XG-BoT: An explainable deep graph neural network for botnet detection and forensics

In this paper, we propose XG-BoT, an explainable deep graph neural network model for botnet node detection. The proposed comprises a detector and explainer automatic forensics. XG-BoT can effectively detect malicious nodes in large-scale networks. Specifically, it utilizes grouped reversible residual connection with isomorphism to learn expressive representations from communication graphs. explainer, based on the GNNExplainer saliency map perform forensics by highlighting suspicious flows related nodes. We evaluated using real-world, datasets. Overall, outperforms state-of-the-art approaches terms of key evaluation metrics. Additionally, demonstrate that explainers generate useful explanations