Computational Efficiency Trade-offs in Semi-Supervised and Unsupervised GNN Anomaly Detection

This report synthesises findings from 8 peer-reviewed papers addressing the following research question: What is the computational efficiency trade-off between semi-supervised methods like Mul-GAD and fully unsupervised GNN anomaly detectors (e.g., DOMINANT) when scaling across graph sizes from 10K to. Abstract Graph neural networks (GNNs) have emerged as a powerful tool for effectively mining and learning from graph-structured data, with applications spanning numerous domains. However, most research focuses on static graphs, neglecting the dynamic nature of real-world. 10 claims were extracted from source literature; 9 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 8.2/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: What is the computational efficiency trade-off between semi-supervised methods like Mul-GAD and fully unsupervised GNN anomaly detectors (e.g., DOMINANT) when scaling across graph sizes from 10K to 100K nodes, as measured by inference latency per node?

Autonomous literature synthesis. Automated review score: 8.2/10. Full text and citation available at Assignee Research.