AICAttack：基于注意力优化的对抗性图像字幕攻击
AICAttack: Adversarial Image Captioning Attack with Attention-Based Optimization
AI总结：
针对图像字幕模型对抗鲁棒性研究不足的问题，提出黑盒攻击策略AICAttack，通过注意力机制选攻击像素、定制差分进化优化扰动，在基准数据集上实现了优于前沿技术的攻击成功率。
摘要：
深度学习研究的最新进展已在计算机视觉（CV）和自然语言处理（NLP）的诸多任务中取得显著成果。图像字幕是CV与NLP的交叉领域问题，目前相关模型针对对抗攻击的鲁棒性尚未得到充分研究。本文提出一种新型对抗攻击策略——AICAttack（基于注意力的图像字幕攻击），旨在通过对图像施加微小扰动来攻击图像字幕模型。该算法在黑盒攻击场景下运行，无需获取目标模型的架构、参数或梯度信息。我们引入了基于注意力的候选选择机制，用于识别最优攻击像素，随后采用定制化差分进化方法优化像素RGB值的扰动。我们在基准数据集上针对多个受害模型开展了大量实验，验证了AICAttack的有效性。实验结果表明，该方法的攻击成功率始终更高，性能优于当前前沿技术。
英文摘要：
Recent advances in deep learning research have shown remarkable achievements across many tasks in computer vision (CV) and natural language processing (NLP). At the intersection of CV and NLP is the problem of image captioning, where the related models' robustness against adversarial attacks has not been well studied. This paper presents a novel adversarial attack strategy, AICAttack (Attention-based Image Captioning Attack), designed to attack image captioning models through subtle perturbations on images. Operating within a black-box attack scenario, our algorithm requires no access to the target model's architecture, parameters, or gradient information. We introduce an attention-based candidate selection mechanism that identifies the optimal pixels to attack, followed by a customised differential evolution method to optimise the perturbations of pixels' RGB values. We demonstrate AICAttack's effectiveness through extensive experiments on benchmark datasets against multiple victim models. The experimental results demonstrate that our method outperforms current leading-edge techniques by achieving consistently higher attack success rates.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
