A Learning-Based Attack Framework to Break SOTA Poisoning Defenses in Federated Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yang, Yuxin, Li, Qiang, Nie, Chenfei, Hong, Yuan, Pang, Meng, Wang, Binghui
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913444562731008
author Yang, Yuxin
Li, Qiang
Nie, Chenfei
Hong, Yuan
Pang, Meng
Wang, Binghui
author_facet Yang, Yuxin
Li, Qiang
Nie, Chenfei
Hong, Yuan
Pang, Meng
Wang, Binghui
contents Federated Learning (FL) is a novel client-server distributed learning framework that can protect data privacy. However, recent works show that FL is vulnerable to poisoning attacks. Many defenses with robust aggregators (AGRs) are proposed to mitigate the issue, but they are all broken by advanced attacks. Very recently, some renewed robust AGRs are designed, typically with novel clipping or/and filtering strate-gies, and they show promising defense performance against the advanced poisoning attacks. In this paper, we show that these novel robust AGRs are also vulnerable to carefully designed poisoning attacks. Specifically, we observe that breaking these robust AGRs reduces to bypassing the clipping or/and filtering of malicious clients, and propose an optimization-based attack framework to leverage this observation. Under the framework, we then design the customized attack against each robust AGR. Extensive experiments on multiple datasets and threat models verify our proposed optimization-based attack can break the SOTA AGRs. We hence call for novel defenses against poisoning attacks to FL. Code is available at: https://github.com/Yuxin104/ BreakSTOAPoisoningDefenses.
format Preprint
id arxiv_https___arxiv_org_abs_2407_15267
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Learning-Based Attack Framework to Break SOTA Poisoning Defenses in Federated Learning
Yang, Yuxin
Li, Qiang
Nie, Chenfei
Hong, Yuan
Pang, Meng
Wang, Binghui
Cryptography and Security
Federated Learning (FL) is a novel client-server distributed learning framework that can protect data privacy. However, recent works show that FL is vulnerable to poisoning attacks. Many defenses with robust aggregators (AGRs) are proposed to mitigate the issue, but they are all broken by advanced attacks. Very recently, some renewed robust AGRs are designed, typically with novel clipping or/and filtering strate-gies, and they show promising defense performance against the advanced poisoning attacks. In this paper, we show that these novel robust AGRs are also vulnerable to carefully designed poisoning attacks. Specifically, we observe that breaking these robust AGRs reduces to bypassing the clipping or/and filtering of malicious clients, and propose an optimization-based attack framework to leverage this observation. Under the framework, we then design the customized attack against each robust AGR. Extensive experiments on multiple datasets and threat models verify our proposed optimization-based attack can break the SOTA AGRs. We hence call for novel defenses against poisoning attacks to FL. Code is available at: https://github.com/Yuxin104/ BreakSTOAPoisoningDefenses.
title A Learning-Based Attack Framework to Break SOTA Poisoning Defenses in Federated Learning
topic Cryptography and Security
url https://arxiv.org/abs/2407.15267