On Feature Decorrelation in Self-Supervised Learning

被引:39
|
作者
Hua, Tianyu [1 ,2 ]
Wang, Wenxiao [1 ]
Xue, Zihui [2 ,3 ]
Ren, Sucheng [2 ,5 ]
Wang, Yue [4 ]
Zhao, Hang [1 ,2 ]
机构
[1] Tsinghua Univ, Beijing, Peoples R China
[2] Shanghai Qi Zhi Inst, Shanghai, Peoples R China
[3] UT Austin, Austin, TX USA
[4] MIT, Cambridge, MA 02139 USA
[5] South China Univ Technol, Guangzhou, Peoples R China
关键词
D O I
10.1109/ICCV48922.2021.00946
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In self-supervised representation learning, a common idea behind most of the state-of-the-art approaches is to enforce the robustness of the representations to predefined augmentations. A potential issue of this idea is the existence of completely collapsed solutions (i.e., constant features), which are typically avoided implicitly by carefully chosen implementation details. In this work, we study a relatively concise framework containing the most common components from recent approaches. We verify the existence of complete collapse and discover another reachable collapse pattern that is usually overlooked, namely dimensional collapse. We connect dimensional collapse with strong correlations between axes and consider such connection as a strong motivation for feature decorrelation (i.e., standardizing the covariance matrix). The gains from feature decorrelation are verified empirically to highlight the importance and the potential of this insight.
引用
收藏
页码:9578 / 9588
页数:11
相关论文
共 50 条
  • [41] Bridging the Gap from Asymmetry Tricks to Decorrelation Principles in Non-contrastive Self-supervised Learning
    Liu, Kang-Jun
    Suganuma, Masanori
    Okatani, Takayuki
    [J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
  • [42] Temporal Feature Alignment in Contrastive Self-Supervised Learning for Human Activity Recognition
    Khaertdinov, Bulat
    Asteriadis, Stylianos
    [J]. 2022 IEEE INTERNATIONAL JOINT CONFERENCE ON BIOMETRICS (IJCB), 2022,
  • [43] Domain Adaptation With Self-Supervised Learning and Feature Clustering for Intelligent Fault Diagnosis
    Lu, Nannan
    Xiao, Hanhan
    Ma, Zhanguo
    Yan, Tong
    Han, Min
    [J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2024, 35 (06) : 7657 - 7670
  • [44] Self-Supervised Learning of Neural Implicit Feature Fields for Camera Pose Refinement
    Pietrantoni, Maxime
    Csurka, Gabriela
    Humenberger, Martin
    Sattler, Torsten
    [J]. 2024 INTERNATIONAL CONFERENCE IN 3D VISION, 3DV 2024, 2024, : 484 - 494
  • [45] Joint data and feature augmentation for self-supervised representation learning on point clouds
    Lu, Zhuheng
    Dai, Yuewei
    Li, Weiqing
    Su, Zhiyong
    [J]. GRAPHICAL MODELS, 2023, 129
  • [46] Self-Supervised Visual Terrain Classification From Unsupervised Acoustic Feature Learning
    Zurn, Jannik
    Burgard, Wolfram
    Valada, Abhinav
    [J]. IEEE TRANSACTIONS ON ROBOTICS, 2021, 37 (02) : 466 - 481
  • [47] Self-Supervised Discriminative Feature Learning for Deep Multi-View Clustering
    Xu, Jie
    Ren, Yazhou
    Tang, Huayi
    Yang, Zhimeng
    Pan, Lili
    Yang, Yang
    Pu, Xiaorong
    Yu, Philip S.
    He, Lifang
    [J]. IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2023, 35 (07) : 7470 - 7482
  • [48] Self-supervised dynamic and static feature representation learning method for flotation monitoring
    Ai, Mingxi
    Xie, Yongfang
    Tang, Zhaohui
    Wu, Jiande
    Li, Peng
    Zhang, Jin
    [J]. POWDER TECHNOLOGY, 2024, 442
  • [49] Semantics-Consistent Feature Search for Self-Supervised Visual Representation Learning
    Song, Kaiyou
    Zhang, Shan
    Luo, Zimeng
    Wang, Tong
    Xie, Jin
    [J]. 2023 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2023), 2023, : 16053 - 16062
  • [50] Self-Supervised Feature Learning for Long-Term Metric Visual Localization
    Chen, Yuxuan
    Barfoot, Timothy D.
    [J]. IEEE ROBOTICS AND AUTOMATION LETTERS, 2023, 8 (02) : 472 - 479