mlsanityData to 14 Oct 2026

How to Craft Backdoors with Unlabeled Data Alone?

arXiv 2404.06694 ↗First posted 10 Apr 2024cs.LG cs.AI cs.CRcited by 1

Not on the map. Why?

Abstract (arXiv)

Relying only on unlabeled data, Self-supervised learning (SSL) can learn rich features in an economical and scalable way. As the drive-horse for building foundation models, SSL has received a lot of attention recently with wide applications, which also raises security concerns where backdoor attack is a major type of threat: if the released dataset is maliciously poisoned, backdoored SSL models can behave badly when triggers are injected to test samples. The goal of this work is to investigate this potential risk. We notice that existing backdoors all require a considerable amount of labeled data that may not be available for SSL. To circumvent this limitation, we explore a more restrictive setting called no-label backdoors, where we only have access to the unlabeled data alone, where the key challenge is how to select the proper poison set without using label information. We propose two strategies for poison selection: clustering-based selection using pseudolabels, and contrastive selection derived from the mutual information principle. Experiments on CIFAR-10 and ImageNet-100 show that both no-label backdoors are effective on many SSL methods and outperform random poisoning by a large margin. Code will be available at https://github.com/PKU-ML/nlb.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 24 of its 34 references that are in the map. First posted on arXiv 10 Apr 2024.

Builds on

24 of 34 references.

References · 24

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  3. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  4. 2020Language Models are Few-Shot Learnerscited by 32,419
  5. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  6. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  7. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,513
  8. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  9. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,750
  10. 2016Pointer Sentinel Mixture Modelscited by 3,096
  11. 2020Improved Baselines with Momentum Contrastive Learningcited by 2,287
  12. 2021Barlow Twins: Self-Supervised Learning via Redundancy Reductioncited by 1,838
12 more
  1. 2021An Empirical Study of Training Self-Supervised Vision Transformerscited by 1,616
  2. 2018Learning deep representations by mutual information estimation and maximizationcited by 1,273
  3. 2017BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chaincited by 1,106
  4. 2017Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoningcited by 1,055
  5. 2017Revisiting Unreasonable Effectiveness of Data in Deep Learning Eracited by 919
  6. 2019On Variational Bounds of Mutual Informationcited by 559
  7. 2021Poisoning and Backdooring Contrastive Learningcited by 151
  8. 2021BadEncoder: Backdoor Attacks to Pre-trained Encoders in Self-Supervised Learningcited by 123
  9. 2021Backdoor Attacks on Self-Supervised Learningcited by 90
  10. 2022Fine-Tuning Is All You Need to Mitigate Backdoor Attackscited by 38
  11. 2022An Embarrassingly Simple Backdoor Attack on Self-supervised Learningcited by 31
  12. 2023Defending Against Patch-based Backdoor Attacks on Self-Supervised Learningcited by 20

Built on by

Cited by year

Citing papers per year: 2024: 1.

1 paper.

  1. 2024Perturb and Recover: Fine-tuning for Effective Backdoor Removal from CLIPcited by 4

Similar, not cited

Close in wording, no citation either way

  1. 2023Erasing Self-Supervised Learning Backdoor by Cluster Activation Masking
  2. 2022The Perils of Learning From Unlabeled Data: Backdoor Attacks on Semi-supervised Learning
  3. 2023Towards Understanding How Self-training Tolerates Data Backdoor Poisoning
  4. 2024Invisible Backdoor Attack against Self-supervised Learning
  5. 2022Rethinking Backdoor Data Poisoning Attacks in the Context of Semi-Supervised Learning
  6. 2023Trojaning semi-supervised learning model via poisoning wild images on the web
  7. 2022Training set cleansing of backdoor poisoning by self-supervised representation learning
  8. 2024Defending Against Repetitive Backdoor Attacks on Semi-supervised Learning through Lens of Rate-Distortion-Perception Trade-off