• Home
  • Search
  • Uncertainty-Aware Learning against Label Noise on Imbalanced Datasets
  • Cite Icon52
  • https://doi.org/10.1609/aaai.v36i6.20654Copy DOI Icon

Uncertainty-Aware Learning against Label Noise on Imbalanced Datasets

Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Learning against label noise is a vital topic to guarantee a reliable performance for deep neural networks.Recent research usually refers to dynamic noise modeling with model output probabilities and loss values, and then separates clean and noisy samples.These methods have gained notable success. However, unlike cherry-picked data, existing approaches often cannot perform well when facing imbalanced datasets, a common scenario in the real world.We thoroughly investigate this phenomenon and point out two major issues that hinder the performance, i.e., inter-class loss distribution discrepancy and misleading predictions due to uncertainty.The first issue is that existing methods often perform class-agnostic noise modeling. However, loss distributions show a significant discrepancy among classes under class imbalance, and class-agnostic noise modeling can easily get confused with noisy samples and samples in minority classes.The second issue refers to that models may output misleading predictions due to epistemic uncertainty and aleatoric uncertainty, thus existing methods that rely solely on the output probabilities may fail to distinguish confident samples. Inspired by our observations, we propose an Uncertainty-aware Label Correction framework(ULC) to handle label noise on imbalanced datasets. First, we perform epistemic uncertainty-aware class-specific noise modeling to identify trustworthy clean samples and refine/discard highly confident true/corrupted labels.Then, we introduce aleatoric uncertainty in the subsequent learning process to prevent noise accumulation in the label noise modeling process. We conduct experiments on several synthetic and real-world datasets. The results demonstrate the effectiveness of the proposed method, especially on imbalanced datasets.

Similar Papers
  • Research Article
  • Citations12

Dynamic Loss For Robust Learning.

  • Dec 01, 2023
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Shenwang Jiang +4
  • Research Article
  • Citations11

Gradient-aware learning for joint biases: Label noise and class imbalance

  • Dec 18, 2023
  • Neural Networks
  • Shichuan Zhang +4
  • Research Article
  • Citations31

Noise-adaptive synthetic oversampling technique

  • Mar 18, 2021
  • Applied Intelligence
  • Minh Thanh Vo +3
  • Research Article
  • Citations1

Unleashing the potential of open-set noisy samples against label noise for medical image classification.

  • Oct 01, 2025
  • Medical image analysis
  • Zehui Liao +3
  • Research Article
  • Citations10

Regroup Median Loss for Combating Label Noise

  • Mar 24, 2024
  • Proceedings of the AAAI Conference on Artificial Intelligence
  • Fengpeng Li +3
  • Book Chapter
  • Citations1

CCLM: Class-Conditional Label Noise Modelling

  • Jan 01, 2023
  • Albert Tatjer +3
  • Research Article
  • Citations358

Image classification with deep learning in the presence of noisy labels: A survey

  • Jan 13, 2021
  • Knowledge-Based Systems
  • Görkem Algan +1
  • PDF
  • Research Article
  • Citations19

RN-Autoencoder: Reduced Noise Autoencoder for classifying imbalanced cancer genomic data

  • Jan 30, 2023
  • Journal of Biological Engineering
  • Ahmed Arafa +3
  • Research Article
  • Citations5

Label noise detection under the noise at random model with ensemble filters

  • Sep 05, 2022
  • Intelligent Data Analysis
  • Kecia G Moura +2
  • Conference Article
  • Citations1

Ensemble Methods for Label Noise Detection Under the Noisy at Random Model

  • Oct 01, 2018
  • Kecia G De Moura +2
  • Research Article
  • Citations4

WCM based hybrid pre-processing algorithm for class imbalanced dataset

  • Sep 15, 2021
  • Journal of Intelligent & Fuzzy Systems
  • Deepika Singh +2
  • Research Article
  • Citations9

A comparative study on noise filtering of imbalanced data sets

  • Jul 24, 2024
  • Knowledge-Based Systems
  • Szilvia Szeghalmy +1
  • Research Article
  • Citations2

Active Label Refinement for Robust Training of Imbalanced Medical Image Classification Tasks in the Presence of High Label Noise.

  • Jan 01, 2024
  • Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
  • Bidur Khanal +3
  • Research Article

Injury Severity Prediction for Older Driver Accidents via Denoised Cascade Framework and Probability Calibration

  • Apr 20, 2026
  • World Electric Vehicle Journal
  • Yiyong Pan +4
  • PDF
  • Research Article
  • Citations5

Performance of Supervised Learning Algorithms on Imbalanced Class Datasets

  • Aug 01, 2021
  • Journal of Physics: Conference Series
  • Nur Anisah Binti Ramli +3
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.