ROBUST FEATURE LEARNING AGAINST NOISY LABELS

被引:0
|
作者
Tai, Tsung-Ming [1 ,2 ]
Jhang, Yun-Jie [3 ]
Hwang, Wen-Jyi [4 ]
机构
[1] Free Univ Bozen Bolzano, Bolzano, Italy
[2] NVIDIA AI Technol Ctr, Hsinchu, Taiwan
[3] Natl Tsing Hua Univ, Int Intercollegiate PhD Program, Hsinchu, Taiwan
[4] Natl Taiwan Normal Univ, Dept Comp Sci & Informat Engn, Taipei, Taiwan
来源
2023 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, ICIP | 2023年
关键词
Image classification; noisy labels; robust feature learning;
D O I
10.1109/ICIP49359.2023.10222264
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Supervised learning of deep neural networks heavily relies on large-scale datasets annotated by high-quality labels. In contrast, mislabeled samples can significantly degrade the generalization of models and result in memorizing samples, further learning erroneous associations of data contents to incorrect annotations. To this end, this paper proposes an efficient approach to tackle noisy labels by learning robust feature representation based on unsupervised augmentation restoration and cluster regularization. In addition, progressive self-bootstrapping is introduced to minimize the negative impact of supervision from noisy labels. Our proposed design is generic and flexible in applying to existing classification architectures with minimal overheads. Experimental results show that our proposed method can efficiently and effectively enhance model robustness under severely noisy labels.
引用
收藏
页码:2235 / 2239
页数:5
相关论文
共 50 条
  • [1] Towards harnessing feature embedding for robust learning with noisy labels
    Zhang, Chuang
    Shen, Li
    Yang, Jian
    Gong, Chen
    MACHINE LEARNING, 2022, 111 (09) : 3181 - 3201
  • [2] Towards harnessing feature embedding for robust learning with noisy labels
    Chuang Zhang
    Li Shen
    Jian Yang
    Chen Gong
    Machine Learning, 2022, 111 : 3181 - 3201
  • [3] Robust Federated Learning With Noisy Labels
    Yang, Seunghan
    Park, Hyoungseob
    Byun, Junyoung
    Kim, Changick
    IEEE INTELLIGENT SYSTEMS, 2022, 37 (02) : 35 - 43
  • [4] Robust Collaborative Learning with Noisy Labels
    Sun, Mengying
    Xing, Jing
    Chen, Bin
    Zhou, Jiayu
    20TH IEEE INTERNATIONAL CONFERENCE ON DATA MINING (ICDM 2020), 2020, : 1274 - 1279
  • [5] Learning to rectify for robust learning with noisy labels
    Sun, Haoliang
    Guo, Chenhui
    Wei, Qi
    Han, Zhongyi
    Yin, Yilong
    PATTERN RECOGNITION, 2022, 124
  • [6] Robust Image Classification With Noisy Labels by Negative Learning and Feature Space Renormalization
    Wu, Hao
    Sun, Jun
    IEEE TRANSACTIONS ON MULTIMEDIA, 2024, 26 : 9280 - 9291
  • [7] Robust Federated Learning with Parameter Classification and Weighted Aggregation against Noisy Labels
    Li, Qun
    Duan, Congying
    Chen, Siguang
    IEEE CONFERENCE ON GLOBAL COMMUNICATIONS, GLOBECOM, 2023, : 2445 - 2450
  • [8] Penalty based robust learning with noisy labels
    Kong, Kyeongbo
    Lee, Junggi
    Kwak, Youngchul
    Cho, Young-Rae
    Kim, Seong-Eun
    Song, Woo-Jin
    NEUROCOMPUTING, 2022, 489 : 112 - 127
  • [9] Learning with noisy labels for robust fatigue detection
    Wang, Mei
    Hu, Ruimin
    Zhu, Xiaojie
    Zhu, Dongliang
    Wang, Xiaochen
    KNOWLEDGE-BASED SYSTEMS, 2024, 300
  • [10] Robust Training for Speaker Verification against Noisy Labels
    Fang, Zhihua
    He, Liang
    Ma, Hanhan
    Guo, Xiaochen
    Li, Lin
    INTERSPEECH 2023, 2023, : 3192 - 3196