Domain generalization based on domain-specific adversarial learning

被引:1
|
作者
Wang, Ziping [1 ,2 ]
Zhang, Xiaohang [1 ,2 ]
Li, Zhengren [3 ]
Chen, Fei [3 ]
机构
[1] Beijing Univ Posts & Telecommun, Sch Econ & Management, Beijing 100876, Peoples R China
[2] Beijing Univ Posts & Telecommun, Key Lab Trustworthy Distributed Comp & Serv, Beijing 100876, Peoples R China
[3] Beijing Univ Posts & Telecommun, Sch Modern Posts, Beijing 100876, Peoples R China
基金
中国国家自然科学基金;
关键词
Domain generalization; Adversarial learning; Distribution alignment; Transfer learning;
D O I
10.1007/s10489-024-05423-z
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Deep learning models often suffer from degraded performance when the distributions of the training and testing data differ (i.e., domain shift). Domain generalization (DG) techniques can help improve the generalization performance for unseen target domains by using multiple source domains. The recently developed domain generalization methods focus on extracting domain-invariant features from all source domains. However, some task-relevant discriminative information can be removed during this process. In addition, the various source domains are treated equally ignoring the negative impacts of distant source domains. Both problems can lead to unsatisfactory performance. This paper proposed a domain-specific adversarial neural network (DSANN) based on adversarial learning to learn effective feature representations and reduce the influence of distantsource domains. The DSANN introduces a reference distribution that is adaptively generated during training. Additionally, domain-invariant features are extracted through a domain-specific adversarial learning process , in which each source domain distribution is aligned only with the reference distribution instead of all the other source domains. Moreover, the DSANN also aligns the outputs of multiple classifiers and adopts the weighted average of the predictions; thus, the employed label classifiers can become more robust to unknown domain shifts. Experiments conducted on popular benchmark datasets demonstrate that our proposed method can achieve remarkable generalization performance and has better classification accuracy than the existing DG algorithms.
引用
收藏
页码:4878 / 4889
页数:12
相关论文
共 50 条
  • [41] Untangling Crosscutting Concerns in Domain-specific Languages with Domain-specific Join Points
    Dinkelaker, Tom
    Monperrus, Martin
    Mezini, Mira
    DSAL09: DOMAIN-SPECIFIC ASPECT LANGUAGES, 2009, : 1 - 5
  • [42] Adversarial data splitting for domain generalization
    Gu, Xiang
    Sun, Jian
    Xu, Zongben
    SCIENCE CHINA-INFORMATION SCIENCES, 2024, 67 (05)
  • [43] An Approach for Domain-Specific Design Pattern Identification Based on Domain Ontology
    Gkantouna, Vassiliki
    Papaioannou, Vaios
    Tzimas, Giannis
    Sabic, Zlatan
    ARTIFICIAL INTELLIGENCE APPLICATIONS AND INNOVATIONS (AIAI 2019), 2019, 560 : 125 - 137
  • [44] Feature Stylization Adversarial Domain Generalization
    Hu, Zhengzhong
    2023 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN, 2023,
  • [45] Adversarial data splitting for domain generalization
    Xiang GU
    Jian SUN
    Zongben XU
    Science China(Information Sciences), 2024, 67 (05) : 28 - 42
  • [46] Adversarial data splitting for domain generalization
    Xiang Gu
    Jian Sun
    Zongben Xu
    Science China Information Sciences, 2024, 67
  • [47] Adversarial Reconstruction Loss for Domain Generalization
    Bekkouch, Imad Eddine Ibrahim
    Nicolae, Dragos Constantin
    Khan, Adil
    Kazmi, S. M. Ahsan
    Khattak, Asad Masood
    Ibragimov, Bulat
    IEEE ACCESS, 2021, 9 : 42424 - 42437
  • [48] A Domain-Specific Language for Aviation Domain Interoperability
    Comitz, Paul
    2013 INTEGRATED COMMUNICATIONS, NAVIGATION AND SURVEILLANCE CONFERENCE (ICNS), 2013,
  • [49] DOMAIN-SPECIFIC SOCIAL JUDGMENTS AND DOMAIN AMBIGUITIES
    TURIEL, E
    MERRILL-PALMER QUARTERLY-JOURNAL OF DEVELOPMENTAL PSYCHOLOGY, 1989, 35 (01): : 89 - 114
  • [50] Lifelong Learning of Topics and Domain-Specific Word Embeddings
    Qin, Xiaorui
    Lu, Yuyin
    Chen, Yufu
    Rao, Yanghui
    FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, ACL-IJCNLP 2021, 2021, : 2294 - 2309