GraphShield: Advanced dynamic graph-based malware detection using graph neural networks

doi:10.1016/j.eswa.2025.129812

GraphShield: Advanced dynamic graph-based malware detection using graph neural networks

2025-09-26

0

PRE

AI

Bander Ali Saleh Al-rimy

A

Alaa Mohasseb

DOI：10.1016/j.eswa.2025.129812

Original

Original request for help

Save

Abstract

En 中文

• GraphShield achieves 99.5 % F1-score for malware detection using GNNs. • BERT-based API clustering reduces dimensionality while preserving semantics. • Hybrid GNN architectures outperform traditional models by 58 % in accuracy. • GNN Explainer identifies critical API subgraphs for transparent malware analysis. • Dynamic graph modeling captures long-range behavioral dependencies in malware.

GraphShield: Advanced dynamic graph-based malware detection using graph neural networks

Abstract

Journal

Organization

Cited Papers