arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2403.07673cs.CR
收藏

面向GAN图像翻译的模型提取攻击中的领域偏移缓解

Towards Model Extraction Attacks in GAN-Based Image Translation via Domain Shift Mitigation

Xiangtan University(湘潭大学);University of Technology Sydney(悉尼科技大学);Griffith University(格里菲斯大学);Huazhong University of Science and Technology(华中科技大学);City University of Macau(澳门城市大学)

Di Mi, Yanjun Zhang, Leo Yu Zhang, Shengshan Hu, Qi Zhong, Haizhuan Yuan, Shirui Pan

AI总结:

本文研究了GAN图像翻译任务中模型提取攻击的威胁,通过缓解领域偏移来改进攻击效果,实验表明新方法在多个图像翻译任务中表现优异,验证了攻击对实际API的严重性。

摘要:

模型提取攻击(MEAs)使攻击者能够通过远程查询受害深度神经网络(DNN)的API服务来复制其功能,对基于查询的DNN服务的安全性和完整性构成严重威胁。尽管大多数当前关于MEAs的研究主要集中在神经分类器上,但图像到图像翻译(I2IT)任务在日常活动中日益普及。然而,为DNN分类器开发的MEA技术无法直接应用于I2IT场景,导致I2IT模型对MEA攻击的脆弱性常被低估。本文从新视角揭示了I2IT任务中MEA的威胁。不同于传统方法通过缩小攻击者查询与受害者训练样本之间的分布差距,我们选择缓解由不同分布引起的效应,即领域偏移。这通过引入一个新的正则化项来惩罚高频噪声,并寻找一个更平坦的最小值以避免过度拟合偏移分布。在不同图像翻译任务(包括图像超分辨率和风格迁移)上对不同主干受害模型进行了大量实验,新设计在所有指标上均显著优于基线。此外,一些实际的I2IT API也被验证对我们的攻击极其脆弱,强调了增强防御和可能修订API发布政策的必要性。

英文摘要:

Model extraction attacks (MEAs) enable an attacker to replicate the functionality of a victim deep neural network (DNN) model by only querying its API service remotely, posing a severe threat to the security and integrity of pay-per-query DNN-based services. Although the majority of current research on MEAs has primarily concentrated on neural classifiers, there is a growing prevalence of image-to-image translation (I2IT) tasks in our everyday activities. However, techniques developed for MEA of DNN classifiers cannot be directly transferred to the case of I2IT, rendering the vulnerability of I2IT models to MEA attacks often underestimated. This paper unveils the threat of MEA in I2IT tasks from a new perspective. Diverging from the traditional approach of bridging the distribution gap between attacker queries and victim training samples, we opt to mitigate the effect caused by the different distributions, known as the domain shift. This is achieved by introducing a new regularization term that penalizes high-frequency noise, and seeking a flatter minimum to avoid overfitting to the shifted distribution. Extensive experiments on different image translation tasks, including image super-resolution and style transfer, are performed on different backbone victim models, and the new design consistently outperforms the baseline by a large margin across all metrics. A few real-life I2IT APIs are also verified to be extremely vulnerable to our attack, emphasizing the need for enhanced defenses and potentially revised API publishing policies.

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。

↑