Instance Correction for Learning with Open-set Noisy Labels

Abstract

The problem of open-set noisy labels denotes that part of training data have a different label space that does not contain the true class. Lots of approaches, e.g., loss correction and label correction, cannot handle such open-set noisy labels well, since they need training data and test data to share the same label space, which does not hold for learning with open-set noisy labels. The state-of-the-art methods thus employ the sample selection approach to handle open-set noisy labels, which tries to select clean data from noisy data for network parameters updates. The discarded data are seen to be mislabeled and do not participate in training. Such an approach is intuitive and reasonable at first glance. However, a natural question could be raised "can such data only be discarded during training?". In this paper, we show that the answer is no. Specifically, we discuss that the instances of discarded data could consist of some meaningful information for generalization. For this reason, we do not abandon such data, but use instance correction to modify the instances of the discarded data, which makes the predictions for the discarded data consistent with given labels. Instance correction are performed by targeted adversarial attacks. The corrected data are then exploited for training to help generalization. In addition to the analytical results, a series of empirical evidences are provided to justify our claims.

References in corpus (20)

• DivideMix: Learning with Noisy Labels as Semi-supervised Learning
• Characterizing Adversarial Subspaces Using Local Intrinsic Dimensionality
• Adversarial Examples Are Not Bugs, They Are Features
• A Downsampled Variant of ImageNet as an Alternative to the CIFAR datasets
• A Closer Look at Memorization in Deep Networks
• WebVision Database: Visual Learning and Understanding from Web Data
• On the Convergence and Robustness of Adversarial Training
• Free Lunch for Few-shot Learning: Distribution Calibration
• advertorch v0.1: An Adversarial Robustness Toolbox based on PyTorch
• Normalized Loss Functions for Deep Learning with Noisy Labels
• Beyond Synthetic Noise: Deep Learning on Controlled Noisy Labels
• Joint Optimization Framework for Learning with Noisy Labels
• Understanding and Utilizing Deep Neural Networks Trained with Noisy Labels
• SIGUA: Forgetting May Make Learning with Noisy Labels More Robust
• Improving Adversarial Robustness via Channel-wise Activation Suppressing
• Identifying Mislabeled Data using the Area Under the Margin Ranking
• Error-Bounded Correction of Noisy Labels
• Learning with Feature-Dependent Label Noise: A Progressive Approach
• Meta Transition Adaptation for Robust Deep Learning with Noisy Labels
• Class2Simi: A Noise Reduction Perspective on Learning with Noisy Labels