Learning with Noisy labels via Self-supervised Adversarial Noisy Masking

TL;DR

This paper introduces a novel adversarial noisy masking technique that leverages deep feature map analysis and self-supervised reconstruction to improve deep learning robustness against noisy labels.

Contribution

It proposes a new regularization method using label quality guided masking and self-supervised reconstruction, addressing noisy labels from a feature distribution perspective.

Findings

Significant performance improvements over state-of-the-art methods.

Effective in both synthetic and real-world noisy datasets.

Enhances model robustness by preventing overfitting to noisy labels.

Abstract

Collecting large-scale datasets is crucial for training deep models, annotating the data, however, inevitably yields noisy labels, which poses challenges to deep learning algorithms. Previous efforts tend to mitigate this problem via identifying and removing noisy samples or correcting their labels according to the statistical properties (e.g., loss values) among training samples. In this paper, we aim to tackle this problem from a new perspective, delving into the deep feature maps, we empirically find that models trained with clean and mislabeled samples manifest distinguishable activation feature distributions. From this observation, a novel robust training approach termed adversarial noisy masking is proposed. The idea is to regularize deep features with a label quality guided masking scheme, which adaptively modulates the input data and label simultaneously, preventing the model to overfit noisy samples. Further, an auxiliary task is designed to reconstruct input data, it naturally provides noise-free self-supervised signals to reinforce the generalization ability of deep models. The proposed method is simple and flexible, it is tested on both synthetic and real-world noisy datasets, where significant improvements are achieved over previous state-of-the-art methods.