mlsanityData to 7 Oct 2026

Learning Noise Transition Matrix from Only Noisy Labels via Total Variation Regularization

arXiv 2102.02414 ↗First posted 4 Feb 2021stat.ML cs.LGcited by 62

Not on the map. Why?

Abstract (arXiv)

Many weakly supervised classification methods employ a noise transition matrix to capture the class-conditional label corruption. To estimate the transition matrix from noisy data, existing methods often need to estimate the noisy class-posterior, which could be unreliable due to the overconfidence of neural networks. In this work, we propose a theoretically grounded method that can estimate the noise transition matrix and learn a classifier simultaneously, without relying on the error-prone noisy class-posterior estimation. Concretely, inspired by the characteristics of the stochastic label corruption process, we propose total variation regularization, which encourages the predicted probabilities to be more distinguishable from each other. Under mild assumptions, the proposed method yields a consistent estimator of the transition matrix. We show the effectiveness of the proposed method through experiments on benchmark and real-world datasets.

Key facts. Cited by 62 papers in the map (data through 7 Oct 2026). Builds on 42 of its 76 references that are in the map. First posted on arXiv 4 Feb 2021.

Builds on

42 of 76 references.

References · 42

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  4. 2017On Calibration of Modern Neural Networkscited by 4,860
  5. 2016Understanding deep learning requires rethinking generalizationcited by 1,657
  6. 2018Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labelscited by 1,053
  7. 2018Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labelscited by 994
  8. 2017A Closer Look at Memorization in Deep Networkscited by 936
  9. 2020DivideMix: Learning with Noisy Labels as Semi-supervised Learningcited by 691
  10. 2016Making Deep Neural Networks Robust to Label Noise: a Loss Correction Approachcited by 663
  11. 2017MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labelscited by 652
  12. 2019Symmetric Cross Entropy for Robust Learning with Noisy Labelscited by 499
30 more
  1. 2017Robust Loss Functions under Label Noise for Deep Neural Networkscited by 441
  2. 2019How does Disagreement Help Generalization against Label Corruption?cited by 390
  3. 2019Beyond temperature scaling: Obtaining well-calibrated multiclass probabilities with Dirichlet calibrationcited by 359
  4. 2018Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problemcited by 323
  5. 2020Combating noisy labels by agreement: A joint training method with co-regularizationcited by 289
  6. 2020Consistent Estimators for Learning to Defer to an Expertcited by 262
  7. 2020Normalized Loss Functions for Deep Learning with Noisy Labelscited by 257
  8. 2017Decoupling "when to update" from "how to update"cited by 230
  9. 2019Are Anchor Points Really Indispensable in Label-Noise Learning?cited by 208
  10. 2019SELF: Learning to Filter Noisy Labels with Self-Ensemblingcited by 164
  11. 2018Learning to Learn from Noisy Labeled Datacited by 160
  12. 2019Peer Loss Functions: Learning from Noisy Labels without Knowing Noise Ratescited by 142
  13. 2020Dual T: Reducing Estimation Error for Transition Matrix in Label-noise Learningcited by 142
  14. 2015Learning with Symmetric Label Noise: The Importance of Being Unhingedcited by 120
  15. 2018Iterative Learning with Open-set Noisy Labelscited by 108
  16. 2017Learning with Biased Complementary Labelscited by 93
  17. 2018Masking: A New Perspective of Noisy Supervisioncited by 93
  18. 2017Learning from Complementary Labelscited by 88
  19. 2019Curriculum Loss: Robust Learning and Generalization against Label Corruptioncited by 87
  20. 2019Combating Label Noise in Deep Learning Using Abstentioncited by 86
  21. 2020A Topological Filter for Learning with Label Noisecited by 76
  22. 2013Classification with Asymmetric Label Noise: Consistency and Maximal Denoisingcited by 67
  23. 2016Mixture Proportion Estimation via Kernel Embedding of Distributionscited by 67
  24. 2020Intra Order-preserving Functions for Calibration of Multi-Class Neural Networkscited by 65
  25. 2020Confidence Scores Make Instance-dependent Label-noise Learning Possiblecited by 63
  26. 2017Learning with Bounded Instance- and Label-dependent Label Noisecited by 57
  27. 2020Classification with Rejection Based on Cost-sensitive Classificationcited by 55
  28. 2019On Symmetric Losses for Learning from Corrupted Labelscited by 53
  29. 2018On the Minimal Supervision for Training Any Binary Classifier from Only Unlabeled Datacited by 43
  30. 2020Robustness of Accuracy Metric and its Inspirations in Learning with Noisy Labelscited by 20

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 6; 2022: 12; 2023: 14; 2024: 10; 2025: 13; 2026: 7.

62 papers.

Sort:

By area:

  1. 2023f-Divergence Minimization for Sequence-Level Knowledge Distillationcited by 112
  2. 2022Robust Training under Label Noise by Over-parameterizationcited by 93
  3. 2021Sample Selection with Uncertainty of Losses for Learning with Noisy Labelscited by 79
  4. 2022Identifiability of Label Noise Transition Matrixcited by 34
  5. 2023Understanding and Mitigating the Label Noise in Pre-training on Downstream Taskscited by 27
  6. 2022Few-shot Learning with Noisy Labelscited by 23
  7. 2021Understanding Instance-Level Label Noise: Disparate Impacts and Treatmentscited by 22
  8. 2022From Noisy Prediction to True Label: Noisy Prediction Calibration via Generative Modelcited by 21
  9. 2021Learning with Noisy Labels by Efficient Transition Matrix Estimation to Combat Label Miscorrectioncited by 20
  10. 2022Towards Robust Adaptive Object Detection under Noisy Annotationscited by 20
  11. 2022SimT: Handling Open-set Noise for Domain Adaptive Semantic Segmentationcited by 17
  12. 2022Instance-Dependent Noisy Label Learning via Graphical Modellingcited by 17
38 more
  1. 2024Label-Noise Robust Diffusion Modelscited by 16
  2. 2023Fine-Grained Classification with Noisy Labelscited by 15
  3. 2021Being Properly Impropercited by 11
  4. 2021Label Distributionally Robust Losses for Multi-class Classification: Consistency, Robustness and Adaptivitycited by 11
  5. 2024Dissecting Sample Hardness: A Fine-Grained Analysis of Hardness Characterization Methods for Data-Centric AIcited by 11
  6. 2025Hide and Seek in Noise Labels: Noise-Robust Collaborative Active Learning with LLM-Powered Assistancecited by 11
  7. 2023DyGen: Learning from Noisy Labels via Dynamics-Enhanced Generative Modelingcited by 10
  8. 2023Imprecise Label Learning: A Unified Framework for Learning with Various Imprecise Label Configurationscited by 9
  9. 2024Slight Corruption in Pre-training Data Makes Better Diffusion Modelscited by 9
  10. 2023Regroup Median Loss for Combating Label Noisecited by 8
  11. 2024Impact of Noisy Supervision in Foundation Model Learningcited by 8
  12. 2022Optical Remote Sensing Image Understanding with Weak Supervision: Concepts, Methods, and Perspectivescited by 6
  13. 2022Label Noise-Robust Learning using a Confidence-Based Sieving Strategycited by 6
  14. 2022ProSelfLC: Progressive Self Label Correction Towards A Low-Temperature Entropy Statecited by 5
  15. 2024De-Confusing Pseudo-Labels in Source-Free Domain Adaptationcited by 5
  16. 2025SelectMix: Enhancing Label Noise Robustness through Targeted Sample Mixingcited by 5
  17. 2023Instance-dependent Noisy-label Learning with Graphical Model Based Noise-rate Estimationcited by 4
  18. 2024Learning under Temporal Label Noisecited by 4
  19. 2024Dirichlet-based Per-Sample Weighting by Transition Matrix for Noisy Label Learningcited by 4
  20. 2025Conformal Prediction of Classifiers with Many Classes based on Noisy Labelscited by 4
  21. 2025Calibrating Pre-trained Language Classifiers on LLM-generated Noisy Labels via Iterative Refinementcited by 4
  22. 2021Elucidating Robust Learning with Uncertainty-Aware Corruption Pattern Estimationcited by 3
  23. 2024Learning Causal Transition Matrix for Instance-dependent Label Noisecited by 3
  24. 2025RCC-PFL: Robust Client Clustering under Noisy Labels in Personalized Federated Learningcited by 3
  25. 2025Guiding Noisy Label Conditional Diffusion Models with Score-based Discriminator Correctioncited by 3
  26. 2025Rethinking Consistent Multi-Label Classification Under Inexact Supervisioncited by 3
  27. 2022Tackling Instance-Dependent Label Noise with Dynamic Distribution Calibrationcited by 2
  28. 2023Binary Classification with Instance and Label Dependent Label Noisecited by 2
  29. 2023LNL+K: Enhancing Learning with Noisy Labels Through Noise Source Knowledge Integrationcited by 2
  30. 2025Noise-Aware Generalization: Robustness to In-Domain Noise and Out-of-Domain Generalizationcited by 2
  31. 2025Robust Classification with Noisy Labels Based on Posterior Maximizationcited by 2
  32. 2025Robust Learning of Diffusion Models with Extremely Noisy Conditionscited by 2
  33. 2026FB-NLL: A Feature-Based Approach to Tackle Noisy Labels in Personalized Federated Learningcited by 2
  34. 2023Towards the Identifiability in Noisy Label Learning: A Multinomial Mixture Modelling Approachcited by 1
  35. 2023PASS: Peer-Agreement based Sample Selection for training with Noisy Labelscited by 1
  36. 2023Learning under Label Proportions for Text Classificationcited by 1
  37. 2025Correcting Noisy Multilabel Predictions: Modeling Label Noise through Latent Space Shiftscited by 1
  38. 2025Weakly-Supervised Contrastive Learning for Imprecise Class Labelscited by 1

50 most cited of 62 All 62 (JSON)