mlsanityData to 14 Oct 2026

Enhancing Sample Selection Against Label Noise by Cutting Mislabeled Easy Examples

arXiv 2502.08227 ↗First posted 12 Feb 2025cs.LGcited by 4

Not on the map. Why?

Abstract (arXiv)

Sample selection is a prevalent approach in learning with noisy labels, aiming to identify confident samples for training. Although existing sample selection methods have achieved decent results by reducing the noise rate of the selected subset, they often overlook that not all mislabeled examples harm the model's performance equally. In this paper, we demonstrate that mislabeled examples correctly predicted by the model early in the training process are particularly harmful to model performance. We refer to these examples as Mislabeled Easy Examples (MEEs). To address this, we propose Early Cutting, which introduces a recalibration step that employs the model's later training state to re-select the confident subset identified early in training, thereby avoiding misleading confidence from early learning and effectively filtering out MEEs. Experiments on the CIFAR, WebVision, and full ImageNet-1k datasets demonstrate that our method effectively improves sample selection and model performance by reducing MEEs.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 49 of its 74 references that are in the map. First posted on arXiv 12 Feb 2025.

Builds on

49 of 74 references.

References · 49

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  3. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
  4. 2017Understanding Black-box Predictions via Influence Functionscited by 1,741
  5. 2020Sharpness-Aware Minimization for Efficiently Improving Generalizationcited by 1,417
  6. 2019MixMatch: A Holistic Approach to Semi-Supervised Learningcited by 1,391
  7. 2018Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labelscited by 1,059
  8. 2018Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labelscited by 997
  9. 2017A Closer Look at Memorization in Deep Networkscited by 942
  10. 2020DivideMix: Learning with Noisy Labels as Semi-supervised Learningcited by 692
  11. 2018An Empirical Study of Example Forgetting during Deep Neural Network Learningcited by 678
  12. 2020What Neural Networks Memorize and Why: Discovering the Long Tail via Influence Estimationcited by 497
37 more
  1. 2019Does Learning Require Memorization? A Short Tale about a Long Tailcited by 460
  2. 2020Early-Learning Regularization Prevents Memorization of Noisy Labelscited by 416
  3. 2019How does Disagreement Help Generalization against Label Corruption?cited by 390
  4. 2020Combating noisy labels by agreement: A joint training method with co-regularizationcited by 289
  5. 2017WebVision Database: Visual Learning and Understanding from Web Datacited by 262
  6. 2021Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotationscited by 256
  7. 2022TinyViT: Fast Pretraining Distillation for Small Vision Transformerscited by 234
  8. 2018Sensitivity and Generalization in Neural Networks: an Empirical Studycited by 189
  9. 2019SELF: Learning to Filter Noisy Labels with Self-Ensemblingcited by 164
  10. 2020Learning with Instance-Dependent Label Noise: A Sample Sieve Approachcited by 158
  11. 2020A Survey of Label-noise Representation Learning: Past, Present and Futurecited by 122
  12. 2021Understanding and Improving Early Stopping for Learning with Noisy Labelscited by 113
  13. 2020Estimating Example Difficulty Using Variance of Gradientscited by 97
  14. 2020Characterizing Structural Regularities of Labeled Data in Overparameterized Modelscited by 93
  15. 2021Sample Selection with Uncertainty of Losses for Learning with Noisy Labelscited by 79
  16. 2013Classification with Asymmetric Label Noise: Consistency and Maximal Denoisingcited by 68
  17. 2021Open-set Label Noise Can Improve Robustness Against Inherent Label Noisecited by 59
  18. 2021Co-learning: Learning from Noisy Labels with Self-supervisioncited by 51
  19. 2019How does Early Stopping Help Generalization against Label Noise?cited by 41
  20. 2022Characterizing Datapoints via Second-Split Forgettingcited by 37
  21. 2022Self-Filtering: A Noise-Aware Sample Selection for Label Noise with Confidence Penalizationcited by 33
  22. 2023Machine Vision Therapy: Multimodal Large Language Models Can Enhance Visual Robustness via Denoising In-Context Learningcited by 27
  23. 2024Why is SAM Robust to Label Noise?cited by 21
  24. 2025Early Stopping Against Label Noise Without Validation Datacited by 21
  25. 2023On the Over-Memorization During Natural, Robust and Catastrophic Overfittingcited by 20
  26. 2022Identifying Hard Noise in Long-Tailed Sample Distributioncited by 19
  27. 2020Robust Learning by Self-Transition for Handling Noisy Labelscited by 18
  28. 2022Label-Noise Learning with Intrinsically Long-Tailed Datacited by 17
  29. 2023Late Stopping: Avoiding Confidently Learning from Mislabeled Examplescited by 17
  30. 2023BadLabel: A Robust Perspective on Evaluating and Enhancing Label-noise Learningcited by 16
  31. 2023InstanT: Semi-supervised Learning with Instance-dependent Thresholdscited by 13
  32. 2023Transferring Annotator- and Instance-dependent Transition Matrix for Learning from Crowdscited by 11
  33. 2025Instance-dependent Early Stoppingcited by 10
  34. 2023Regularly Truncated M-estimators for Learning with Noisy Labelscited by 8
  35. 2023Regroup Median Loss for Combating Label Noisecited by 8
  36. 2024Tackling Noisy Labels with Network Parameter Additive Decompositioncited by 8
  37. 2022ProSelfLC: Progressive Self Label Correction Towards A Low-Temperature Entropy Statecited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 3.

4 papers.

  1. 2025The Role of Diversity in In-Context Learning for Large Language Modelscited by 6
  2. 2026Mitigating Mismatch within Reference-based Preference Optimizationcited by 1
  3. 2026MeGU: Machine-Guided Unlearning with Target Feature Disentanglementcited by 0
  4. 2026Radial-Angular Geometry for Reliable Update Diagnosis in Noisy-Label Learningcited by 0