Learning advisor networks for noisy image classification

TL;DR

This paper proposes a novel advisor network approach for noisy image classification that weights features rather than loss, enabling better utilization of noisy data and achieving state-of-the-art results on benchmark datasets.

Contribution

Introduces a new advisor network that weights features directly, trained with meta-learning, to improve noisy label handling in image classification.

Findings

Achieved state-of-the-art results on CIFAR10 and CIFAR100 with synthetic noise.

Performed well on Clothing1M with real-world noisy labels.

Outperformed existing noise-robust methods.

Abstract

In this paper, we introduced the novel concept of advisor network to address the problem of noisy labels in image classification. Deep neural networks (DNN) are prone to performance reduction and overfitting problems on training data with noisy annotations. Weighting loss methods aim to mitigate the influence of noisy labels during the training, completely removing their contribution. This discarding process prevents DNNs from learning wrong associations between images and their correct labels but reduces the amount of data used, especially when most of the samples have noisy labels. Differently, our method weighs the feature extracted directly from the classifier without altering the loss value of each data. The advisor helps to focus only on some part of the information present in mislabeled examples, allowing the classifier to leverage that data as well. We trained it with a meta-learning strategy so that it can adapt throughout the training of the main model. We tested our method on CIFAR10 and CIFAR100 with synthetic noise, and on Clothing1M which contains real-world noise, reporting state-of-the-art results.