mlsanityData to 14 Oct 2026

Measuring the False Sense of Security

arXiv 2204.04778 ↗First posted 10 Apr 2022cs.LG cs.CRcited by 0

Not on the map. Why?

Abstract (arXiv)

Recently, several papers have demonstrated how widespread gradient masking is amongst proposed adversarial defenses. Defenses that rely on this phenomenon are considered failed, and can easily be broken. Despite this, there has been little investigation into ways of measuring the phenomenon of gradient masking and enabling comparisons of its extent amongst different networks. In this work, we investigate gradient masking under the lens of its mensurability, departing from the idea that it is a binary phenomenon. We propose and motivate several metrics for it, performing extensive empirical tests on defenses suspected of exhibiting different degrees of gradient masking. These are computationally cheaper than strong attacks, enable comparisons between models, and do not require the large time investment of tailor-made attacks for specific models. Our results reveal metrics that are successful in measuring the extent of gradient masking across different networks

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 20 of its 25 references that are in the map. First posted on arXiv 10 Apr 2022.

Builds on

20 of 25 references.

References · 20

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  2. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  3. 2020Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attackscited by 1,451
  4. 2015DeepFool: a simple and accurate method to fool deep neural networkscited by 1,318
  5. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,268
  6. 2016Practical Black-Box Attacks against Machine Learningcited by 903
  7. 2017Ensemble Adversarial Training: Attacks and Defensescited by 840
  8. 2020Fast is better than free: Revisiting adversarial trainingcited by 727
  9. 2020RobustBench: a standardized adversarial robustness benchmarkcited by 661
  10. 2020On Adaptive Attacks to Adversarial Example Defensescited by 555
  11. 2019Unlabeled Data Improves Adversarial Robustnesscited by 416
  12. 2019Using Pre-Training Can Improve Model Robustness and Uncertaintycited by 373
8 more
  1. 2020Uncovering the Limits of Adversarial Training against Norm-Bounded Adversarial Examplescited by 278
  2. 2018Adversarial Risk and the Dangers of Evaluating Against Weak Attackscited by 221
  3. 2020Geometry-aware Instance-reweighted Adversarial Trainingcited by 198
  4. 2018Robustness via curvature regularization, and vice versacited by 175
  5. 2019Robust Learning with Jacobian Regularizationcited by 129
  6. 2019Defense Against Adversarial Attacks Using Feature Scattering-based Adversarial Trainingcited by 95
  7. 2018Improving DNN Robustness to Adversarial Attacks using Jacobian Regularizationcited by 88
  8. 2019Adversarial Training is a Form of Data-dependent Operator Norm Regularizationcited by 4

Built on by

Not cited yet.