,128-138.
MLA	Jingwei, Li,et al."Cross-attention-map-based regularization for adversarial domain adaptation".NEURAL NETWORKS 145(2022):128-138.