Saved in:
Bibliographic Details
Main Authors: Liu, Yongxiang, Peng, Bowen, Liu, Li, Li, Xiang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.13891
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912999563853824
author Liu, Yongxiang
Peng, Bowen
Liu, Li
Li, Xiang
author_facet Liu, Yongxiang
Peng, Bowen
Liu, Li
Li, Xiang
contents Transferable Targeted Attacks (TTAs) face significant challenges due to severe overfitting to surrogate models. Recent breakthroughs heavily rely on large-scale training data of victim models, while data-free solutions, \textit{i.e.}, image transformation-involved gradient optimization, often depend on black-box feedback for method design and tuning. These dependencies violate black-box transfer settings and compromise threat evaluation fairness. In this paper, we propose two blind estimation measures, self-alignment and self-transferability, to analyze per-transformation effectiveness and cross-transformation correlations under strict black-box constraints. Our findings challenge conventional assumptions: (1) Attacking simple scaling transformations uniquely enhances targeted transferability, outperforming other basic transformations and rivaling leading complex methods; (2) Geometric and color transformations exhibit high internal redundancy despite weak inter-category correlations. These insights drive the design and tuning of S$^4$ST (Strong, Self-transferable, faSt, Simple Scale Transformation), which integrates dimensionally consistent scaling, complementary low-redundancy transformations, and block-wise operations. Extensive evaluations across diverse architectures, training distributions, and tasks show that S$^{4}$ST achieves state-of-the-art effectiveness-efficiency balance without data dependency. We reveal that scaling's effectiveness stems from visual data's multi-scale nature and ubiquitous scale augmentation during training, rendering such augmentation a double-edged sword. Further validations on medical imaging and face verification confirm the framework's strong generalization.
format Preprint
id arxiv_https___arxiv_org_abs_2410_13891
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle S$^4$ST: A Strong, Self-transferable, faSt, and Simple Scale Transformation for Transferable Targeted Attack
Liu, Yongxiang
Peng, Bowen
Liu, Li
Li, Xiang
Cryptography and Security
Artificial Intelligence
Transferable Targeted Attacks (TTAs) face significant challenges due to severe overfitting to surrogate models. Recent breakthroughs heavily rely on large-scale training data of victim models, while data-free solutions, \textit{i.e.}, image transformation-involved gradient optimization, often depend on black-box feedback for method design and tuning. These dependencies violate black-box transfer settings and compromise threat evaluation fairness. In this paper, we propose two blind estimation measures, self-alignment and self-transferability, to analyze per-transformation effectiveness and cross-transformation correlations under strict black-box constraints. Our findings challenge conventional assumptions: (1) Attacking simple scaling transformations uniquely enhances targeted transferability, outperforming other basic transformations and rivaling leading complex methods; (2) Geometric and color transformations exhibit high internal redundancy despite weak inter-category correlations. These insights drive the design and tuning of S$^4$ST (Strong, Self-transferable, faSt, Simple Scale Transformation), which integrates dimensionally consistent scaling, complementary low-redundancy transformations, and block-wise operations. Extensive evaluations across diverse architectures, training distributions, and tasks show that S$^{4}$ST achieves state-of-the-art effectiveness-efficiency balance without data dependency. We reveal that scaling's effectiveness stems from visual data's multi-scale nature and ubiquitous scale augmentation during training, rendering such augmentation a double-edged sword. Further validations on medical imaging and face verification confirm the framework's strong generalization.
title S$^4$ST: A Strong, Self-transferable, faSt, and Simple Scale Transformation for Transferable Targeted Attack
topic Cryptography and Security
Artificial Intelligence
url https://arxiv.org/abs/2410.13891