Domain Adaptation with Adversarial Training on Penultimate Activations

被引:0
|
作者
Sun, Tao [1 ]
Lu, Cheng [2 ]
Ling, Haibin [1 ]
机构
[1] SUNY Stony Brook, Stony Brook, NY 11794 USA
[2] XPeng Motors, Guangzhou, Peoples R China
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Enhancing model prediction confidence on target data is an important objective in Unsupervised Domain Adaptation (UDA). In this paper, we explore adversarial training on penultimate activations, i.e., input features of the final linear classification layer. We show that this strategy is more efficient and better correlated with the objective of boosting prediction confidence than adversarial training on input images or intermediate features, as used in previous works. Furthermore, with activation normalization commonly used in domain adaptation to reduce domain gap, we derive two variants and systematically analyze the effects of normalization on our adversarial training. This is illustrated both in theory and through empirical analysis on real adaptation tasks. Extensive experiments are conducted on popular UDA benchmarks under both standard setting and source-data free setting. The results validate that our method achieves the best scores against previous arts. Code is available at https://github.com/tsun/APA.
引用
收藏
页码:9935 / 9943
页数:9
相关论文
共 50 条
  • [41] Unsupervised Domain Adaptation for Dialogue Sequence Labeling Based on Hierarchical Adversarial Training
    Orihashi, Shota
    Ihori, Mana
    Tanaka, Tomohiro
    Masumura, Ryo
    INTERSPEECH 2020, 2020, : 1575 - 1579
  • [42] Self-Training Based Adversarial Domain Adaptation for Radio Signal Recognition
    Liang, Zhi
    Xie, Jian
    Yang, Xin
    Tao, Mingliang
    Wang, Ling
    IEEE COMMUNICATIONS LETTERS, 2022, 26 (11) : 2646 - 2650
  • [43] Visualizing and Analyzing the Topology of Neuron Activations in Deep Adversarial Training
    Zhou, Youjia
    Zhou, Yi
    Ding, Jie
    Wang, Bei
    TOPOLOGICAL, ALGEBRAIC AND GEOMETRIC LEARNING WORKSHOPS 2023, VOL 221, 2023, 221
  • [44] Unsupervised Domain Adaptation for Dysarthric Speech Detection via Domain Adversarial Training and Mutual Information Minimization
    Wang, Disong
    Deng, Liqun
    Yeung, Yu Ting
    Chen, Xiao
    Liu, Xunying
    Meng, Helen
    INTERSPEECH 2021, 2021, : 2956 - 2960
  • [45] Unsupervised domain adaptation with adversarial distribution adaptation network
    Zhou, Qiang
    Zhou, Wen'an
    Wang, Shirui
    Xing, Ying
    NEURAL COMPUTING & APPLICATIONS, 2021, 33 (13): : 7709 - 7721
  • [46] Adversarial Multiple Source Domain Adaptation
    Zhao, Han
    Zhang, Shanghang
    Wu, Guanhang
    Costeira, Joao P.
    Moura, Jose M. F.
    Gordon, Geoffrey J.
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018), 2018, 31
  • [47] Adversarial Reweighting for Partial Domain Adaptation
    Gu, Xiang
    Yu, Xi
    Yang, Yan
    Sun, Jian
    Xu, Zongben
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021), 2021, 34
  • [48] Targeted adversarial discriminative domain adaptation
    Chen, Hua-Mei
    Savakis, Andreas
    Diehl, Ashley
    Blasch, Erik
    Wei, Sixiao
    Chen, Genshe
    JOURNAL OF APPLIED REMOTE SENSING, 2021, 15 (03)
  • [49] Unsupervised domain adaptation with adversarial distribution adaptation network
    Qiang Zhou
    Wen’an Zhou
    Shirui Wang
    Ying Xing
    Neural Computing and Applications, 2021, 33 : 7709 - 7721
  • [50] Targeted Adversarial Discriminative Domain Adaptation
    Chen, Hua-Mei
    Savakis, Andreas
    Diehl, Ashley
    Blasch, Erik
    Wei, Sixiao
    Chen, Genshe
    GEOSPATIAL INFORMATICS XI, 2021, 11733