Do We Really Need Gold Samples for Sample Weighting Under Label Noise?

Abstract

Learning with labels noise has gained significant traction recently due to the sensitivity of deep neural networks under label noise under common loss functions. Losses that are theoretically robust to label noise, however, often makes training difficult. Consequently, several recently proposed methods, such as Meta-Weight-Net (MW-Net), use a small number of unbiased, clean samples to learn a weighting function that downweights samples that are likely to have corrupted labels under the meta-learning framework. However, obtaining such a set of clean samples is not always feasible in practice. In this paper, we analytically show that one can easily train MW-Net without access to clean samples simply by using a loss function that is robust to label noise, such as mean absolute error, as the meta objective to train the weighting network. We experimentally show that our method beats all existing methods that do not use clean samples and performs on-par with methods that use gold samples on benchmark datasets across various noise types and noise rates.

arXiv commentWACV 2021

References in corpus (17)

• Prototypical Networks for Few-shot Learning
• Focal Loss for Dense Object Detection
• Image Classification with Deep Learning in the Presence of Noisy Labels: A Survey
• Making Risk Minimization Tolerant to Label Noise
• Masking: A New Perspective of Noisy Supervision
• How does Disagreement Help Generalization against Label Corruption?
• Robust Loss Functions under Label Noise for Deep Neural Networks
• Symmetric Cross Entropy for Robust Learning with Noisy Labels
• Using Trusted Data to Train Deep Networks on Labels Corrupted by Severe Noise
• Bilevel Programming for Hyperparameter Optimization and Meta-Learning
• Probabilistic End-to-end Noise Correction for Learning with Noisy Labels
• On Symmetric Losses for Learning from Corrupted Labels
• Learning from Noisy Labels with Distillation
• Double Ramp Loss Based Reject Option Classifier
• Loss factorization, weakly supervised learning and label noise robustness
• L_DMI: An Information-theoretic Noise-robust Loss Function
• Iterative Learning with Open-set Noisy Labels