Label-Noise Learning with Intrinsically Long-Tailed Data

TL;DR

This paper introduces a novel two-stage sample selection framework called TABASCO designed to improve label-noise learning in long-tailed datasets, effectively distinguishing clean from noisy samples especially in tail classes.

Contribution

The paper proposes a new bi-dimensional sample selection method tailored for long-tailed data with label noise, addressing limitations of existing methods in imbalanced scenarios.

Findings

TABASCO outperforms existing methods on benchmark datasets.

The dual-metric approach improves noise separation in tail classes.

Extensive experiments validate the effectiveness of the proposed framework.

Abstract

Label noise is one of the key factors that lead to the poor generalization of deep learning models. Existing label-noise learning methods usually assume that the ground-truth classes of the training data are balanced. However, the real-world data is often imbalanced, leading to the inconsistency between observed and intrinsic class distribution with label noises. In this case, it is hard to distinguish clean samples from noisy samples on the intrinsic tail classes with the unknown intrinsic class distribution. In this paper, we propose a learning framework for label-noise learning with intrinsically long-tailed data. Specifically, we propose two-stage bi-dimensional sample selection (TABASCO) to better separate clean samples from noisy samples, especially for the tail classes. TABASCO consists of two new separation metrics that complement each other to compensate for the limitation of using a single metric in sample separation. Extensive experiments on benchmarks demonstrate the effectiveness of our method. Our code is available at https://github.com/Wakings/TABASCO.