Iterative Learning with Open-set Noisy Labels

TL;DR

This paper introduces an iterative learning framework for CNNs that effectively handles datasets with open-set noisy labels, improving robustness by detecting noise, learning discriminative features, and reweighting samples.

Contribution

It proposes a novel open-set noisy label training method using iterative detection, a Siamese network for dissimilarity enforcement, and a reweighting module to enhance robustness.

Findings

Effective on CIFAR-10, ImageNet, and web datasets.

Robustly handles high proportions of open-set and closed-set noise.

Improves CNN training accuracy in noisy label scenarios.

Abstract

Large-scale datasets possessing clean label annotations are crucial for training Convolutional Neural Networks (CNNs). However, labeling large-scale data can be very costly and error-prone, and even high-quality datasets are likely to contain noisy (incorrect) labels. Existing works usually employ a closed-set assumption, whereby the samples associated with noisy labels possess a true class contained within the set of known classes in the training data. However, such an assumption is too restrictive for many applications, since samples associated with noisy labels might in fact possess a true class that is not present in the training data. We refer to this more complex scenario as the \textbf{open-set noisy label} problem and show that it is nontrivial in order to make accurate predictions. To address this problem, we propose a novel iterative learning framework for training CNNs on datasets with open-set noisy labels. Our approach detects noisy labels and learns deep discriminative features in an iterative fashion. To benefit from the noisy label detection, we design a Siamese network to encourage clean labels and noisy labels to be dissimilar. A reweighting module is also applied to simultaneously emphasize the learning from clean labels and reduce the effect caused by noisy labels. Experiments on CIFAR-10, ImageNet and real-world noisy (web-search) datasets demonstrate that our proposed model can robustly train CNNs in the presence of a high proportion of open-set as well as closed-set noisy labels.