收藏切换
Federated Learning Adaptive Gradient Accumulation Backdoor Attack
收藏切换
Guifen Zhang1, 3, 4, Qinchun Su2, 3, 4, Hongzi Li2, 3, 4, Lina Ge1, 3, 4, *
Journal of Computer Research and Development | 2026, 63(9) : 2386 - 2400
Less
收藏切换
Journal of Computer Research and Development | 2026, 63(9): 2386-2400
• Network and Information Security •
Federated Learning Adaptive Gradient Accumulation Backdoor Attack
Full
Guifen Zhang1, 3, 4, Qinchun Su2, 3, 4, Hongzi Li2, 3, 4, Lina Ge1, 3, 4, *
Affiliations
  • 1School of Artificial Intelligence, Guangxi Minzu University, Nanning 530006
  • 2School of Physics and Electronic Information, Guangxi Minzu University, Nanning 530006
  • 3Key Laboratory of Intelligent Computing Network and Big Data Information Security (Guangxi Minzu University), Nanning 530006
  • 4Guangxi Key Laboratory of Hybrid Computation and IC Design Analysis (Guangxi Minzu University), Nanning 530006
  • born in 1974. Master, associate professor, master supervisor. Member of CCF. Her main research interests include information security and federated learning

    born in 1998. Master. His main research interests include network and information security, and federated learning

    born in 2000. Master. His main research interests include network and information security, and federated learning

    born in 1969. PhD, professor, PhD supervisor. Distinguished member of CCF. Her main research interests include artificial intelligence and information security

About Author:

born in 1974. Master, associate professor, master supervisor. Member of CCF. Her main research interests include information security and federated learning

born in 1998. Master. His main research interests include network and information security, and federated learning

born in 2000. Master. His main research interests include network and information security, and federated learning

born in 1969. PhD, professor, PhD supervisor. Distinguished member of CCF. Her main research interests include artificial intelligence and information security

Published: 2026-09-01 doi: 10.7544/issn1000-1239.202550637
Outline
收藏切换

Adaptive gradient accumulation backdoor attack (AGABA) framework is designed for federated learning (FL) systems, integrating a parameterized adaptive subblock trigger (AST) with dynamic transparency and a multistage gradient accumulation (MGA) mechanism. This framework effectively addresses the persistent trade-off dilemma between stealthiness and persistence that plagues traditional backdoor attacks in the FL environment. AST decomposes the complete trigger into multiple independent components via dynamic transparency control and distributed subblock superposition technology, allowing malicious clients to collaboratively construct a global trigger pattern while maintaining a high level of attack stealthiness. MGA mechanism adopts a three-phase attack strategy including initial accumulation, gradient accumulation and attack execution, and combines it with a parameter importance-aware mechanism. It realizes the latent hiding and targeted activation of malicious updates in model aggregation through cross-round progressive gradient accumulation. Moreover, AGABA adopts momentum-accelerated gradient divergence propagation and adaptive memory factor adjustment, ensuring that all malicious gradient updates fall within the legitimate distribution interval of the FL system and thus evading statistical anomaly-based detection mechanisms effectively. Experiments show that with 20% malicious clients participating in FL training, AGABA can maintain a favorable backdoor attack success rate even under the protection of various mainstream defense mechanisms, and its overall performance outperforms single backdoor attack methods.

federated learning  /  distributed  /  centralized  /  backdoor attacks  /  gradient accumulation  /  adaptive attacks
张桂芬, 苏秦春, 李鸿梓, 葛丽娜. 联邦学习自适应梯度累积后门攻击. 计算机研究与发展, 2026 , 63 (9) : 2386 -2400 . DOI: 10.7544/issn1000-1239.202550637
Guifen Zhang, Qinchun Su, Hongzi Li, Lina Ge. Federated Learning Adaptive Gradient Accumulation Backdoor Attack[J]. Journal of Computer Research and Development, 2026 , 63 (9) : 2386 -2400 . DOI: 10.7544/issn1000-1239.202550637
Year 2026 volume 63 Issue 9
81
39
Cite this Article
BibTeX
Article Info
doi: 10.7544/issn1000-1239.202550637
  • Receive Date:2025-08-27
  • Online Date:2026-11-17
  • Published:2026-09-01
Article Data
Affiliations
History
  • Received:2025-08-27
  • Revised:2026-05-07
  • Accepted:2026-05-14
Affiliations
    1School of Artificial Intelligence, Guangxi Minzu University, Nanning 530006
    2School of Physics and Electronic Information, Guangxi Minzu University, Nanning 530006
    3Key Laboratory of Intelligent Computing Network and Big Data Information Security (Guangxi Minzu University), Nanning 530006
    4Guangxi Key Laboratory of Hybrid Computation and IC Design Analysis (Guangxi Minzu University), Nanning 530006
References
Share
https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/10.7544/issn1000-1239.202550637
Share to
QR

Scan QR to access full text

Cite this article
BibTeX
Citations
表12种不同金属材料的力学参数
科
Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)
属
Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏
  • BibTeX
  • EndNote
  • RefWorks
  • TxT