Data to 7 Oct 2026

Online Adversarial Purification based on Self-Supervision

arXiv 2101.09387 ↗First posted 23 Jan 2021cs.LG cs.CR cs.CVcited by 45

Not on the map. Why?

Abstract (arXiv)

Deep neural networks are known to be vulnerable to adversarial examples, where a perturbation in the input space leads to an amplified shift in the latent network representation. In this paper, we combine canonical supervised learning with self-supervised representation learning, and present Self-supervised Online Adversarial Purification (SOAP), a novel defense strategy that uses a self-supervised loss to purify adversarial examples at test-time. Our approach leverages the label-independent nature of self-supervised signals and counters the adversarial perturbation with respect to the self-supervised tasks. SOAP yields competitive robust accuracy against state-of-the-art adversarial training and purification methods, with considerably less training complexity. In addition, our approach is robust even when adversaries are given knowledge of the purification defense strategy. To the best of our knowledge, our paper is the first that generalizes the idea of using self-supervised signals to perform online test-time purification.

Key facts. Cited by 45 papers in the map (data through 7 Oct 2026). Builds on 32 of its 36 references that are in the map. First posted on arXiv 23 Jan 2021.

Builds on

32 of 36 references.

References · 32

2015Deep Residual Learning for Image Recognitioncited by 51,809
2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,482
2014Explaining and Harnessing Adversarial Examplescited by 7,287
2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
2016Wide Residual Networkscited by 3,057
2016Towards Evaluating the Robustness of Neural Networkscited by 2,976
2018Unsupervised Representation Learning by Predicting Image Rotationscited by 1,586
2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,510
2016Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzlescited by 1,379
2015DeepFool: a simple and accurate method to fool deep neural networkscited by 1,315
2016Colorful Image Colorizationcited by 1,113

20 more

2015Unsupervised Visual Representation Learning by Context Predictioncited by 1,062
2016Temporal Ensembling for Semi-Supervised Learningcited by 1,054
2016Adversarial Machine Learning at Scalecited by 934
2016Practical Black-Box Attacks against Machine Learningcited by 893
2017Ensemble Adversarial Training: Attacks and Defensescited by 833
2019Using Self-Supervised Learning Can Improve Model Robustness and Uncertaintycited by 511
2016Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learningcited by 441
2018Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Modelscited by 362
2017Defense against Adversarial Attacks Using High-Level Representation Guided Denoisercited by 329
2014Discriminative Unsupervised Feature Learning with Exemplar Convolutional Neural Networkscited by 299
2017MagNet: a Two-Pronged Defense against Adversarial Examplescited by 268
2017PixelDefend: Leveraging Generative Models to Understand and Defend against Adversarial Examplescited by 243
2018Adversarial Logit Pairingcited by 224
2017Multi-task Self-Supervised Visual Learningcited by 200
2020A Self-supervised Approach for Adversarial Robustnesscited by 180
2018Disentangling Adversarial Robustness and Generalizationcited by 171
2020Adversarial Robustness: From Self-Supervised Pre-Training to Fine-Tuningcited by 152
2014Towards Deep Neural Network Architectures Robust to Adversarial Examplescited by 146
2019Metric Learning for Adversarial Robustnesscited by 77
2019PuVAE: A Variational Autoencoder to Purify Adversarial Examplescited by 28

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 9; 2022: 8; 2023: 14; 2024: 8; 2025: 4; 2026: 2.

45 papers.

Sort:

By area:

2022Diffusion Models for Adversarial Purificationcited by 469
2021Adversarial purification with Score-based generative modelscited by 132
2023Diffusion Models for Imperceptible and Transferable Adversarial Attackcited by 93
2022Guided Diffusion Model for Adversarial Purificationcited by 90
2023TrojDiff: Trojan Attacks on Diffusion Models with Diverse Targetscited by 88
2023How Deep Learning Sees the World: A Survey on Adversarial Attacks & Defensescited by 65
2022Evaluating the Adversarial Robustness of Adaptive Test-time Defensescited by 64
2021Adversarially Robust One-class Novelty Detectioncited by 24
2024ADBM: Adversarial diffusion bridge model for reliable adversarial purificationcited by 24
2023Defending against Adversarial Audio via Diffusion Modelcited by 20
2021Fighting Gradients with Gradients: Dynamic Defenses against Adversarial Attackscited by 19
2021COMPOSER: Compositional Reasoning of Group Activity in Videos with Keypoint-Only Modalitycited by 19

33 more

2023Enhancing Adversarial Robustness via Score-Based Optimizationcited by 19
2023MimicDiffusion: Purifying Adversarial Perturbation via Mimicking Clean Diffusion Modelcited by 18
2022Text Adversarial Purification as Defense against Adversarial Attackscited by 15
2023The Best Defense is a Good Offense: Adversarial Augmentation against Adversarial Attackscited by 15
2023Defenses in Adversarial Machine Learning: A Surveycited by 14
2021AID-Purifier: A Light Auxiliary Network for Boosting Adversarial Defensecited by 10
2022Robust Perception through Equivariancecited by 9
2024Demonstration of an Adversarial Attack Against a Multimodal Vision Language Model for Pathology Imagingcited by 9
2022Active Image Indexingcited by 8
2022Adversarial Purification with the Manifold Hypothesiscited by 8
2023Beyond Pretrained Features: Noisy Image Modeling Provides Adversarial Defensecited by 7
2024Improving Accuracy-robustness Trade-off via Pixel Reweighted Adversarial Trainingcited by 6
2025Trustworthy Machine Learning via Memorization and the Granular Long-Tail: A Survey on Interactions, Tradeoffs, and Beyondcited by 6
2023Robust Feature Inference: A Test-time Defense Strategy using Spectral Projectionscited by 3
2024Interpretability-Guided Test-Time Adversarial Defensecited by 3
2024Achieving Domain-Independent Certified Robustness via Knowledge Continuitycited by 3
2024BridgePure: Limited Protection Leakage Can Break Black-Box Data Protectioncited by 3
2025Adversarial-Guided Diffusion for Multimodal LLM Attackscited by 3
2022Exploiting the Potential of Datasets: A Data-Centric Approach for Model Robustnesscited by 2
2023Improving Adversarial Robustness of DEQs with Explicit Regulations Along the Neural Dynamicscited by 2
2024DiffPAD: Denoising Diffusion-based Adversarial Patch Decontaminationcited by 2
2023Language Guided Adversarial Purificationcited by 1
2025Adversarial Examples in Environment Perception for Automated Driving (Review)cited by 1
2025Rectifying Adversarial Sample with Low Entropy Prior for Test-Time Defensecited by 1
2026Beyond Uniform Sampling: Synergistic Active Learning and Input Denoising for Robust Neural Operatorscited by 1
2021NCIS: Neural Contextual Iterative Smoothing for Purifying Adversarial Perturbationscited by 0
2021Adversarial Purification through Representation Disentanglementcited by 0
2021Enhanced countering adversarial attacks via input denoising and feature restoringcited by 0
2021Improving Robustness with Image Filteringcited by 0
2023Blending adversarial training and representation-conditional purification via aggregation improves adversarial robustnesscited by 0
2023Advancing Adversarial Robustness Through Adversarial Logit Updatecited by 0
2024Invariance-powered Trustworthy Defense via Remove Then Restorecited by 0
2026PurSAMERE: Reliable Adversarial Purification via Sharpness-Aware Minimization of Expected Reconstruction Errorcited by 0