mlsanityData to 7 Oct 2026

Preventing Manifold Intrusion with Locality: Local Mixup

arXiv 2201.04368 ↗First posted 12 Jan 2022cs.LGcited by 11

Not on the map. Why?

Abstract (arXiv)

Mixup is a data-dependent regularization technique that consists in linearly interpolating input samples and associated outputs. It has been shown to improve accuracy when used to train on standard machine learning datasets. However, authors have pointed out that Mixup can produce out-of-distribution virtual samples and even contradictions in the augmented training set, potentially resulting in adversarial effects. In this paper, we introduce Local Mixup in which distant input samples are weighted down when computing the loss. In constrained settings we demonstrate that Local Mixup can create a trade-off between bias and variance, with the extreme cases reducing to vanilla training and classical Mixup. Using standardized computer vision benchmarks , we also show that Local Mixup can improve test accuracy.

Key facts. Cited by 11 papers in the map (data through 7 Oct 2026). Builds on 20 of its 36 references that are in the map. First posted on arXiv 12 Jan 2022.

Builds on

20 of 36 references.

References · 20

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2016Densely Connected Convolutional Networkscited by 6,691
  3. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  4. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,283
  5. 2019CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Featurescited by 2,875
  6. 2017Improved Regularization of Convolutional Neural Networks with Cutoutcited by 1,623
  7. 2016Colorful Image Colorizationcited by 1,113
  8. 2019AugMix: A Simple Data Processing Method to Improve Robustness and Uncertaintycited by 897
  9. 2018AutoAugment: Learning Augmentation Policies from Datacited by 728
  10. 2019A Baseline for Few-Shot Image Classificationcited by 298
  11. 2020Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal Mixupcited by 260
  12. 2018MixUp as Locally Linear Out-Of-Manifold Regularizationcited by 158
8 more
  1. 2020Remix: Rebalanced Mixupcited by 144
  2. 2019Charting the Right Manifold: Manifold Mixup for Few-shot Learningcited by 135
  3. 2020On Mixup Regularizationcited by 82
  4. 2020Variational Autoencoders with Riemannian Brownian Motion Priorscited by 34
  5. 2021MixMo: Mixing Multiple Inputs for Multiple Outputs via Deep Subnetworkscited by 32
  6. 2016CIFAR-10: KNN-based Ensemble of Classifierscited by 16
  7. 2021k-Mixup Regularization for Deep Learning via Optimal Transportcited by 14
  8. 2020Enhancing Mixup-based Semi-Supervised Learning with Explicit Lipschitz Regularizationcited by 3

Built on by

Cited by year

Citing papers per year: 2022: 3; 2023: 4; 2024: 3; 2025: 1.

11 papers.

  1. 2022C-Mixup: Improving Generalization in Regressioncited by 57
  2. 2022A Survey of Mix-based Data Augmentation: Taxonomy, Methods, Applications, and Explainabilitycited by 30
  3. 2024A Survey on Mixup Augmentations and Beyondcited by 18
  4. 2023IPMix: Label-Preserving Data Augmentation Method for Training Robust Classifierscited by 14
  5. 2023Anchor Data Augmentationcited by 5
  6. 2023Augment on Manifold: Mixup Regularization with UMAPcited by 3
  7. 2022SoftEdge: Regularizing Graph Classification with Random Soft Edgescited by 2
  8. 2024SaVe-TAG: LLM-based Interpolation for Long-Tailed Text-Attributed Graphscited by 1
  9. 2024LH-Mix: Local Hierarchy Correlation Guided Mixup over Hierarchical Prompt Tuningcited by 1
  10. 2025A Generalized Theory of Mixup for Structure-Preserving Synthetic Datacited by 1
  11. 2023ID-MixGCL: Identity Mixup for Graph Contrastive Learningcited by 0