Boosting Adversarial Transferability via Commonality-Oriented Gradient Optimization

doi:10.1007/978-981-95-5693-9_5

2026-01-01

PRE

DOI：10.1007/978-981-95-5693-9_5

原文链接

原文求助

摘要

En 中文

探索有效且可迁移的对抗样本对于理解视觉变换器（ViTs）的特性与机制至关重要。然而，源自替代模型的对抗样本在黑盒设置中通常表现出较差的迁移性，这是由于过拟合所致。现有方法通过多样化扰动输入或应用替代模型内的统一梯度正则化来提升迁移性，但它们未能充分利用在相同任务上训练的替代模型所共有的共享特征和独特特征，导致迁移性能次优。因此，增强替代模型共享信息的扰动并抑制与个体特性相关的扰动，是提升迁移性的有效途径。为此，我们提出了一种面向共性的梯度优化策略（COGO），包括两个组成部分：共性增强（CE）和个体性抑制（IS）。CE扰动中低频区域，利用了在相同数据集上训练的ViTs倾向于更多依赖中低频信息进行分类的事实。IS采用自适应阈值来评估反向传播梯度与模型个体性之间的相关性，并据此为梯度分配权重。大量实验表明，COGO显著提升了对抗攻击的迁移成功率，超越了当前最先进的方法。

Keyword:

Adversarial Example

Vision Transformer

期刊

PATTERN RECOGNITION AND COMPUTER VISION, PRCV 2025, PT II

IF:

论文数:

被引数:

机构

State University System of Florida

学者数:

12.8W

论文数: 10.9W

被引数: 130

tongji university

学者数:

7.9W

论文数: 6.0W

被引数: 98

引用论文

Segment Anything Model Is a Good Teacher for Local Feature Learning Segment anything model 是局部特征学习的一个好老师。

IEEE Transactions on Image Processing

IF0

2025-01-01

PREAI

Wu,Jingqian; Xu,Rongtao; Wood-Doughty,Zach; Wang,Changwei; Xu,Shibiao; Lam,Edmund Y.

Rethinking the Inception Architecture for Computer Vision 重新思考计算机视觉的Inception体系结构

2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

IF0

2016-06-01

OAAI

Christian Szegedy; Vincent Vanhoucke; Sergey Ioffe; Jon Shlens; Zbigniew Wojna

Rethinking Spatial Dimensions of Vision Transformers 视觉变形金刚空间维度的再思考

2021 IEEE/CVF International Conference on Computer Vision (ICCV)

IF0

2021-10-01

OAAI

Byeongho Heo; Sangdoo Yun; Dongyoon Han; Sanghyuk Chun; Junsuk Choe; Seong Joon Oh

Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning Inception-v4、Inception-ResNet和剩余连接对学习的影响

Proceedings of the AAAI Conference on Artificial Intelligence

IF0

2017-02-12

OAAI

Christian Szegedy; Sergey Ioffe; Vincent Vanhoucke; Alexander Alemi

Boosting Adversarial Attacks with Momentum 用动量增强对抗性攻击

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

IF0

2018-06-01

OAAI

Yinpeng Dong; Fangzhou Liao; Tianyu Pang; Hang Su; Jun Zhu; Xiaolin Hu; Jianguo Li

ImageNet Large Scale Visual Recognition Challenge ImageNet大规模视觉识别挑战

INTERNATIONAL JOURNAL OF COMPUTER VISION

IF9.3

2015-04-11

2.7W

PREAI

Russakovsky, Olga; Deng, Jia; Su, Hao; Krause, Jonathan; Satheesh, Sanjeev; Ma, Sean; Huang, Zhiheng; Karpathy, Andrej; Khosla, Aditya; Bernstein, Michael; Berg, Alexander C.; Fei-Fei, Li

ConViT: improving vision transformers with soft convolutional inductive biases*

Journal of Statistical Mechanics: Theory and Experiment

IF0

2022-11-24

OAAI

Stéphane d’Ascoli; Hugo Touvron; Matthew L Leavitt; Ari S Morcos; Giulio Biroli; Levent Sagun

Going deeper with Image Transformers 更深入地了解图像变形金刚

2021 IEEE/CVF International Conference on Computer Vision (ICCV)

IF0

2021-10-01

OAAI

Hugo Touvron; Matthieu Cord; Alexandre Sablayrolles; Gabriel Synnaeve; Herve Jegou