mlsanityData to 7 Oct 2026

Rethinking Rotation in Self-Supervised Contrastive Learning: Adaptive Positive or Negative Data Augmentation

arXiv 2210.12681 ↗First posted 23 Oct 2022cs.CVcited by 3

Not on the map. Why?

Abstract (arXiv)

Rotation is frequently listed as a candidate for data augmentation in contrastive learning but seldom provides satisfactory improvements. We argue that this is because the rotated image is always treated as either positive or negative. The semantics of an image can be rotation-invariant or rotation-variant, so whether the rotated image is treated as positive or negative should be determined based on the content of the image. Therefore, we propose a novel augmentation strategy, adaptive Positive or Negative Data Augmentation (PNDA), in which an original and its rotated image are a positive pair if they are semantically close and a negative pair if they are semantically different. To achieve PNDA, we first determine whether rotation is positive or negative on an image-by-image basis in an unsupervised way. Then, we apply PNDA to contrastive learning frameworks. Our experiments showed that PNDA improves the performance of contrastive learning. The code is available at url https://github.com/AtsuMiyai/rethinkingrotation.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 36 of its 40 references that are in the map. First posted on arXiv 23 Oct 2022.

Builds on

36 of 40 references.

References · 36

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  3. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,482
  4. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,339
  5. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
  6. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  7. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
  8. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,712
  9. 2019CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Featurescited by 2,875
  10. 2020Exploring Simple Siamese Representation Learningcited by 2,853
  11. 2020Unsupervised Learning of Visual Features by Contrasting Cluster Assignmentscited by 2,845
  12. 2020Supervised Contrastive Learningcited by 2,803
24 more
  1. 2020Improved Baselines with Momentum Contrastive Learningcited by 2,282
  2. 2018Unsupervised Representation Learning by Predicting Image Rotationscited by 1,586
  3. 2017Accurate, Large Minibatch SGD: Training ImageNet in 1 Hourcited by 1,479
  4. 2016Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzlescited by 1,379
  5. 2016Group Equivariant Convolutional Networkscited by 1,154
  6. 2020What Makes for Good Views for Contrastive Learning?cited by 843
  7. 2017Large Batch Training of Convolutional Networkscited by 529
  8. 2020CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instancescited by 441
  9. 2020Noise or Signal: The Role of Image Backgrounds in Object Recognitioncited by 342
  10. 2021With a Little Help from My Friends: Nearest-Neighbor Contrastive Learning of Visual Representationscited by 324
  11. 2020Whitening for Self-Supervised Representation Learningcited by 278
  12. 2020Universal Domain Adaptation through Self Supervisioncited by 172
  13. 2020Boosting Contrastive Self-Supervised Learning with False Negative Cancellationcited by 119
  14. 2021ReSSL: Relational Self-Supervised Learning with Weak Augmentationcited by 83
  15. 2021Weakly Supervised Contrastive Learningcited by 81
  16. 2021Equivariant Contrastive Learningcited by 65
  17. 2021Understanding Negative Samples in Instance Discriminative Self-supervised Representation Learningcited by 47
  18. 2021Negative Data Augmentationcited by 46
  19. 2021Robust Contrastive Learning Using Negative Samples with Diminished Semanticscited by 43
  20. 2021Solving Inefficiency of Self-supervised Representation Learningcited by 41
  21. 2021Residual Relaxation for Multi-view Representation Learningcited by 30
  22. 2016Warped Convolutions: Efficient Invariance to Spatial Transformationscited by 24
  23. 2021Novelty Detection via Contrastive Learning with Negative Data Augmentationcited by 10
  24. 2022Towards Efficient and Effective Self-Supervised Learning of Visual Representationscited by 6

Built on by

Cited by year

Citing papers per year: 2022: 1; 2023: 1; 2024: 1.

3 papers.

  1. 2024Guidelines for Augmentation Selection in Contrastive Learning for Time Series Classificationcited by 5
  2. 2023Contrastive Learning of View-Invariant Representations for Facial Expressions Recognitioncited by 2
  3. 2022Ladder Siamese Network: a Method and Insights for Multi-level Self-Supervised Learningcited by 1