Peak-controlled logits poisoning attack in federated distillation

doi:10.1007/s10791-025-09742-8

	Peak-controlled logits poisoning attack in federated distillation
	Tang, Yuhan 1,2; Wu, Zhiyuan 3,4; Gao, Bo 1,2; Wen, Tian 3; Wang, Yuwei 3; Sun, Sheng 3
	2025-10-22
发表期刊	DISCOVER COMPUTING
ISSN	2948-2984
卷号	28 期号:1 页码:18
摘要	Federated Distillation (FD) is an innovative distributed machine learning paradigm that enables efficient and flexible cross-device knowledge transfer through knowledge distillation, without the need to upload large-scale model parameters to a central server. Although FD has attracted increasing attention in recent years, its security aspects remain relatively underexplored. Existing attack methods targeting traditional federated learning mainly focus on the transmission of model parameters and gradients, while attacks specifically designed for the unnormalized outputs (logits) in the emerging FD paradigm are still lacking. To fill this research gap and contribute to the enhancement of FD's security, we previously proposed the Federated Distillation Logits Attack (FDLA), which manipulates the logits transmitted during communication to mislead and degrade the performance of client models. However, FDLA has limitations in controlling its impact on participants with different roles or identities and lacks a systematic investigation into the effects of malicious interventions at various stages of knowledge transfer. To overcome these limitations, we propose a more advanced and controllable logits poisoning method-Peak-Controlled Federated Distillation Logits Attack (PCFDLA). PCFDLA enhances the effectiveness of FDLA by precisely controlling the peak values of logits to adjust the intensity of the attack. This method generates highly misleading perturbations that achieve stronger attack performance while maintaining a similar level of stealthiness to FDLA when detection is based on differences in model parameters. Moreover, we introduce a novel evaluation metric to more comprehensively assess the performance of such attacks. Experimental results show that PCFDLA significantly increases the destructive impact on victim models while maintaining high stealth. It consistently achieves superior performance across multiple datasets, highlighting its potential threat to the security of federated distillation systems.
关键词	Federated learning Knowledge distillation Knowledge transfer Poisoning attack Misleading attack
DOI	10.1007/s10791-025-09742-8
收录类别	SCI
语种	英语
资助项目	Fundamental Research Funds for the Central Universities
WOS研究方向	Computer Science
WOS类目	Computer Science, Information Systems
WOS记录号	WOS:001598387200001
出版者	SPRINGER
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/41633
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Gao, Bo
作者单位	1.Beijing Jiaotong Univ, Engn Res Ctr Network Management Technol High Speed, Sch Comp Sci & Technol, Minist Educ, Beijing, Peoples R China 2.Beijing Jiaotong Univ, Collaborat Innovat Ctr Railway Traff Safety, Beijing, Peoples R China 3.Chinese Acad Sci, Inst Comp Technol, Beijing, Peoples R China 4.Univ Chinese Acad Sci, Beijing, Peoples R China
推荐引用方式 GB/T 7714	Tang, Yuhan,Wu, Zhiyuan,Gao, Bo,et al. Peak-controlled logits poisoning attack in federated distillation[J]. DISCOVER COMPUTING,2025,28(1):18.
APA	Tang, Yuhan,Wu, Zhiyuan,Gao, Bo,Wen, Tian,Wang, Yuwei,&Sun, Sheng.(2025).Peak-controlled logits poisoning attack in federated distillation.DISCOVER COMPUTING,28(1),18.
MLA	Tang, Yuhan,et al."Peak-controlled logits poisoning attack in federated distillation".DISCOVER COMPUTING 28.1(2025):18.