mlsanityData to 14 Oct 2026

How does Early Stopping Help Generalization against Label Noise?

arXiv 1911.08059 ↗First posted 19 Nov 2019cs.LG stat.MLcited by 41

Not on the map. Why?

Abstract (arXiv)

Noisy labels are very common in real-world training data, which lead to poor generalization on test data because of overfitting to the noisy labels. In this paper, we claim that such overfitting can be avoided by "early stopping" training a deep neural network before the noisy labels are severely memorized. Then, we resume training the early stopped network using a "maximal safe set," which maintains a collection of almost certainly true-labeled samples at each epoch since the early stop point. Putting them all together, our novel two-phase training method, called Prestopping, realizes noise-free training under any type of label noise for practical use. Extensive experiments using four image benchmark data sets verify that our method significantly outperforms four state-of-the-art methods in test error by 0.4-8.2 percent points under existence of real-world noise.

Key facts. Cited by 41 papers in the map (data through 14 Oct 2026). First posted on arXiv 19 Nov 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 13; 2022: 6; 2023: 7; 2024: 7; 2025: 5; 2026: 3.

41 papers.

Sort:

By area:

  1. 2021Deep Learning with Label Differential Privacycited by 117
  2. 2022Robust Training under Label Noise by Over-parameterizationcited by 93
  3. 2021FINE Samples for Learning with Noisy Labelscited by 82
  4. 2021Understanding Instance-Level Label Noise: Disparate Impacts and Treatmentscited by 22
  5. 2025Early Stopping Against Label Noise Without Validation Datacited by 21
  6. 2021Learning with Noisy Labels by Efficient Transition Matrix Estimation to Combat Label Miscorrectioncited by 20
  7. 2022Instance-Dependent Noisy Label Learning via Graphical Modellingcited by 17
  8. 2021Disparity Between Batches as a Signal for Early Stoppingcited by 9
  9. 2021Sample Prior Guided Robust Model Learning to Suppress Noisy Labelscited by 8
  10. 2023Improving Medical Image Classification in Noisy Labels Using Only Self-supervised Pretrainingcited by 8
  11. 2024Tackling Noisy Labels with Network Parameter Additive Decompositioncited by 8
  12. 2021Unsupervised Embedding Adaptation via Early-Stage Feature Reconstruction for Few-Shot Classificationcited by 7
29 more
  1. 2022Label Noise-Robust Learning using a Confidence-Based Sieving Strategycited by 6
  2. 2023Lower Generalization Bounds for GD and SGD in Smooth Stochastic Convex Optimizationcited by 6
  3. 2021Confidence Adaptive Regularization for Deep Learning with Noisy Labelscited by 5
  4. 2022A Survey of Dataset Refinement for Problems in Computer Vision Datasetscited by 5
  5. 2024PLReMix: Combating Noisy Labels with Pseudo-Label Relaxed Contrastive Representation Learningcited by 5
  6. 2021Learning to Combat Noisy Labels via Classification Marginscited by 4
  7. 2022A Gift from Label Smoothing: Robust Training with Adaptive Label Smoothing via Auxiliary Classifier under Label Noisecited by 4
  8. 2023Making Binary Classification from Multiple Unlabeled Datasets Almost Free of Supervisioncited by 4
  9. 2025Enhancing Sample Selection Against Label Noise by Cutting Mislabeled Easy Examplescited by 4
  10. 2021How Does Heterogeneous Label Noise Impact Generalization in Neural Nets?cited by 3
  11. 2025Is the Information Bottleneck Robust Enough? Towards Label-Noise Resistant Information Bottleneck Learningcited by 3
  12. 2026A Human-in-the-Loop Label Error Detection Framework Applied to Arabic-Script HTR Datasetscited by 3
  13. 2022CNTN: Cyclic Noise-tolerant Network for Gait Recognitioncited by 2
  14. 2023Pushing the Accuracy-Group Robustness Frontier with Introspective Self-playcited by 2
  15. 2023Training on Synthetic Data Beats Real Data in Multimodal Relation Extractioncited by 2
  16. 2024Learning from Different Samples: A Source-free Framework for Semi-supervised Domain Adaptationcited by 2
  17. 2021Double Descent Optimization Pattern and Aliasing: Caveats of Noisy Labelscited by 1
  18. 2024How does self-supervised pretraining improve robustness against noisy labels across various medical image classification datasets?cited by 1
  19. 2024Paint Outside the Box: Synthesizing and Selecting Training Data for Visual Groundingcited by 1
  20. 2026Symmetrization of Loss Functions for Robust Training of Neural Networks in the Presence of Noisy Labelscited by 1
  21. 2021Generation and Analysis of Feature-Dependent Pseudo Noise for Training Deep Neural Networkscited by 0
  22. 2021Influential Rank: A New Perspective of Post-training for Robust Model against Noisy Labelscited by 0
  23. 2023A Curriculum View of Robust Loss Functionscited by 0
  24. 2023Omnipotent Adversarial Training in the Wildcited by 0
  25. 2024Prioritizing Informative Features and Examples for Deep Learning from Noisy Datacited by 0
  26. 2024ConsistentFeature: A Plug-and-Play Component for Neural Network Regularizationcited by 0
  27. 2025Source-free Semantic Regularization Learning for Semi-supervised Domain Adaptationcited by 0
  28. 2025FedEFC: Federated Learning Using Enhanced Forward Correction Against Noisy Labelscited by 0
  29. 2026ACD-U: Asymmetric co-teaching with machine unlearning for robust learning with noisy labelscited by 0