IEICE Transactions on Information and Systems
Online ISSN : 1745-1361
Print ISSN : 0916-8532
Special Section on Enriched Multimedia — Media technologies opening up the future —
Unbiased Pseudo-Labeling for Learning with Noisy Labels
Ryota HIGASHIMOTOSoh YOSHIDATakashi HORIHATAMitsuji MUNEYASU
Author information
JOURNAL FREE ACCESS

2024 Volume E107.D Issue 1 Pages 44-48

Details
Abstract

Noisy labels in training data can significantly harm the performance of deep neural networks (DNNs). Recent research on learning with noisy labels uses a property of DNNs called the memorization effect to divide the training data into a set of data with reliable labels and a set of data with unreliable labels. Methods introducing semi-supervised learning strategies discard the unreliable labels and assign pseudo-labels generated from the confident predictions of the model. So far, this semi-supervised strategy has yielded the best results in this field. However, we observe that even when models are trained on balanced data, the distribution of the pseudo-labels can still exhibit an imbalance that is driven by data similarity. Additionally, a data bias is seen that originates from the division of the training data using the semi-supervised method. If we address both types of bias that arise from pseudo-labels, we can avoid the decrease in generalization performance caused by biased noisy pseudo-labels. We propose a learning method with noisy labels that introduces unbiased pseudo-labeling based on causal inference. The proposed method achieves significant accuracy gains in experiments at high noise rates on the standard benchmarks CIFAR-10 and CIFAR-100.

Content from these authors
© 2024 The Institute of Electronics, Information and Communication Engineers
Previous article Next article
feedback
Top