arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2411.07036cs.CR
收藏

ProP:面向过参数化模型的高效后门检测方法——基于传播扰动

ProP: Efficient Backdoor Detection via Propagation Perturbation for Overparametrized Models

Tao Ren, Qiongxiu Li

AI总结:

针对过参数化模型的后门攻击威胁,本文提出ProP方法,通过传播扰动与统计输出分布,利用良性分数在无先验知识下高效检测后门模型及其目标类别,实验证明其精度与效率优于现有方法。

摘要:

后门攻击对机器学习模型的安全性构成了严峻挑战,尤其是对深度神经网络等过参数化模型。本文提出ProP(Propagation Perturbation),一种新颖且可扩展的后门检测方法。该方法利用统计输出分布来识别后门模型及其目标类别,而无需依赖穷举式优化策略。ProP引入了一个新指标——良性分数(benign score),用于量化输出分布并有效区分良性模型与后门模型。与现有方法不同,ProP在最小假设下运行,不需要触发器或恶意样本的先验知识,使其高度适用于现实场景。在多种流行后门攻击上的广泛实验验证表明,ProP实现了高检测精度与计算效率,优于现有方法。这些结果凸显了ProP作为鲁棒且实用的后门检测方案的潜力。

英文摘要:

Backdoor attacks pose significant challenges to the security of machine learning models, particularly for overparameterized models like deep neural networks. In this paper, we propose ProP (Propagation Perturbation), a novel and scalable backdoor detection method that leverages statistical output distributions to identify backdoored models and their target classes without relying on exhausive optimization strategies. ProP introduces a new metric, the benign score, to quantify output distributions and effectively distinguish between benign and backdoored models. Unlike existing approaches, ProP operates with minimal assumptions, requiring no prior knowledge of triggers or malicious samples, making it highly applicable to real-world scenarios. Extensive experimental validation across multiple popular backdoor attacks demonstrates that ProP achieves high detection accuracy and computational efficiency, outperforming existing methods. These results highlight ProP's potential as a robust and practical solution for backdoor detection.

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。

↑