Estimating Noisy Class Posterior with Part-level Labels for Noisy Label Learning

TL;DR

This paper introduces a novel approach for noisy label learning that leverages part-level labels to improve the estimation of noisy class posteriors, leading to more accurate classifiers.

Contribution

It proposes a method that incorporates part-level labels and a single-to-multiple transition matrix to enhance noisy class posterior estimation.

Findings

Improves estimation accuracy of noisy class posteriors.

Enhances classifier performance on synthetic and real-world noisy datasets.

Theoretically grounded and empirically validated.

Abstract

In noisy label learning, estimating noisy class posteriors plays a fundamental role for developing consistent classifiers, as it forms the basis for estimating clean class posteriors and the transition matrix. Existing methods typically learn noisy class posteriors by training a classification model with noisy labels. However, when labels are incorrect, these models may be misled to overemphasize the feature parts that do not reflect the instance characteristics, resulting in significant errors in estimating noisy class posteriors. To address this issue, this paper proposes to augment the supervised information with part-level labels, encouraging the model to focus on and integrate richer information from various parts. Specifically, our method first partitions features into distinct parts by cropping instances, yielding part-level labels associated with these various parts. Subsequently, we introduce a novel single-to-multiple transition matrix to model the relationship between the noisy and part-level labels, which incorporates part-level labels into a classifier-consistent framework. Utilizing this framework with part-level labels, we can learn the noisy class posteriors more precisely by guiding the model to integrate information from various parts, ultimately improving the classification performance. Our method is theoretically sound, while experiments show that it is empirically effective in synthetic and real-world noisy benchmarks.