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 2301.04554cs.CVcs.AI
收藏

通过基于密度的聚类和质心分析实现后门攻击的通用检测

Universal Detection of Backdoor Attacks via Density-based Clustering and Centroids Analysis

University of Siena(锡耶纳大学)

Wei Guo, Benedetta Tondi, Mauro Barni

AI总结:

提出CCA-UD通用后门防御,通过密度聚类训练样本并分析聚类质心特征诱发的误分类行为检测中毒数据,实验证明其跨攻击类型有效且优于现有方法。

摘要:

我们提出了一种基于聚类和质心分析(CCA-UD)的针对后门攻击的通用防御方法。该防御的目标是通过检查训练数据集来揭示深度神经网络模型是否遭受后门攻击。CCA-UD首先借助基于密度的聚类对训练集中的样本进行聚类。随后,它采用一种新颖策略来检测中毒聚类的存在。所提出策略基于一种普遍的误分类行为:当被分析聚类的代表性样本的特征被添加到良性样本中时,会观察到这种行为。诱发误分类错误的能力是中毒样本的一个普遍特征,因此所提出的防御与攻击类型无关。这标志着其与现有防御方法的显著区别:现有防御要么只能抵御某些类型的后门攻击,要么只有在满足关于投毒比例或攻击者所用触发信号类型的某些条件时才有效。在若干分类任务和网络架构上开展的实验,考虑了不同类型的后门攻击(使用干净标签或损坏标签)以及触发信号,包括全局和局部触发信号,以及样本特定和来源特定触发器;结果表明,所提出方法在所有情况下都能非常有效地防御后门攻击,并且始终优于最先进技术。

英文摘要:

We propose a Universal Defence against backdoor attacks based on Clustering and Centroids Analysis (CCA-UD). The goal of the defence is to reveal whether a Deep Neural Network model is subject to a backdoor attack by inspecting the training dataset. CCA-UD first clusters the samples of the training set by means of density-based clustering. Then, it applies a novel strategy to detect the presence of poisoned clusters. The proposed strategy is based on a general misclassification behaviour observed when the features of a representative example of the analysed cluster are added to benign samples. The capability of inducing a misclassification error is a general characteristic of poisoned samples, hence the proposed defence is attack-agnostic. This marks a significant difference with respect to existing defences, that, either can defend against only some types of backdoor attacks, or are effective only when some conditions on the poisoning ratio or the kind of triggering signal used by the attacker are satisfied. Experiments carried out on several classification tasks and network architectures, considering different types of backdoor attacks (with either clean or corrupted labels), and triggering signals, including both global and local triggering signals, as well as sample-specific and source-specific triggers, reveal that the proposed method is very effective to defend against backdoor attacks in all the cases, always outperforming the state of the art techniques.

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。

↑