mlsanityData to 7 Oct 2026

Invariance-powered Trustworthy Defense via Remove Then Restore

arXiv 2402.00304 ↗First posted 1 Feb 2024cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Adversarial attacks pose a challenge to the deployment of deep neural networks (DNNs), while previous defense models overlook the generalization to various attacks. Inspired by targeted therapies for cancer, we view adversarial samples as local lesions of natural benign samples, because a key finding is that salient attack in an adversarial sample dominates the attacking process, while trivial attack unexpectedly provides trustworthy evidence for obtaining generalizable robustness. Based on this finding, a Pixel Surgery and Semantic Regeneration (PSSR) model following the targeted therapy mechanism is developed, which has three merits: 1) To remove the salient attack, a score-based Pixel Surgery module is proposed, which retains the trivial attack as a kind of invariance information. 2) To restore the discriminative content, a Semantic Regeneration module based on a conditional alignment extrapolator is proposed, which achieves pixel and semantic consistency. 3) To further harmonize robustness and accuracy, an intractable problem, a self-augmentation regularizer with adversarial R-drop is designed. Experiments on numerous benchmarks show the superiority of PSSR.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 22 of its 55 references that are in the map. First posted on arXiv 1 Feb 2024.

Builds on

22 of 55 references.

References · 22

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  2. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  3. 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,348
  4. 2014Conditional Generative Adversarial Netscited by 2,152
  5. 2016Adversarial Machine Learning at Scalecited by 934
  6. 2017Decision-Based Adversarial Attacks: Reliable Attacks Against Black-Box Machine Learning Modelscited by 459
  7. 2017Mitigating Adversarial Effects Through Randomizationcited by 382
  8. 2018Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Modelscited by 362
  9. 2018Certified Defenses against Adversarial Examplescited by 305
  10. 2017PixelDefend: Leveraging Generative Models to Understand and Defend against Adversarial Examplescited by 243
  11. 2020Bag of Tricks for Adversarial Trainingcited by 199
  12. 2019Are Labels Required for Improving Adversarial Robustness?cited by 162
10 more
  1. 2014Towards Deep Neural Network Architectures Robust to Adversarial Examplescited by 146
  2. 2019advertorch v0.1: An Adversarial Robustness Toolbox based on PyTorchcited by 69
  3. 2021Online Adversarial Purification based on Self-Supervisioncited by 45
  4. 2020Towards Frequency-Based Explanation for Robust CNNcited by 29
  5. 2021A Frequency Perspective of Adversarial Robustnesscited by 28
  6. 2017Fraternal Dropoutcited by 14
  7. 2017Training Ensembles to Detect Adversarial Examplescited by 10
  8. 2022LPF-Defense: 3D Adversarial Defense based on Frequency Analysiscited by 10
  9. 2016Dropout with Expectation-linear Regularizationcited by 9
  10. 2022What Do Adversarially trained Neural Networks Focus: A Fourier Domain-based Studycited by 3

Built on by

Not cited yet.