Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - OHD
T2 - An Online Category-Aware Framework for Learning with Noisy Labels under Long-Tailed Distribution
AU - Zhao, Qihao
AU - Zhang, Fan
AU - Hu, Wei
AU - Feng, Songhe
AU - Liu, Jun
PY - 2024/5/1
Y1 - 2024/5/1
N2 - Recently, many effective methods have emerged to address the robustness problem of Deep Neural Networks (DNNs) trained with noisy labels. However, existing work on learning with noisy labels (LNL) mainly focuses on balanced datasets, while real-world scenarios usually also exhibit a long-tailed distribution (LTD). In this paper, we propose an online category-aware approach to mitigate the impact of noisy labels and LTD on the robustness of DNNs. First, the category frequency of clean samples used to rebalance the feature space cannot be obtained directly in the presence of noisy samples. We design a novel category-aware Online Joint Distribution to dynamically estimate the category frequency of clean samples. Second, previous LNL methods were category-agnostic. These methods would easily be confused with noisy samples and tail categories' samples under LTD. Based on this observation, we propose a Harmonizing Factor strategy to exploit more information from the category-aware online joint distribution. This strategy provides more accurate estimates of clean samples between noisy samples and samples with tail categories. Finally, we propose Dynamic Cost-sensitive Learning, which utilizes the loss and category frequency of the estimated clean samples to address both LNL and LTD. Compared to extensive state-of-the-art methods, our strategy consistently improves the generalization performance of DNNs on several synthetic datasets and two real-world datasets.
AB - Recently, many effective methods have emerged to address the robustness problem of Deep Neural Networks (DNNs) trained with noisy labels. However, existing work on learning with noisy labels (LNL) mainly focuses on balanced datasets, while real-world scenarios usually also exhibit a long-tailed distribution (LTD). In this paper, we propose an online category-aware approach to mitigate the impact of noisy labels and LTD on the robustness of DNNs. First, the category frequency of clean samples used to rebalance the feature space cannot be obtained directly in the presence of noisy samples. We design a novel category-aware Online Joint Distribution to dynamically estimate the category frequency of clean samples. Second, previous LNL methods were category-agnostic. These methods would easily be confused with noisy samples and tail categories' samples under LTD. Based on this observation, we propose a Harmonizing Factor strategy to exploit more information from the category-aware online joint distribution. This strategy provides more accurate estimates of clean samples between noisy samples and samples with tail categories. Finally, we propose Dynamic Cost-sensitive Learning, which utilizes the loss and category frequency of the estimated clean samples to address both LNL and LTD. Compared to extensive state-of-the-art methods, our strategy consistently improves the generalization performance of DNNs on several synthetic datasets and two real-world datasets.
KW - Deep neural networks
KW - image classification
KW - learning with noisy labels
KW - long-tailed distribution
U2 - 10.1109/TCSVT.2023.3321733
DO - 10.1109/TCSVT.2023.3321733
M3 - Journal article
AN - SCOPUS:85174848345
VL - 34
SP - 3806
EP - 3818
JO - IEEE Transactions on Circuits and Systems for Video Technology
JF - IEEE Transactions on Circuits and Systems for Video Technology
SN - 1051-8215
IS - 5
ER -