Mining Label Distribution Drift in Unsupervised Domain Adaptation

被引:0
作者
Li, Peizhao [1 ]
Ding, Zhengming [2 ]
Liu, Hongfu [1 ]
机构
[1] Brandeis Univ, Waltham, MA 02254 USA
[2] Tulane Univ, New Orleans, LA USA
来源
ADVANCES IN ARTIFICIAL INTELLIGENCE, AI 2023, PT I | 2024年 / 14471卷
关键词
Unsupervised Domain Adaptation; Label Distribution Drift; Transfer Learning; Deep Learning;
D O I
10.1007/978-981-99-8388-9_29
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Unsupervised domain adaptation targets to transfer task-related knowledge from labeled source domain to unlabeled target domain. Although tremendous efforts have been made to minimize domain divergence, most existing methods only partially manage by aligning feature representations from diverse domains. Beyond the discrepancy in data distribution, the gap between source and target label distribution, recognized as label distribution drift, is another crucial factor raising domain divergence, and has been under insufficient exploration. From this perspective, we first reveal how label distribution drift brings negative influence. Next, we propose Label distribution Matching Domain Adversarial Network (LMDAN) to handle data distribution shift and label distribution drift jointly. In LMDAN, label distribution drift is addressed by a source sample weighting strategy, which selects samples that contribute to positive adaptation and avoid adverse effects brought by the mismatched samples. Experiments show that LMDAN delivers superior performance under considerable label distribution drift.
引用
收藏
页码:354 / 366
页数:13
相关论文
共 44 条
  • [1] Normalized Wasserstein for Mixture Distributions with Applications in Adversarial Learning and Domain Adaptation
    Balaji, Yogesh
    Chellappa, Rama
    Feizi, Soheil
    [J]. 2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, : 6509 - 6517
  • [2] A theory of learning from different domains
    Ben-David, Shai
    Blitzer, John
    Crammer, Koby
    Kulesza, Alex
    Pereira, Fernando
    Vaughan, Jennifer Wortman
    [J]. MACHINE LEARNING, 2010, 79 (1-2) : 151 - 175
  • [3] Re-weighted Adversarial Adaptation Network for Unsupervised Domain Adaptation
    Chen, Qingchao
    Liu, Yang
    Wang, Zhaowen
    Wassell, Ian
    Chetty, Kevin
    [J]. 2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, : 7976 - 7985
  • [4] Chen XY, 2019, PR MACH LEARN RES, V97
  • [5] Chhabra A., 2022, 11 INT C LEARNING RE
  • [6] Unsupervised Domain Adaptation via Regularized Conditional Alignment
    Cicek, Safa
    Soatto, Stefano
    [J]. 2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, : 1416 - 1425
  • [7] Optimal Transport for Domain Adaptation
    Courty, Nicolas
    Flamary, Remi
    Tuia, Devis
    Rakotomamonjy, Alain
    [J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2017, 39 (09) : 1853 - 1865
  • [8] DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptation
    Damodaran, Bharath Bhushan
    Kellenberger, Benjamin
    Flamary, Remi
    Tuia, Devis
    Courty, Nicolas
    [J]. COMPUTER VISION - ECCV 2018, PT IV, 2018, 11208 : 467 - 483
  • [9] Graph Adaptive Knowledge Transfer for Unsupervised Domain Adaptation
    Ding, Zhengming
    Li, Sheng
    Shao, Ming
    Fu, Yun
    [J]. COMPUTER VISION - ECCV 2018, PT II, 2018, 11206 : 36 - 52
  • [10] Deep Domain Generalization With Structured Low-Rank Constraint
    Ding, Zhengming
    Fu, Yun
    [J]. IEEE TRANSACTIONS ON IMAGE PROCESSING, 2018, 27 (01) : 304 - 313