arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2010.08138cs.CRcs.CV
收藏

输入感知动态后门攻击

Input-Aware Dynamic Backdoor Attack

VinAI Research(VinAI研究院);Hanoi University of Science and Technology(河内理工大学);VinUniversity

Anh Nguyen, Anh Tran

摘要:

近年来,神经后门攻击被认为是深度学习系统的潜在安全威胁。这样的系统虽然在干净数据上实现了最先进的性能,但在具有预定义触发器的输入上执行异常。然而,当前的后门技术依赖于统一的触发模式,这些模式很容易被当前的防御方法检测和缓解。在这项工作中,我们提出了一种新的后门攻击技术,其中的触发器不同的输入。为了实现这一目标,我们实现了一个输入感知触发器驱动的多样性损失。一种新的交叉触发器测试应用于强制触发器不可重用性,使后门验证不可能。实验表明,我们的方法是有效的,在各种攻击场景以及多个数据集。我们进一步证明了我们的后门可以绕过最先进的防御方法。一个著名的神经网络检查器的分析再次证明了所提出的攻击的隐蔽性。我们的代码可在https://github.com/VinAIResearch/input-aware-backdoor-attack-release上公开获取。

英文摘要:

In recent years, neural backdoor attack has been considered to be a potential security threat to deep learning systems. Such systems, while achieving the state-of-the-art performance on clean data, perform abnormally on inputs with predefined triggers. Current backdoor techniques, however, rely on uniform trigger patterns, which are easily detected and mitigated by current defense methods. In this work, we propose a novel backdoor attack technique in which the triggers vary from input to input. To achieve this goal, we implement an input-aware trigger generator driven by diversity loss. A novel cross-trigger test is applied to enforce trigger nonreusablity, making backdoor verification impossible. Experiments show that our method is efficient in various attack scenarios as well as multiple datasets. We further demonstrate that our backdoor can bypass the state of the art defense methods. An analysis with a famous neural network inspector again proves the stealthiness of the proposed attack. Our code is publicly available at https://github.com/VinAIResearch/input-aware-backdoor-attack-release.

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。

↑