FLARE: A Backdoor Attack to Federated Learning with Refined Evasion

Federated Learning (FL) is vulnerable to backdoor attacks in which attackers inject malicious behaviors into the global model. To counter these attacks, existing works mainly introduce sophisticated defenses by analyzing model parameters and utilizing robust aggregation strategies. However, we find...

Full description

Saved in:
Bibliographic Details
Main Authors: Qingya Wang, Yi Wu, Haojun Xuan, Huishu Wu
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Mathematics
Subjects:
Online Access:https://www.mdpi.com/2227-7390/12/23/3751
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850060413898588160
author Qingya Wang
Yi Wu
Haojun Xuan
Huishu Wu
author_facet Qingya Wang
Yi Wu
Haojun Xuan
Huishu Wu
author_sort Qingya Wang
collection DOAJ
description Federated Learning (FL) is vulnerable to backdoor attacks in which attackers inject malicious behaviors into the global model. To counter these attacks, existing works mainly introduce sophisticated defenses by analyzing model parameters and utilizing robust aggregation strategies. However, we find that FL systems can still be attacked by exploiting their inherent complexity. In this paper, we propose a novel three-stage backdoor attack strategy named FLARE: A Backdoor Attack to Federated Learning with Refined Evasion, which is designed to operate under the radar of conventional defense strategies. Our proposal begins with a trigger inspection stage to leverage the initial susceptibilities of FL systems, followed by a trigger insertion stage where the synthesized trigger is stealthily embedded at a low poisoning rate. Finally, the trigger is amplified to increase the attack’s success rate during the backdoor activation stage. Experiments on the effectiveness of FLARE show significant enhancements in both the stealthiness and success rate of backdoor attacks across multiple federated learning environments. In particular, the success rate of our backdoor attack can be improved by up to 45× compared to existing methods.
format Article
id doaj-art-618d248ad7a549a9b6d707cb37ba061f
institution DOAJ
issn 2227-7390
language English
publishDate 2024-11-01
publisher MDPI AG
record_format Article
series Mathematics
spelling doaj-art-618d248ad7a549a9b6d707cb37ba061f2025-08-20T02:50:33ZengMDPI AGMathematics2227-73902024-11-011223375110.3390/math12233751FLARE: A Backdoor Attack to Federated Learning with Refined EvasionQingya Wang0Yi Wu1Haojun Xuan2Huishu Wu3Faculty of Law, University of Montreal, 2900 Edouard Montpetit Blvd, Montreal, QC H3T 1J4, CanadaChina Academy of Information and Communications Technology, Beijing 100191, ChinaSchool of Cyberspace Science and Technology, Beijing Institute of Technology, Beijing 100081, ChinaFaculty of Law, University of Montreal, 2900 Edouard Montpetit Blvd, Montreal, QC H3T 1J4, CanadaFederated Learning (FL) is vulnerable to backdoor attacks in which attackers inject malicious behaviors into the global model. To counter these attacks, existing works mainly introduce sophisticated defenses by analyzing model parameters and utilizing robust aggregation strategies. However, we find that FL systems can still be attacked by exploiting their inherent complexity. In this paper, we propose a novel three-stage backdoor attack strategy named FLARE: A Backdoor Attack to Federated Learning with Refined Evasion, which is designed to operate under the radar of conventional defense strategies. Our proposal begins with a trigger inspection stage to leverage the initial susceptibilities of FL systems, followed by a trigger insertion stage where the synthesized trigger is stealthily embedded at a low poisoning rate. Finally, the trigger is amplified to increase the attack’s success rate during the backdoor activation stage. Experiments on the effectiveness of FLARE show significant enhancements in both the stealthiness and success rate of backdoor attacks across multiple federated learning environments. In particular, the success rate of our backdoor attack can be improved by up to 45× compared to existing methods.https://www.mdpi.com/2227-7390/12/23/3751federated learningbackdoor attackstrigger inspectionpoisoning rate
spellingShingle Qingya Wang
Yi Wu
Haojun Xuan
Huishu Wu
FLARE: A Backdoor Attack to Federated Learning with Refined Evasion
Mathematics
federated learning
backdoor attacks
trigger inspection
poisoning rate
title FLARE: A Backdoor Attack to Federated Learning with Refined Evasion
title_full FLARE: A Backdoor Attack to Federated Learning with Refined Evasion
title_fullStr FLARE: A Backdoor Attack to Federated Learning with Refined Evasion
title_full_unstemmed FLARE: A Backdoor Attack to Federated Learning with Refined Evasion
title_short FLARE: A Backdoor Attack to Federated Learning with Refined Evasion
title_sort flare a backdoor attack to federated learning with refined evasion
topic federated learning
backdoor attacks
trigger inspection
poisoning rate
url https://www.mdpi.com/2227-7390/12/23/3751
work_keys_str_mv AT qingyawang flareabackdoorattacktofederatedlearningwithrefinedevasion
AT yiwu flareabackdoorattacktofederatedlearningwithrefinedevasion
AT haojunxuan flareabackdoorattacktofederatedlearningwithrefinedevasion
AT huishuwu flareabackdoorattacktofederatedlearningwithrefinedevasion