Open set label noise learning with robust sample selection and margin-guided module

TL;DR

This paper proposes RSS-MGM, a novel method for learning with open set label noise that improves sample selection and distinguishes open from closed set noise using margin functions, outperforming existing methods.

Contribution

Introduces RSS-MGM, combining robust sample selection and margin-guided filtering to effectively handle open set label noise in deep learning.

Findings

Outperforms state-of-the-art label noise methods on benchmark datasets.

More accurately distinguishes open set from closed set label noise.

Effective in real-world noisy datasets like CIFAR-100N-C and Food101N.

Abstract

In recent years, the remarkable success of deep neural networks (DNNs) in computer vision is largely due to large-scale, high-quality labeled datasets. Training directly on real-world datasets with label noise may result in overfitting. The traditional method is limited to deal with closed set label noise, where noisy training data has true class labels within the known label space. However, there are some real-world datasets containing open set label noise, which means that some samples belong to an unknown class outside the known label space. To address the open set label noise problem, we introduce a method based on Robust Sample Selection and Margin-Guided Module (RSS-MGM). Firstly, unlike the prior clean sample selection approach, which only select a limited number of clean samples, a robust sample selection module combines small loss selection or high-confidence sample selection to obtain more clean samples. Secondly, to efficiently distinguish open set label noise and closed set ones, margin functions are designed to filter open-set data and closed set data. Thirdly, different processing methods are selected for different types of samples in order to fully utilize the data's prior information and optimize the whole model. Furthermore, extensive experimental results with noisy labeled data from benchmark datasets and real-world datasets, such as CIFAR-100N-C, CIFAR80N-O, WebFG-469, and Food101N, indicate that our approach outperforms many state-of-the-art label noise learning methods. Especially, it can more accurately divide open set label noise samples and closed set ones.