Cutting through buggy adversarial example defenses: fixing 1 line of code breaks Sabre
Not on the map. Why?
Abstract (arXiv)
Sabre is a defense to adversarial examples that was accepted at IEEE S&P 2024. We first reveal significant flaws in the evaluation that point to clear signs of gradient masking. We then show the cause of this gradient masking: a bug in the original evaluation code. By fixing a single line of code in the original repository, we reduce Sabre's robust accuracy to 0%. In response to this, the authors modify the defense and introduce a new defense component not described in the original paper. But this fix contains a second bug; modifying one more line of code reduces robust accuracy to below baseline levels. After we released the first version of our paper online, the authors introduced another change to the defense; by commenting out one line of code during attack we reduce the robust accuracy to 0% again.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 8 of its 9 references that are in the map. First posted on arXiv 6 May 2024.
Builds on
8 of 9 references.
References · 8
- 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
- 2020Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attackscited by 1,451
- 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,268
- 2020On Adaptive Attacks to Adversarial Example Defensescited by 555
- 2017Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networkscited by 486
- 2019On Evaluating Adversarial Robustnesscited by 447
- 2019ME-Net: Towards Effective Adversarial Robustness with Matrix Estimationcited by 65
- 2023A LLM Assisted Exploitation of AI-Guardiancited by 15
Built on by
Citing papers per year: 2024: 2.
2 papers.