CVAICLLGMay 8, 2025

X-Transfer Attacks: Towards Super Transferable Adversarial Attacks on CLIP

arXiv:2505.05528v318 citationsh-index: 12Has CodeICML
Originality Highly original
AI Analysis

This addresses a critical security concern for CLIP models used in diverse downstream tasks and vision-language models, representing a strong specific gain rather than an incremental improvement.

The paper tackles the vulnerability of CLIP models to adversarial attacks by introducing X-Transfer, a method that generates a universal adversarial perturbation with super transferability across data, domains, models, and tasks, significantly outperforming previous state-of-the-art methods.

As Contrastive Language-Image Pre-training (CLIP) models are increasingly adopted for diverse downstream tasks and integrated into large vision-language models (VLMs), their susceptibility to adversarial perturbations has emerged as a critical concern. In this work, we introduce \textbf{X-Transfer}, a novel attack method that exposes a universal adversarial vulnerability in CLIP. X-Transfer generates a Universal Adversarial Perturbation (UAP) capable of deceiving various CLIP encoders and downstream VLMs across different samples, tasks, and domains. We refer to this property as \textbf{super transferability}--a single perturbation achieving cross-data, cross-domain, cross-model, and cross-task adversarial transferability simultaneously. This is achieved through \textbf{surrogate scaling}, a key innovation of our approach. Unlike existing methods that rely on fixed surrogate models, which are computationally intensive to scale, X-Transfer employs an efficient surrogate scaling strategy that dynamically selects a small subset of suitable surrogates from a large search space. Extensive evaluations demonstrate that X-Transfer significantly outperforms previous state-of-the-art UAP methods, establishing a new benchmark for adversarial transferability across CLIP models. The code is publicly available in our \href{https://github.com/HanxunH/XTransferBench}{GitHub repository}.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes