LongReMix: Robust Learning with High Confidence Samples in a Noisy Label Environment

03/06/2021
by   Filipe R. Cordeiro, et al.
18

Deep neural network models are robust to a limited amount of label noise, but their ability to memorise noisy labels in high noise rate problems is still an open issue. The most competitive noisy-label learning algorithms rely on a 2-stage process comprising an unsupervised learning to classify training samples as clean or noisy, followed by a semi-supervised learning that minimises the empirical vicinal risk (EVR) using a labelled set formed by samples classified as clean, and an unlabelled set with samples classified as noisy. In this paper, we hypothesise that the generalisation of such 2-stage noisy-label learning methods depends on the precision of the unsupervised classifier and the size of the training set to minimise the EVR. We empirically validate these two hypotheses and propose the new 2-stage noisy-label training algorithm LongReMix. We test LongReMix on the noisy-label benchmarks CIFAR-10, CIFAR-100, WebVision, Clothing1M, and Food101-N. The results show that our LongReMix generalises better than competing approaches, particularly in high label noise problems. Furthermore, our approach achieves state-of-the-art performance in most datasets. The code will be available upon paper acceptance.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
10/17/2022

Bootstrapping the Relationship Between Images and Their Clean and Noisy Labels

Many state-of-the-art noisy-label learning methods rely on learning mech...
research
10/22/2021

PropMix: Hard Sample Filtering and Proportional MixUp for Learning with Noisy Labels

The most competitive noisy label learning methods rely on an unsupervise...
research
05/31/2023

Noisy-label Learning with Sample Selection based on Noise Rate Estimate

Noisy-labels are challenging for deep learning due to the high capacity ...
research
09/29/2021

Robust Temporal Ensembling for Learning with Noisy Labels

Successful training of deep neural networks with noisy labels is an esse...
research
03/09/2022

Noisy Label Learning for Security Defects

Data-driven software engineering processes, such as vulnerability predic...
research
07/21/2022

ProMix: Combating Label Noise via Maximizing Clean Sample Utility

The ability to train deep neural networks under label noise is appealing...
research
11/20/2018

Limited Gradient Descent: Learning With Noisy Labels

Label noise may handicap the generalization of classifiers, and it is an...

Please sign up or login with your details

Forgot password? Click here to reset