从变换到迁移：提升视觉语言预训练模型上的对抗攻击迁移性
Transform to Transfer: Boosting Adversarial Attack Transferability on Vision-Language Pre-training Models
可学习的变换组合机制（自适应选择最优变换以最大化输入多样性）可直接迁移到任何黑盒迁移攻击流程中，替代手工固定的变换堆叠；积分梯度用于缓解对源模型的过拟合，也是提升跨模型迁移性的通用手段。
原文依据 a learnable transformation mechanism that adaptively selects optimal combinations of transformations to maximize input diversity
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论视觉语言预训练（VLP）模型虽在多模态任务上取得最优性能，却对多模态对抗样本表现出显著脆弱性。在 VLP 模型的黑盒攻击场景中，关键挑战在于这些对抗样本的迁移性有限。现有提升迁移性的方法往往过度依赖源模型，并依赖有限且固定的变换技术。为克服这些局限，我们提出一种新的「从变换到迁移攻击」（TTA）方法。该方法引入可学习的变换机制，自适应地选择最优变换组合以最大化输入多样性，并引入积分梯度以缓解对源模型的过度依赖，从而改进攻击优化过程。大量实验表明，TTA 在下游任务中取得出色的攻击性能，在不同 VLP 架构上均优于当前最优的攻击方法。
英文原文摘要 1059 字符
Vision-Language Pre-training (VLP) models, while achieving state-of-the-art performance on various multimodal tasks, exhibit significant vulnerability to multimodal adversarial examples. In black-box attack scenarios of VLP models, a key challenge lies in the limited transferability of these adversarial examples. Existing methods to enhance transferability often suffer from an excessive dependence on the source model and a reliance on limited and fixed transformation techniques. To overcome these limitations, we propose a novel Transform to Transfer Attack (TTA) method. Our approach introduces a learnable transformation mechanism that adaptively selects optimal combinations of transformations to maximize input diversity, and incorporates integrated gradients to mitigate over-reliance on the source model, thereby refining the attack optimization process. Extensive experiments demonstrate that TTA achieves outstanding attack performance in downstream tasks, outperforming current state-of-the-art attack methods across different VLP architectures.