Reusing the Task-specific Classifier as a Discriminator: Discriminator-free Adversarial Domain Adaptation

被引:56
|
作者
Chen, Lin [1 ]
Chen, Huaian [1 ]
Wei, Zhixiang [1 ]
Jin, Xin [1 ]
Tan, Xiao [1 ]
Jin, Yi [1 ]
Chen, Enhong [1 ]
机构
[1] Univ Sci & Technol China, Hefei, Peoples R China
基金
中国国家自然科学基金;
关键词
D O I
10.1109/CVPR52688.2022.00704
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Adversarial learning has achieved remarkable performances for unsupervised domain adaptation (UDA). Existing adversarial UDA methods typically adopt an additional discriminator to play the min-max game with a feature extractor. However, most of these methods failed to effectively leverage the predicted discriminative information, and thus cause mode collapse for generator. In this work, we address this problem from a different perspective and design a simple yet effective adversarial paradigm in the form of a discriminator-free adversarial learning network (DALN), wherein the category classifier is reused as a discriminator, which achieves explicit domain alignment and category distinguishment through a unified objective, enabling the DALN to leverage the predicted discriminative information for sufficient feature alignment. Basically, we introduce a Nuclear-norm Wasserstein discrepancy (NWD) that has definite guidance meaning for performing discrimination. Such NWD can be coupled with the classifier to serve as a discriminator satisfying the K-Lipschitz constraint without the requirements of additional weight clipping or gradient penalty strategy. Without bells and whistles, DALN compares favorably against the existing state-of-the-art (SOTA) methods on a variety of public datasets. Moreover, as a plug-and-play technique, NWD can be directly used as a generic regularizer to benefit existing UDA algorithms. Code is available at https://github.com/xiaoachen98/DALN.
引用
收藏
页码:7171 / 7180
页数:10
相关论文
共 50 条
  • [1] Discriminator-Free Generative Adversarial Attack
    Lu, Shaohao
    Xian, Yuqiao
    Yan, Ke
    Hu, Yi
    Sun, Xing
    Guo, Xiaowei
    Huang, Feiyue
    Zheng, Wei-Shi
    PROCEEDINGS OF THE 29TH ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA, MM 2021, 2021, : 1544 - 1552
  • [2] ADVERSARIAL DOMAIN ADAPTATION WITH A DOMAIN SIMILARITY DISCRIMINATOR FOR SEMANTIC SEGMENTATION OF URBAN AREAS
    Yan, Liang
    Fan, Bin
    Xiang, Shiming
    Pan, Chunhong
    2018 25TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), 2018, : 1583 - 1587
  • [3] Re-energizing Domain Discriminator with Sample Relabeling for Adversarial Domain Adaptation
    Jin, Xin
    Lan, Cuiling
    Zeng, Wenjun
    Chen, Zhibo
    2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, : 9154 - 9163
  • [4] Informative discriminator for domain adaptation
    Kurmi, Vinod K.
    Subramanian, Venkatesh K.
    Namboodiri, Vinay P.
    IMAGE AND VISION COMPUTING, 2021, 111
  • [5] Exploring dropout discriminator for domain adaptation
    Kurmi, Vinod Kumar
    Subramanian, Venkatesh K.
    Namboodiri, Vinay P.
    NEUROCOMPUTING, 2021, 457 : 168 - 181
  • [6] Adversarial task-specific learning
    Fu, Xin
    Zhao, Yao
    Liu, Ting
    Wei, Yunchao
    Li, Jianan
    Wei, Shikui
    NEUROCOMPUTING, 2019, 362 : 118 - 128
  • [7] Unifying TopDown Views by Task-Specific Domain Adaptation
    Lin, Jianzhe
    Yu, Tianze
    Mou, Lichao
    Zhu, Xiaoxiang
    Ward, Rabab Kreidieh
    Wang, Z. Jane
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2021, 59 (06): : 4689 - 4702
  • [8] Domain Adaptation via a Task-Specific Classifier Framework for Remote Sensing Cross-Scene Classification
    Zheng, Zhendong
    Zhong, Yanfei
    Su, Yu
    Ma, Ailong
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2022, 60
  • [9] Disentangled Discriminator for Unsupervised Domain Adaptation on Object Detection
    Zhu, Yangguang
    Guo, Ping
    Wei, Haoran
    Zhao, Xin
    Wu, Xiangbin
    2023 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2023, : 5685 - 5691
  • [10] TMDA: Task-Specific Multi-Source Domain Adaptation via Clustering Embedded Adversarial Training
    Wang, Haotian
    Yang, Wenjing
    Lin, Zhipeng
    Yu, Yue
    2019 19TH IEEE INTERNATIONAL CONFERENCE ON DATA MINING (ICDM 2019), 2019, : 1372 - 1377