GNDPL: graph-neural and deep representation learning-based vulnerability detection in power systems

Current vulnerability detection methods in power systems often suffer from high false positives and limited coverage, especially in complex environments. To tackle these issues, this paper introduces GNDPL, a framework that combines Graph Neural Networks (GNN) with Deep Feature Representation Learning (DFRL). By leveraging multiple code representations and lexical features, GNDPL enhances detection accuracy and robustness. Experiments show it outperforms traditional methods and other machine learning models, making it a strong tool for securing modern power systems.