Next Article in Journal
Measurements of LoRaWAN Technology in Urban Scenarios: A Data Descriptor
Previous Article in Journal
Information Quality Assessment for Data Fusion Systems
Article

A Framework Using Contrastive Learning for Classification with Noisy Labels

R&D Department, EURA NOVA, 1435 Mont-Saint-Guibert, Belgium
*
Author to whom correspondence should be addressed.
These authors contributed equally to this work.
Academic Editors: M. Rashedur Rahman, Ken Barker, Luis Paulo F. Garcia, Nabeel Mohammed and Sifat Momen
Received: 29 April 2021 / Revised: 3 June 2021 / Accepted: 5 June 2021 / Published: 9 June 2021
(This article belongs to the Special Issue Machine Learning with Label Noise)
We propose a framework using contrastive learning as a pre-training task to perform image classification in the presence of noisy labels. Recent strategies, such as pseudo-labeling, sample selection with Gaussian Mixture models, and weighted supervised contrastive learning have, been combined into a fine-tuning phase following the pre-training. In this paper, we provide an extensive empirical study showing that a preliminary contrastive learning step brings a significant gain in performance when using different loss functions: non robust, robust, and early-learning regularized. Our experiments performed on standard benchmarks and real-world datasets demonstrate that: (i) the contrastive pre-training increases the robustness of any loss function to noisy labels and (ii) the additional fine-tuning phase can further improve accuracy, but at the cost of additional complexity. View Full-Text
Keywords: noisy labels; image classification; contrastive learning; robust loss noisy labels; image classification; contrastive learning; robust loss
Show Figures

Figure 1

MDPI and ACS Style

Ciortan, M.; Dupuis, R.; Peel, T. A Framework Using Contrastive Learning for Classification with Noisy Labels. Data 2021, 6, 61. https://0-doi-org.brum.beds.ac.uk/10.3390/data6060061

AMA Style

Ciortan M, Dupuis R, Peel T. A Framework Using Contrastive Learning for Classification with Noisy Labels. Data. 2021; 6(6):61. https://0-doi-org.brum.beds.ac.uk/10.3390/data6060061

Chicago/Turabian Style

Ciortan, Madalina, Romain Dupuis, and Thomas Peel. 2021. "A Framework Using Contrastive Learning for Classification with Noisy Labels" Data 6, no. 6: 61. https://0-doi-org.brum.beds.ac.uk/10.3390/data6060061

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop