mlsanityData to 7 Oct 2026

The Effects of Mixed Sample Data Augmentation are Class Dependent

arXiv 2307.09136 ↗First posted 18 Jul 2023cs.CV cs.AIcited by 1

Not on the map. Why?

Abstract (arXiv)

Mixed Sample Data Augmentation (MSDA) techniques, such as Mixup, CutMix, and PuzzleMix, have been widely acknowledged for enhancing performance in a variety of tasks. A previous study reported the class dependency of traditional data augmentation (DA), where certain classes benefit disproportionately compared to others. This paper reveals a class dependent effect of MSDA, where some classes experience improved performance while others experience degraded performance. This research addresses the issue of class dependency in MSDA and proposes an algorithm to mitigate it. The approach involves training on a mixture of MSDA and non-MSDA data, which not only mitigates the negative impact on the affected classes, but also improves overall accuracy. Furthermore, we provide in-depth analysis and discussion of why MSDA introduced class dependencies and which classes are most likely to have them.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 48 of its 70 references that are in the map. First posted on arXiv 18 Jul 2023.

Builds on

48 of 70 references.

References · 48

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  3. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  4. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,845
  5. 2020Training data-efficient image transformers & distillation through attentioncited by 4,560
  6. 2016Wide Residual Networkscited by 3,057
  7. 2019CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Featurescited by 2,875
  8. 2019RandAugment: Practical automated data augmentation with a reduced search spacecited by 2,261
  9. 2016Identity Mappings in Deep Residual Networkscited by 1,960
  10. 2019A Survey on Bias and Fairness in Machine Learningcited by 1,911
  11. 2013Fine-Grained Visual Classification of Aircraftcited by 1,701
  12. 2020Simple Copy-Paste is a Strong Data Augmentation Method for Instance Segmentationcited by 595
36 more
  1. 2021MViTv2: Improved Multiscale Vision Transformers for Classification and Detectioncited by 516
  2. 2020Contrastive learning of global and local features for medical image segmentation with limited annotationscited by 309
  3. 2019On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networkscited by 304
  4. 2020Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal Mixupcited by 260
  5. 2020How Does Mixup Help With Robustness and Generalization?cited by 193
  6. 2021Benchmarking Representation Learning for Natural World Image Collectionscited by 167
  7. 2021A survey on datasets for fairness-aware machine learningcited by 157
  8. 2020SaliencyMix: A Saliency Guided Data Augmentation Strategy for Better Regularizationcited by 149
  9. 2018Classification with Fairness Constraints: A Meta-Algorithm with Provable Guaranteescited by 146
  10. 2021Co-Mixup: Saliency Guided Joint Mixup with Supermodular Diversitycited by 127
  11. 2019Augmenting Data with Mixup for Sentence Classification: An Empirical Studycited by 125
  12. 2022Bias Mitigation for Machine Learning Classifiers: A Comprehensive Surveycited by 125
  13. 2018Data Augmentation by Pairing Samples for Images Classificationcited by 96
  14. 2021A Survey on Bias in Visual Datasetscited by 84
  15. 2021Context Decoupling Augmentation for Weakly Supervised Semantic Segmentationcited by 69
  16. 2021AutoMix: Unveiling the Power of Mixup for Stronger Classifierscited by 69
  17. 2020ResizeMix: Mixing Data with Preserved Object Information and True Labelscited by 68
  18. 2022The Effects of Regularization and Data Augmentation are Class Dependentcited by 67
  19. 2022Regional Semantic Contrast and Aggregation for Weakly Supervised Semantic Segmentationcited by 64
  20. 2021When and How Mixup Improves Calibrationcited by 59
  21. 2020SuperMix: Supervising the Mixing Data Augmentationcited by 58
  22. 2020SnapMix: Semantically Proportional Mixing for Augmenting Fine-grained Datacited by 56
  23. 2020VideoMix: Rethinking Data Augmentation for Video Classificationcited by 55
  24. 2020Combining Ensembles and Data Augmentation can Harm your Calibrationcited by 44
  25. 2020Structured Consistency Loss for semi-supervised semantic segmentationcited by 39
  26. 2021Analyzing Overfitting under Class Imbalance in Neural Networks for Image Segmentationcited by 34
  27. 2021AlignMixup: Improving Representations By Interpolating Aligned Featurescited by 33
  28. 2020Local Additivity Based Data Augmentation for Semi-supervised NERcited by 31
  29. 2018Proxy Fairnesscited by 29
  30. 2019Overfitting of neural nets under class imbalance: Analysis and improvements for segmentationcited by 27
  31. 2019Improving Robustness of Deep Learning Based Knee MRI Segmentation: Mixup and Adversarial Domain Adaptationcited by 27
  32. 2022A Unified Analysis of Mixed Sample Data Augmentation: A Loss Function Perspectivecited by 27
  33. 2021Towards Understanding the Data Dependency of Mixup-style Trainingcited by 19
  34. 2019Semi-supervised semantic segmentation needs strong, varied perturbationscited by 14
  35. 2020Learning Temporally Invariant and Localizable Features via Data Augmentation for Video Recognitioncited by 13
  36. 2021Weakly-guided Self-supervised Pretraining for Temporal Activity Detectioncited by 5

Built on by

Cited by year

Citing papers per year: 2024: 1.

1 paper.

  1. 2024Class-wise Generalization Error: an Information-Theoretic Analysiscited by 2