arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2007.03767cs.LGcs.CRstat.ML
收藏

学习率稳健的联邦学习后门防御

Defending against Backdoors in Federated Learning with Robust Learning Rate

The University of Texas at Dallas(德克萨斯大学达拉斯分校)

Mustafa Safa Ozdayi, Murat Kantarcioglu, Yulia R. Gel

摘要:

联邦学习(FL)允许一组代理协作训练模型,而无需共享其潜在的敏感数据。这使得FL适合隐私保护应用。同时,由于数据的分散性和不可见性,FL容易受到对抗攻击的影响。针对FL的一种重要攻击是后门攻击。在后门攻击中,攻击者试图在训练期间将后门功能嵌入模型,该后门功能稍后可以被激活以引起期望的错误分类。为了防止后门攻击,我们提出了一个轻量级的防御,需要最小的变化FL协议。在一个高层次上,我们的防御是基于仔细调整聚合服务器的学习率,每维和每轮,基于代理的更新的标志信息。我们首先推测在FL设置中进行成功的后门攻击的必要步骤,然后,明确地制定基于我们的推测的防御。通过实验,我们提供了支持我们的猜想的经验证据,我们测试我们的防御后门攻击在不同的设置。我们观察到,要么后门被完全消除,要么其准确性显着降低。总的来说,我们的实验表明,我们的防御显着优于一些最近提出的防御文献。我们通过对训练模型的准确性产生最小的影响来实现这一点。此外,我们还提供了我们提出的计划的收敛速度分析。

英文摘要:

Federated learning (FL) allows a set of agents to collaboratively train a model without sharing their potentially sensitive data. This makes FL suitable for privacy-preserving applications. At the same time, FL is susceptible to adversarial attacks due to decentralized and unvetted data. One important line of attacks against FL is the backdoor attacks. In a backdoor attack, an adversary tries to embed a backdoor functionality to the model during training that can later be activated to cause a desired misclassification. To prevent backdoor attacks, we propose a lightweight defense that requires minimal change to the FL protocol. At a high level, our defense is based on carefully adjusting the aggregation server's learning rate, per dimension and per round, based on the sign information of agents' updates. We first conjecture the necessary steps to carry a successful backdoor attack in FL setting, and then, explicitly formulate the defense based on our conjecture. Through experiments, we provide empirical evidence that supports our conjecture, and we test our defense against backdoor attacks under different settings. We observe that either backdoor is completely eliminated, or its accuracy is significantly reduced. Overall, our experiments suggest that our defense significantly outperforms some of the recently proposed defenses in the literature. We achieve this by having minimal influence over the accuracy of the trained models. In addition, we also provide convergence rate analysis for our proposed scheme.

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。

↑