Awesome Mixup Methods for Supervised Learning

We summarize mixup methods proposed for supervised visual representation learning from two aspects: sample mixup policy and label mixup policy. We are working a survey of mixup methods. Current list is on updating.

Sample Mixup Methods

Pre-defined Policies

1. MixUp, [ICLR 2018] [code] mixup: Beyond Empirical Risk Minimization.

2. AdaMixup, [AAAI 2019] MixUp as Locally Linear Out-Of-Manifold Regularization.

3. CutMix, [ICCV 2019] [code] CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features.

4. ManifoldMix, [ICML 2019] [code] Manifold Mixup: Better Representations by Interpolating Hidden States.

5. FMix, [Arixv 2020] [code] FMix: Enhancing Mixed Sample Data Augmentation.

6. SmoothMix, [CVPRW 2020] [code] SmoothMix: a Simple Yet Effective Data Augmentation to Train Robust Classifiers.

7. PatchUp, [Arxiv 2020] [code] PatchUp: A Regularization Technique for Convolutional Neural Networks.

8. GridMixup, [Pattern Recognition 2021] [code] GridMix: Strong regularization through local context mapping.

9. SmoothMix, [CVPRW 2020] SmoothMix: A Simple Yet Effective Data Augmentation to Train Robust Classifiers.

10. ResizeMix, [Arixv 2020] [code] ResizeMix: Mixing Data with Preserved Object Information and True Labels.

11. FocusMix, [ICTC 2020] Where to Cut and Paste: Data Regularization with Selective Features.

12. AugMix, [ICLR 2020] [code] AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty.

13. DJMix, [Arxiv 2021] DJMix: Unsupervised Task-agnostic Augmentation for Improving Robustness.

14. PixMix, [Arxiv 2021] [code] PixMix: Dreamlike Pictures Comprehensively Improve Safety Measures.

15. StyleMix, [CVPR 2021] [code] StyleMix: Separating Content and Style for Enhanced Data Augmentation.

16. MixStyle, [ICLR 2021] [code] Domain Generalization with MixStyle.

17. MoEx, [CVPR 2021] [code] On Feature Normalization and Data Augmentation.

18. LocalMix, [AISTATS 2021] Preventing Manifold Intrusion with Locality: Local Mixup.

Saliency-guided Policies

1. SaliencyMix, [ICLR 2021] [code] SaliencyMix: A Saliency Guided Data Augmentation Strategy for Better Regularization.

2. AttentiveMix, [ICASSP 2020] [code] Attentive CutMix: An Enhanced Data Augmentation Approach for Deep Learning Based Image Classification.

3. SnapMix, [AAAI 2021] [code] SnapMix: Semantically Proportional Mixing for Augmenting Fine-grained Data.

4. AttributeMix, [Arxiv 2020] Attribute Mix: Semantic Data Augmentation for Fine Grained Recognition.

5. PuzzleMix, [ICML 2020] [code] Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal Mixup.

6. CoMixup, [ICLR 2021] [code] Co-Mixup: Saliency Guided Joint Mixup with Supermodular Diversity.

7. SuperMix, [CVPR 2021] [code] SuperMix: Supervising the Mixing Data Augmentation.

8. PatchMix, [Arxiv 2021] Evolving Image Compositions for Feature Representation Learning.

9. AutoMix, [Arxiv 2021] [code] Unveiling the Power of Mixup for Stronger Classifiers.

10. AlignMix, [Arxiv 2021] [code] AlignMix: Improving representation by interpolating aligned features.

11. SAMix, [Arxiv 2021] [code] Boosting Discriminative Visual Representation Learning with Scenario-Agnostic Mixup.

12. AutoMix, [ECCV 2020] AutoMix: Mixup Networks for Sample Interpolation via Cooperative Barycenter Learning.

13. StackMix, [Arxiv 2021] StackMix: A complementary Mix algorithm.

14. ScoreMix, [Arxiv 2022] ScoreNet: Learning Non-Uniform Attention and Augmentation for Transformer-Based Histopathological Image Classification.

15. RecursiveMix, [Arxiv 2022] [code] RecursiveMix: Mixed Learning with History.

Label Mixup Methods

1. MixUp, [ICLR 2018] [code] mixup: Beyond Empirical Risk Minimization.

2. CutMix, [ICCV 2019] [code] CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features.

3. MetaMixup, [TNNLS 2021] Metamixup: Learning adaptive interpolation policy of mixup with metalearning.

4. mWH, [Arxiv 2021] [code] Mixup Without Hesitation.

5. CAMixup, [ICLR 2021] [code] Combining Ensembles and Data Augmentation can Harm your Calibration.

6. Saliency Grafting, [AAAI 2022] Saliency Grafting: Innocuous Attribution-Guided Mixup with Calibrated Label Mixing.

7. TransMix, [CVPR 2022] [code] TransMix: Attend to Mix for Vision Transformers.

8. DecoupleMix, [Arxiv 2022,] [code] Decoupled Mixup for Data-efficient Learning.

Contribution

Feel free to send pull requests to add more links! Current contributors include: Siyuan Li (@Lupin1998) and Zicheng Liu (@pone7).