mlsanityData to 7 Oct 2026

Indiscriminate Poisoning Attacks on Unsupervised Contrastive Learning

arXiv 2202.11202 ↗First posted 22 Feb 2022cs.LG cs.AI cs.CRcited by 32

On the map: LLM safety and security → Backdoor and data poisoning attacks

Abstract (arXiv)

Indiscriminate data poisoning attacks are quite effective against supervised learning. However, not much is known about their impact on unsupervised contrastive learning (CL). This paper is the first to consider indiscriminate poisoning attacks of contrastive learning. We propose Contrastive Poisoning (CP), the first effective such attack on CL. We empirically show that Contrastive Poisoning, not only drastically reduces the performance of CL algorithms, but also attacks supervised learning models, making it the most generalizable indiscriminate poisoning attack. We also show that CL algorithms with a momentum encoder are more robust to indiscriminate poisoning, and propose a new countermeasure based on matrix completion. Code is available at: https://github.com/kaiwenzha/contrastive-poisoning.

Key facts. Cited by 32 papers in the map (data through 7 Oct 2026). Placed in Backdoor and data poisoning attacks, one of 12 directions in LLM safety and security. Builds on 44 of its 59 references that are in the map. First posted on arXiv 22 Feb 2022.

Builds on

44 of 59 references.

Closest ancestors · 1

Same direction, newest first

  1. 2021BadEncoder: Backdoor Attacks to Pre-trained Encoders in Self-Supervised Learning1 Aug

Other references · 43

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  3. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  4. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  5. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  6. 2018Representation Learning with Contrastive Predictive Codingcited by 7,662
  7. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,482
  8. 2016Densely Connected Convolutional Networkscited by 6,691
  9. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  10. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  11. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
  12. 2018MobileNetV2: Inverted Residuals and Linear Bottleneckscited by 4,773
31 more
  1. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,712
  2. 2019CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Featurescited by 2,875
  3. 2020Improved Baselines with Momentum Contrastive Learningcited by 2,282
  4. 2020Extracting Training Data from Large Language Modelscited by 2,144
  5. 2017Improved Regularization of Convolutional Neural Networks with Cutoutcited by 1,623
  6. 2021An Empirical Study of Training Self-Supervised Vision Transformerscited by 1,610
  7. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,510
  8. 2020Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hyperspherecited by 1,393
  9. 2020Big Self-Supervised Models are Strong Semi-Supervised Learnerscited by 1,375
  10. 2016Understanding intermediate layers using linear classifier probescited by 1,279
  11. 2020A Survey on Contrastive Self-supervised Learningcited by 553
  12. 2017Wild Patterns: Ten Years After the Rise of Adversarial Machine Learningcited by 411
  13. 2021Big Self-Supervised Models Advance Medical Image Classificationcited by 320
  14. 2020Large image datasets: A pyrrhic win for computer vision?cited by 237
  15. 2017Towards Poisoning of Deep Learning Algorithms with Back-gradient Optimizationcited by 188
  16. 2021Unlearnable Examples: Making Personal Data Unexploitablecited by 188
  17. 2021Poisoning and Backdooring Contrastive Learningcited by 149
  18. 2020Robust Pre-Training by Adversarial Contrastive Learningcited by 144
  19. 2020Adversarial Self-Supervised Contrastive Learningcited by 138
  20. 2021Adversarial Examples Make Strong Poisonscited by 115
  21. 2021LowKey: Leveraging Adversarial Attacks to Protect Social Media Users from Facial Recognitioncited by 92
  22. 2021Backdoor Attacks on Self-Supervised Learningcited by 89
  23. 2021Divide and Contrast: Self-supervised Learning from Uncurated Datacited by 85
  24. 2020On the Effectiveness of Mitigating Data Poisoning Attacks with Gradient Shapingcited by 82
  25. 2021When Does Contrastive Learning Preserve Adversarial Robustness from Pretraining to Finetuning?cited by 76
  26. 2019Learning to Confuse: Generating Training Time Adversarial Data with Auto-Encodercited by 70
  27. 2019ME-Net: Towards Effective Adversarial Robustness with Matrix Estimationcited by 65
  28. 2021Availability Attacks Create Shortcutscited by 60
  29. 2021Better Safe Than Sorry: Preventing Delusive Adversaries with Adversarial Trainingcited by 58
  30. 2021DP-InstaHide: Provably Defusing Poisoning and Backdoor Attacks with Differentially Private Data Augmentationscited by 33
  31. 2020Systematic Evaluation of Backdoor Data Poisoning Attacks on Image Classifierscited by 23

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 2; 2023: 6; 2024: 14; 2025: 5; 2026: 5.

32 papers.

Sort:

By area:

  1. 2022Transferable Unlearnable Examplescited by 41
  2. 2023Image Shortcut Squeezing: Countering Perturbative Availability Poisons with Compressioncited by 39
  3. 2022Unlearnable Clusters: Towards Label-agnostic Unlearnable Examplescited by 35
  4. 2023Stable Unlearnable Example: Enhancing the Robustness of Unlearnable Examples via Stable Error-Minimizing Noisecited by 23
  5. 2023Unlearnable Examples Give a False Sense of Security: Piercing through Unexploitable Data with Learnable Examplescited by 17
  6. 2023Detection and Defense of Unlearnable Examplescited by 13
  7. 2024Semantic Deep Hiding for Robust Unlearnable Examplescited by 13
  8. 2024UnSeg: One Universal Unlearnable Example Generator is Enough against All Image Segmentationcited by 13
  9. 2024Efficient Availability Attacks against Supervised and Contrastive Learning Simultaneouslycited by 12
  10. 2025MTL-UE: Learning to Learn Nothing for Multi-Task Learningcited by 11
  11. 2024Indiscriminate Data Poisoning Attacks on Pre-trained Feature Extractorscited by 8
  12. 2024Medical Unlearnable Examples: Securing Medical Data from Unauthorized Training via Sparsity-Aware Local Maskingcited by 8
20 more
  1. 2023APBench: A Unified Benchmark for Availability Poisoning Attacks and Defensescited by 7
  2. 2024Toward Availability Attacks in 3D Point Cloudscited by 7
  3. 2024DeDe: Detecting Backdoor Samples for SSL Encoders via Decoderscited by 6
  4. 2023Transferable Availability Poisoning Attackscited by 5
  5. 2024Nonlinear Transformations Against Unlearnable Datasetscited by 5
  6. 2026When Priors Backfire: On the Vulnerability of Unlearnable Examples to Pretrainingcited by 5
  7. 2024BridgePure: Limited Protection Leakage Can Break Black-Box Data Protectioncited by 3
  8. 2025T2UE: Generating Unlearnable Examples from Text Descriptionscited by 3
  9. 2025A Linear Approach to Data Poisoningcited by 2
  10. 2025Provable Watermarking for Data Poisoning Attackscited by 2
  11. 2026Why Do Unlearnable Examples Work: A Novel Perspective of Mutual Informationcited by 2
  12. 2024Exploiting Supervised Poison Vulnerability to Strengthen Self-Supervised Defensecited by 1
  13. 2024REPEAT: Improving Uncertainty Estimation in Representation Learning Explainabilitycited by 1
  14. 2026Safety-Efficacy Trade Off: Robustness against Data-Poisoningcited by 1
  15. 2024Hard Work Does Not Always Pay Off: Poisoning Attacks on Neural Architecture Searchcited by 0
  16. 2024Empirical Perturbation Analysis of Linear System Solvers from a Data Poisoning Perspectivecited by 0
  17. 2024Deferred Poisoning: Making the Model More Vulnerable via Hessian Singularizationcited by 0
  18. 2025Learning to Generate Cross-Task Unexploitable Examplescited by 0
  19. 2026SoK: A Comprehensive Analysis of the Current Status of Neural Tangent Generalization Attacks with Research Directionscited by 0
  20. 2026Leveraging Imperfect Restoration for Data Availability Attackcited by 0