Saved in:
Bibliographic Details
Main Author: Fort, Stanislav
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.14496
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929685803302912
author Fort, Stanislav
author_facet Fort, Stanislav
contents This note documents an implementation issue in recent adaptive attacks (Zhang et al. [2024]) against the multi-resolution self-ensemble defense (Fort and Lakshminarayanan [2024]). The implementation allowed adversarial perturbations to exceed the standard $L_\infty = 8/255$ bound by up to a factor of 20$\times$, reaching magnitudes of up to $L_\infty = 160/255$. When attacks are properly constrained within the intended bounds, the defense maintains non-trivial robustness. Beyond highlighting the importance of careful validation in adversarial machine learning research, our analysis reveals an intriguing finding: properly bounded adaptive attacks against strong multi-resolution self-ensembles often align with human perception, suggesting the need to reconsider how we measure adversarial robustness.
format Preprint
id arxiv_https___arxiv_org_abs_2501_14496
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Note on Implementation Errors in Recent Adaptive Attacks Against Multi-Resolution Self-Ensembles
Fort, Stanislav
Cryptography and Security
Computer Vision and Pattern Recognition
Machine Learning
This note documents an implementation issue in recent adaptive attacks (Zhang et al. [2024]) against the multi-resolution self-ensemble defense (Fort and Lakshminarayanan [2024]). The implementation allowed adversarial perturbations to exceed the standard $L_\infty = 8/255$ bound by up to a factor of 20$\times$, reaching magnitudes of up to $L_\infty = 160/255$. When attacks are properly constrained within the intended bounds, the defense maintains non-trivial robustness. Beyond highlighting the importance of careful validation in adversarial machine learning research, our analysis reveals an intriguing finding: properly bounded adaptive attacks against strong multi-resolution self-ensembles often align with human perception, suggesting the need to reconsider how we measure adversarial robustness.
title A Note on Implementation Errors in Recent Adaptive Attacks Against Multi-Resolution Self-Ensembles
topic Cryptography and Security
Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2501.14496