Learning Confident Classifiers in the Presence of Label Noise
Hashmi, Asma Ahmed Zhumabayeva, Aigerim Kotelevskii, Nikita Agafonov, Artem Yaqub, Mohammad Panov, Maxim Takac, Martin
Hashmi, Asma Ahmed
Zhumabayeva, Aigerim
Kotelevskii, Nikita
Agafonov, Artem
Yaqub, Mohammad
Panov, Maxim
Takac, Martin
Supervisor
Department
Machine Learning
Embargo End Date
Type
Conference proceeding
Date
2025
License
Language
English
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
The success of Deep Neural Network (DNN) models significantly depends on the quality of provided annotations. In medical image segmentation, for example, having multiple expert annotations for each data point is standard to minimize subjective annotation bias. Then, the goal of estimation is to filter out the label noise and recover the ground-truth masks, which are not explicitly given. This paper proposes a probabilistic model for noisy observations that allows us to build confident classification and segmentation models. We explicitly model label noise to accomplish this and introduce a new information-based regularization that pushes the network to recover the ground-truth labels. In addition, we adjust the loss function for the segmentation task by prioritizing learning in high-confidence regions where all the annotators agree on labeling. We evaluate the proposed method on a series of classification tasks such as noisy versions of MNIST, CIFAR-10, and Fashion-MNIST datasets, as well as CIFAR-10N, a real-world dataset with noisy human annotations. Additionally, for the segmentation task, we consider several medical imaging datasets, such as LIDC and RIGA, that reflect real-world inter-variability among multiple annotators. Our experiments show that our algorithm outperforms state-of-the-art solutions for the considered classification and segmentation problems. Copyright © 2025 by SIAM.
Citation
A. A. Hashmi et al., “Learning Confident Classifiers in the Presence of Label Noise,” Proc West Mark Ed Assoc Conf, pp. 81–113, Jan. 2025, doi: 10.1137/1.9781611978520.9
Source
2025 SIAM International Conference on Data Mining, SDM 2025
Conference
2025 SIAM International Conference on Data Mining, SDM 2025
Keywords
Subjects
Source
2025 SIAM International Conference on Data Mining, SDM 2025
Publisher
Society for Industrial and Applied Mathematics Publications