Secure Network Attack Detection in Federated Learning via Attention-Enhanced Graph Neural Networks

Federated learning (FL) enables decentralized model training while preserving data privacy, but it remains vulnerable to network attacks that threaten system integrity. This paper proposes an effective network attack detection method for FL using graph neural networks (GNNs) with an attention mechanism. Devices and their communications are modeled as a graph, allowing the GNN to capture structural relationships and spatial dependencies. An attention mechanism is introduced to prioritize critical nodes and improve detection performance. The model is trained in a privacy-preserving manner using federated learning protocols. Experimental results in MATLAB demonstrate that the proposed approach significantly outperforms traditional methods in detection accuracy, convergence speed, and robustness under various network conditions. The method is intended to enhance the security of FL systems while maintaining privacy.

Authors

Publication Details

Journal
Archives des sciences a multidisciplinary journal
Published
2026-11-20
DOI
https://doi.org/10.68304/as/76507
Primary Topic
Privacy-Preserving Technologies in Data
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Secure Network Attack Detection in Federated Learning via Attention-Enhanced Graph Neural Networks

Qingyan Fang, Xide Wang
Archives des sciences a multidisciplinary journal
Privacy-Preserving Technologies in Data
article

Secure Network Attack Detection in Federated Learning via Attention-Enhanced Graph Neural Networks

Qingyan Fang, Xide Wang
article en

Abstract

Federated learning (FL) enables decentralized model training while preserving data privacy, but it remains vulnerable to network attacks that threaten system integrity. This paper proposes an effective network attack detection method for FL using graph neural networks (GNNs) with an attention mechanism. Devices and their communications are modeled as a graph, allowing the GNN to capture structural relationships and spatial dependencies. An attention mechanism is introduced to prioritize critical nodes and improve detection performance. The model is trained in a privacy-preserving manner using federated learning protocols. Experimental results in MATLAB demonstrate that the proposed approach significantly outperforms traditional methods in detection accuracy, convergence speed, and robustness under various network conditions. The method is intended to enhance the security of FL systems while maintaining privacy.

Archives des sciences a multidisciplinary journalVol. 76(5)
Privacy-Preserving Technologies in Data
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon