// the find
safe-graph/graph-fraud-detection-papers
A curated list of Graph/Transformer-based fraud, anomaly, and outlier detection papers & resources
A paper list tracking GNN and transformer-based fraud/anomaly detection research, organized by year and method type. It's a reading list, not a library — no runnable code, just links to papers and their venues. Useful as a literature survey for ML researchers working on fraud detection, graph anomaly detection, or related applied security problems.
Actively maintained through mid-2026 with recent NeurIPS/KDD/ICLR papers already indexed. Covers an unusually broad scope: node-level, edge-level, and graph-level anomaly detection, plus LLM-based approaches that most older lists miss. The companion interactive dashboard (paper_dashboard) lets you filter/search without scrolling through raw markdown. Code links are included where available, which saves time tracking down implementations.
It's purely a list — no summaries, no quality filtering, no indication of which papers actually reproduce well or matter. The code column is sparsely populated and many links go to anonymous OpenReview repos that expire. No non-graph baselines or classical methods for context, so it's useless if you want to understand what GNNs are actually beating. Organizing papers across sections by accepted venue year rather than arXiv date creates confusing placement (2023 papers appearing under 2024 sections).