mlsanityData to 14 Oct 2026

Robust Deep Learning Models Against Semantic-Preserving Adversarial Attack

arXiv 2304.03955 ↗First posted 8 Apr 2023cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Deep learning models can be fooled by small lp-norm adversarial perturbations and natural perturbations in terms of attributes. Although the robustness against each perturbation has been explored, it remains a challenge to address the robustness against joint perturbations effectively. In this paper, we study the robustness of deep learning models against joint perturbations by proposing a novel attack mechanism named Semantic-Preserving Adversarial (SPA) attack, which can then be used to enhance adversarial training. Specifically, we introduce an attribute manipulator to generate natural and human-comprehensible perturbations and a noise generator to generate diverse adversarial noises. Based on such combined noises, we optimize both the attribute value and the diversity variable to generate jointly-perturbed samples. For robust training, we adversarially train the deep learning model against the generated joint perturbations. Empirical results on four benchmarks show that the SPA attack causes a larger performance decline with small l∞ norm-ball constraints compared to existing approaches. Furthermore, our SPA-enhanced training outperforms existing defense methods against such joint perturbations.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 27 of its 34 references that are in the map. First posted on arXiv 8 Apr 2023.

Builds on

27 of 34 references.

References · 27

  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. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,349
  6. 2014Deep Learning Face Attributes in the Wildcited by 3,693
  7. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,998
  8. 2014Sequence to Sequence Learning with Neural Networkscited by 2,986
  9. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,877
  10. 2014Conditional Generative Adversarial Netscited by 2,162
  11. 2016Adversarial examples in the physical worldcited by 1,769
  12. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,517
15 more
  1. 2015Spatial Transformer Networkscited by 1,429
  2. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,268
  3. 2016Adversarial Machine Learning at Scalecited by 936
  4. 2020Fast is better than free: Revisiting adversarial trainingcited by 727
  5. 2020On Adaptive Attacks to Adversarial Example Defensescited by 555
  6. 2020Bag of Tricks for Adversarial Trainingcited by 200
  7. 2019Diversity-Sensitive Conditional Generative Adversarial Networkscited by 77
  8. 2020Understanding Catastrophic Overfitting in Single-step Adversarial Trainingcited by 76
  9. 2021Going Deeper through the Gleason Scoring Scale: An Automatic end-to-end System for Histology Prostate Grading and Cribriform Pattern Detectioncited by 68
  10. 2019Semantic Adversarial Attacks: Parametric Transformations That Fool Deep Classifierscited by 57
  11. 2017NAG: Network for Adversary Generationcited by 51
  12. 2019Are Adversarial Robustness and Common Perturbation Robustness Independent Attributes ?cited by 25
  13. 2020Attribute-Guided Adversarial Training for Robustness to Natural Perturbationscited by 23
  14. 2019A Direct Approach to Robust Deep Learning Using Adversarial Networkscited by 18
  15. 2019Latent Space Factorisation and Manipulation via Matrix Subspace Projectioncited by 9

Built on by

Not cited yet.