mlsanityData to 14 Oct 2026

Attacking Adversarial Defences by Smoothing the Loss Landscape

arXiv 2208.00862 ↗First posted 1 Aug 2022cs.LG cs.CR cs.CVcited by 2

Not on the map. Why?

Abstract (arXiv)

This paper investigates a family of methods for defending against adversarial attacks that owe part of their success to creating a noisy, discontinuous, or otherwise rugged loss landscape that adversaries find difficult to navigate. A common, but not universal, way to achieve this effect is via the use of stochastic neural networks. We show that this is a form of gradient obfuscation, and propose a general extension to gradient-based adversaries based on the Weierstrass transform, which smooths the surface of the loss function and provides more reliable gradient estimates. We further show that the same principle can strengthen gradient-free adversaries. We demonstrate the efficacy of our loss-smoothing method against both stochastic and non-stochastic adversarial defences that exhibit robustness due to this type of obfuscation. Furthermore, we provide analysis of how it interacts with Expectation over Transformation; a popular gradient-sampling method currently used to attack stochastic defences.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 33 of its 46 references that are in the map. First posted on arXiv 1 Aug 2022.

Builds on

33 of 46 references.

References · 33

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  3. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  4. 2013Intriguing properties of neural networkscited by 5,155
  5. 2016Wide Residual Networkscited by 3,070
  6. 2016Adversarial examples in the physical worldcited by 1,769
  7. 2020Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attackscited by 1,451
  8. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,268
  9. 2016Deep Variational Information Bottleneckcited by 930
  10. 2016Practical Black-Box Attacks against Machine Learningcited by 903
  11. 2017ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Modelscited by 666
  12. 2020RobustBench: a standardized adversarial robustness benchmarkcited by 661
21 more
  1. 2020On Adaptive Attacks to Adversarial Example Defensescited by 555
  2. 2016Transferability in Machine Learning: from Phenomena to Black-Box Attacks using Adversarial Samplescited by 446
  3. 2019Unlabeled Data Improves Adversarial Robustnesscited by 416
  4. 2019Using Pre-Training Can Improve Model Robustness and Uncertaintycited by 373
  5. 2019Nesterov Accelerated Gradient and Scale Invariance for Adversarial Attackscited by 355
  6. 2020Uncovering the Limits of Adversarial Training against Norm-Bounded Adversarial Examplescited by 278
  7. 2021Enhancing the Transferability of Adversarial Attacks through Variance Tuningcited by 244
  8. 2021Fixing Data Augmentation to Improve Adversarial Robustnesscited by 204
  9. 2020Geometry-aware Instance-reweighted Adversarial Trainingcited by 198
  10. 2017Formal Guarantees on the Robustness of a Classifier against Adversarial Manipulationcited by 185
  11. 2017Towards Robust Neural Networks via Random Self-ensemblecited by 159
  12. 2018Parametric Noise Injection: Trainable Randomness to Improve Deep Neural Network Robustness against Adversarial Attackcited by 125
  13. 2018Adversarial Attacks and Defences Competitioncited by 96
  14. 2018Adv-BNN: Improved Adversarial Defense through Robust Bayesian Neural Networkcited by 69
  15. 2019Enhancing Adversarial Defense by k-Winners-Take-Allcited by 51
  16. 2021GradDiv: Adversarial Robustness of Randomized Neural Networks via Gradient Diversity Regularizationcited by 34
  17. 2020Learn2Perturb: an End-to-end Feature Perturbation Learning to Improve Adversarial Robustnesscited by 27
  18. 2021Combating Adversaries with Anti-Adversariescited by 27
  19. 2021Indicators of Attack Failure: Debugging and Improving Optimization of Adversarial Examplescited by 23
  20. 2021Improving Neural Network Robustness via Persistency of Excitationcited by 14
  21. 2020Weight-Covariance Alignment for Adversarially Robust Neural Networkscited by 9

Built on by

Cited by year

Citing papers per year: 2023: 1; 2024: 1.

2 papers.

  1. 2024Feature compression is the root cause of adversarial fragility in neural network classifierscited by 1
  2. 2023Physics Inspired Approaches To Understanding Gaussian Processescited by 0