mlsanityData to 14 Oct 2026

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

arXiv 2110.11809 ↗First posted 22 Oct 2021cs.CVcited by 17

Not on the map. Why?

Abstract (arXiv)

The most competitive noisy label learning methods rely on an unsupervised classification of clean and noisy samples, where samples classified as noisy are re-labelled and "MixMatched" with the clean samples. These methods have two issues in large noise rate problems: 1) the noisy set is more likely to contain hard samples that are in-correctly re-labelled, and 2) the number of samples produced by MixMatch tends to be reduced because it is constrained by the small clean set size. In this paper, we introduce the learning algorithm PropMix to handle the issues above. PropMix filters out hard noisy samples, with the goal of increasing the likelihood of correctly re-labelling the easy noisy samples. Also, PropMix places clean and re-labelled easy noisy samples in a training set that is augmented with MixUp, removing the clean set size constraint and including a large proportion of correctly re-labelled easy noisy samples. We also include self-supervised pre-training to improve robustness to high noisy label scenarios. Our experiments show that PropMix has state-of-the-art (SOTA) results on CIFAR-10/-100(with symmetric, asymmetric and semantic label noise), Red Mini-ImageNet (from the Controlled Noisy Web Labels), Clothing1M and WebVision. In severe label noise bench-marks, our results are substantially better than other methods. The code is available athttps://github.com/filipe-research/PropMix.

Key facts. Cited by 17 papers in the map (data through 14 Oct 2026). Builds on 45 of its 57 references that are in the map. First posted on arXiv 22 Oct 2021.

Builds on

45 of 57 references.

References · 45

  1. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,364
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  3. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,513
  4. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
  5. 2020Improved Baselines with Momentum Contrastive Learningcited by 2,287
  6. 2016Identity Mappings in Deep Residual Networkscited by 1,965
  7. 2016Understanding deep learning requires rethinking generalizationcited by 1,674
  8. 2019MixMatch: A Holistic Approach to Semi-Supervised Learningcited by 1,391
  9. 2017A Survey on Deep Learning in Medical Image Analysiscited by 1,186
  10. 2018Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labelscited by 1,059
  11. 2020DivideMix: Learning with Noisy Labels as Semi-supervised Learningcited by 692
  12. 2018Learning to Reweight Examples for Robust Deep Learningcited by 675
33 more
  1. 2016Making Deep Neural Networks Robust to Label Noise: a Loss Correction Approachcited by 665
  2. 2020Image Segmentation Using Deep Learning: A Surveycited by 660
  3. 2017MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labelscited by 653
  4. 2019Symmetric Cross Entropy for Robust Learning with Noisy Labelscited by 500
  5. 2017Robust Loss Functions under Label Noise for Deep Neural Networkscited by 443
  6. 2020Early-Learning Regularization Prevents Memorization of Noisy Labelscited by 416
  7. 2019A Survey of Deep Learning Techniques for Autonomous Drivingcited by 403
  8. 2019How does Disagreement Help Generalization against Label Corruption?cited by 390
  9. 2020SCAN: Learning to Classify Images without Labelscited by 352
  10. 2019Unsupervised Label Noise Modeling and Loss Correctioncited by 328
  11. 2017WebVision Database: Visual Learning and Understanding from Web Datacited by 262
  12. 2020Identifying Mislabeled Data using the Area Under the Margin Rankingcited by 261
  13. 2020Normalized Loss Functions for Deep Learning with Noisy Labelscited by 257
  14. 2017Decoupling "when to update" from "how to update"cited by 230
  15. 2019Probabilistic End-to-end Noise Correction for Learning with Noisy Labelscited by 210
  16. 2019Understanding and Utilizing Deep Neural Networks Trained with Noisy Labelscited by 209
  17. 2017CleanNet: Transfer Learning for Scalable Image Classifier Training with Label Noisecited by 181
  18. 2018Dimensionality-Driven Learning with Noisy Labelscited by 173
  19. 2019SELF: Learning to Filter Noisy Labels with Self-Ensemblingcited by 164
  20. 2018Learning to Learn from Noisy Labeled Datacited by 160
  21. 2019NLNL: Negative Learning for Noisy Labelscited by 151
  22. 2019Deep Self-Learning From Noisy Labelscited by 122
  23. 2020Multi-Objective Interpolation Training for Robustness to Label Noisecited by 86
  24. 2021Augmentation Strategies for Learning with Noisy Labelscited by 86
  25. 2019Robust Inference via Generative Classifiers for Handling Noisy Labelscited by 68
  26. 2014DenseNet: Implementing Efficient ConvNet Descriptor Pyramidscited by 66
  27. 2021Faster Meta Update Strategy for Noise-Robust Deep Learningcited by 38
  28. 2019IMAE for Noise-Robust Learning: Mean Absolute Error Does Not Treat Examples Equally and Gradient Magnitude's Variance Matterscited by 33
  29. 2020ProSelfLC: Progressive Self Label Correction for Training Robust Deep Neural Networkscited by 33
  30. 2018A Semi-Supervised Two-Stage Approach to Learning from Noisy Labelscited by 29
  31. 2020Meta Soft Label Generation for Noisy Labelscited by 21
  32. 2019Towards Robust Learning with Different Label Noise Distributionscited by 14
  33. 2019Photometric Transformer Networks and Label Adjustment for Breast Density Predictioncited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 7; 2023: 5; 2024: 2; 2025: 1; 2026: 2.

17 papers.

Sort:

  1. 2022Instance-Dependent Noisy Label Learning via Graphical Modellingcited by 17
  2. 2022Is your noise correction noisy? PLS: Robustness to label noise with two stage detectioncited by 11
  3. 2023Manifold DivideMix: A Semi-Supervised Contrastive Learning Framework for Severe Label Noisecited by 9
  4. 2022Bootstrapping the Relationship Between Images and Their Clean and Noisy Labelscited by 7
  5. 2022Embedding contrastive unsupervised features to cluster in- and out-of-distribution noise in corrupted image datasetscited by 6
  6. 2023Data Optimization in Deep Learning: A Surveycited by 6
  7. 2023Instance-dependent Noisy-label Learning with Graphical Model Based Noise-rate Estimationcited by 4
  8. 2022A Study on the Impact of Data Augmentation for Training Convolutional Neural Networks in the Presence of Noisy Labelscited by 3
  9. 2024ANNE: Adaptive Nearest Neighbors and Eigenvector-based Sample Selection for Robust Learning with Noisy Labelscited by 3
  10. 2023Learning with Noisy Labels Using Collaborative Sample Selection and Contrastive Semi-Supervised Learningcited by 2
  11. 2022Maximising the Utility of Validation Sets for Imbalanced Noisy-label Meta-learningcited by 1
  12. 2023PASS: Peer-Agreement based Sample Selection for training with Noisy Labelscited by 1
5 more
  1. 2025Adaptively Point-weighting Curriculum Learningcited by 1
  2. 2026Deconstructing the Failure of Ideal Noise Correction: A Three-Pillar Diagnosiscited by 1
  3. 2022Class Prototype-based Cleaner for Label Noise Learningcited by 0
  4. 2024An accurate detection is not all you need to combat label noise in web-noisy datasetscited by 0
  5. 2026Data-Efficient Semantic Segmentation of 3D Point Clouds via Open-Vocabulary Image Segmentation-based Pseudo-Labelingcited by 0