One Surrogate to Fool Them All: Universal, Transferable, and Targeted Adversarial Attacks with CLIP
Open Access
- 19 November 2025
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 3087-3101
Abstract
No abstract availableAll Related Versions
This publication has 25 references indexed in Scilit:
- Robust Overfitting Does Matter: Test-Time Adversarial Purification with FGSM2024
- ImageBind One Embedding Space to Bind Them All2023
- Enhancing the Self-Universality for Transferable Targeted Attacks2023
- Revisiting Residual Networks for Adversarial Robustness2023
- Medical image fusion method by deep learningInternational Journal of Cognitive Computing in Engineering, 2021
- Regional Homogeneity: Towards Learning Transferable Universal Adversarial Perturbations Against Defenses2020
- Boosting Decision-Based Black-Box Adversarial Attacks with Random Sign Flip2020
- Adversarial Robustness: From Self-Supervised Pre-Training to Fine-Tuning2020
- Feature Space Perturbations Yield More Transferable Adversarial Examples2019
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015