Mind the Gap: Open Set Domain Adaptation via Mutual-to-Separate Framework

doi:10.1109/TCSVT.2023.3326862

	Mind the Gap: Open Set Domain Adaptation via Mutual-to-Separate Framework
	Chang, Dongliang 1; Sain, Aneeshan 2; Ma, Zhanyu 1; Song, Yi-Zhe 2; Wang, Ruiping 3; Guo, Jun 1
	2024-06-01
发表期刊	IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
ISSN	1051-8215
卷号	34 期号:6 页码:4159-4174
摘要	Unsupervised domain adaptation aims to leverage labeled data from a source domain to learn a classifier for an unlabeled target domain. Amongst its many variants, open set domain adaptation (OSDA) is perhaps the most challenging one, as it further assumes the presence of unknown classes in the target domain. In this paper, we study OSDA with a particular focus on enriching its ability to traverse across larger domain gaps, and we show that existing state-of-the-art methods suffer a considerable performance drop in the presence of larger domain gaps, especially on a new dataset (PACS) that we re-purposed for OSDA. Exploring this is pivotal for OSDA as with increasing domain shift, identifying unknown samples in the target domain becomes harder for the model, thus making negative transfer between source and target domains more challenging. Accordingly, we propose a Mutual-to-Separate (MTS) framework to address the larger domain gaps. Essentially we design two networks - (a) Sample Separation Network (SSN): which is trained to learn a hyperplane for separating unknown samples from known ones, and (b) Distribution Matching Network (DMN): which is trained to maximise domain confusion between source and target domains without unknown samples under the guidance of the SSN. The key insight lies in how we exploit the mutually beneficial information between these two networks. On closer observation, we see that SSN can reveal which samples in the target domain belong to the unknown class by instance weighting whereas, DMN pushes apart the samples that most likely belong to the unknown class in the target domain, which in turn reduces the difficulty of SSN in identifying unknown samples. It follows that (a) and (b) will mutually supervise each other and alternate until convergence, which can better align the source and target domains in the shared label space. Extensive experiments on five datasets (Office-31, Office-Home, PACS, VisDA, and mini _DomainNet) demonstrate the efficiency of the proposed method. Detailed ablation experiments also validate the effectiveness of each component and the generality of the proposed framework. Codes are available at: https://github.com/PRIS-CV/Mutual-to-Separate.
关键词	Picture archiving and communication systems Training Task analysis Labeling Adaptation models Visualization Information exchange Domain adaptation open set mutual learning transfer learning
DOI	10.1109/TCSVT.2023.3326862
收录类别	SCI
语种	英语
资助项目	Beijing Natural Science Foundation
WOS研究方向	Engineering
WOS类目	Engineering, Electrical & Electronic
WOS记录号	WOS:001241605300060
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/39642
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Ma, Zhanyu
作者单位	1.Beijing Univ Posts & Telecommun, Sch Artificial Intelligence, Pattern Recognit & Intelligent Syst Lab, Beijing 100876, Peoples R China 2.Univ Surrey, Ctr Vis Speech & Signal Proc, Guildford GU2 7XH, England 3.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Chang, Dongliang,Sain, Aneeshan,Ma, Zhanyu,et al. Mind the Gap: Open Set Domain Adaptation via Mutual-to-Separate Framework[J]. IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY,2024,34(6):4159-4174.
APA	Chang, Dongliang,Sain, Aneeshan,Ma, Zhanyu,Song, Yi-Zhe,Wang, Ruiping,&Guo, Jun.(2024).Mind the Gap: Open Set Domain Adaptation via Mutual-to-Separate Framework.IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY,34(6),4159-4174.
MLA	Chang, Dongliang,et al."Mind the Gap: Open Set Domain Adaptation via Mutual-to-Separate Framework".IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 34.6(2024):4159-4174.