Causal Adaptive Meta-Path Gating for Dynamic Heterogeneous Graph-Based Financial Fraud Detection

Authors

  • Seojun Rivera Donald Bren School of Information and Computer Sciences, University of California, Irvine, USA
  • Serena Edwards Department of Computer Science, North Carolina State University, USA
  • Liam Park Department of Computer Science, North Carolina State University, USA

DOI:

https://doi.org/10.54097/4jgjh479

Keywords:

Anti-money laundering, heterogeneous graphs, adaptive meta-paths, temporal representation, financial fraud, synthetic evaluation

Abstract

Collusive financial fraud often appears through heterogeneous relational paths whose explanatory value changes over time. Building on the metapath-guided temporal modeling of Audit-HCL, we propose CAMP-Gate, a lightweight context-adaptive framework that uses only information available at the end of each period: it constructs three causal meta-path views with retained-history and drift descriptors, cross-fits path-specific classifiers by account, and fuses their probabilities with a small multilayer perceptron gate conditioned on account-time context. In a heterogeneous-transaction environment with planted collusion rings and a held-out future split, CAMP-Gate obtains ROC-AUC 0.898, F1 0.580 and Recall@10% 0.597, improving on the best static single-path expert (ROC-AUC 0.887, Recall@10% 0.578) and on uniform temporal fusion at the 10% review budget (0.519), while uniform fusion attains the highest AP. Ablations identify retained path history as the main temporal driver, and the framework offers an interpretable, low-cost basis for budget-constrained alert prioritization.

Downloads

Download data is not yet available.

References

[1] Weber, M., et al. (2019). Anti-Money Laundering in Bitcoin: Experimenting with Graph Convolutional Networks for Financial Forensics. KDD Workshop on Anomaly Detection in Finance. arXiv:1908.02591.

[2] Altman, E., et al. (2023). Realistic Synthetic Financial Transactions for Anti-Money Laundering Models. Advances in Neural Information Processing Systems, 36. https://doi.org/10.52202/075280-1300 DOI: https://doi.org/10.52202/075280-1300

[3] Wang, X., et al. (2019). Heterogeneous Graph Attention Network. In Proceedings of the WWW (pp. 2022–2032). https://doi.org/10.1145/3308558.3313562 DOI: https://doi.org/10.1145/3308558.3313562

[4] Chen, J., Han, X., Zhang, X., & Yang, Y. (2023). Exact concept-level explanations for early detection of financial misstatements. Eurasia Journal of Science and Technology, 5(2), 34–48. DOI: https://doi.org/10.61784/ejst4162

[5] Wang, J., Ge, Y., Liu, J., & Jiang, B. (2024). Early warning of liquidity stress in U.S. banks with seasonally aligned frailty hazards and delayed-feedback conformal alarms. Journal of Trends in Finance and Economics, 1(1), 85–100. DOI: https://doi.org/10.61784/jtfe4078

[6] Xu, D., et al. (2020). Inductive Representation Learning on Temporal Graphs. In Proceedings of the ICLR. arXiv:2002.07962.

[7] Rossi, E., et al. (2020). Temporal Graph Networks for Deep Learning on Dynamic Graphs. arXiv:2006.10637.

[8] Dou, Y., et al. (2020). Enhancing Graph Neural Network-Based Fraud Detectors Against Camouflaged Fraudsters. In Proceedings of the CIKM (pp. 315–324). https://doi.org/10.1145/3340531.3411903 DOI: https://doi.org/10.1145/3340531.3411903

[9] Liu, Y., et al. (2021). Pick and Choose: A GNN-Based Imbalanced Learning Approach for Fraud Detection. In Proceedings of the WWW (pp. 3168–3177). https://doi.org/10.1145/3442381.3449989 DOI: https://doi.org/10.1145/3442381.3449989

[10] Kipf, T. N., & Welling, M. (2017). Semi-Supervised Classification with Graph Convolutional Networks. In Proceedings of the ICLR. arXiv:1609.02907.

[11] Hamilton, W. L., Ying, Z., & Leskovec, J. (2017). Inductive Representation Learning on Large Graphs. Advances in Neural Information Processing Systems, 30.

[12] Veličković, P., et al. (2018). Graph Attention Networks. In Proceedings of the ICLR. arXiv:1710.10903.

[13] Zhou, F. (2026). Moment-constrained side-force allocation for wind-assisted merchant ships: Hydrodynamic performance of retractable anti-leeway fins and flap rudders. Innovation and Technology Studies, 3(1), 27–41. DOI: https://doi.org/10.61784/its4027

[14] Cardoso, M., Saleiro, P., & Bizarro, P. (2022). LaundroGraph: Self-Supervised Graph Representation Learning for Anti-Money Laundering. In Proceedings of the ICAIF. https://doi.org/10.1145/3533271.3561727 DOI: https://doi.org/10.1145/3533271.3561727

[15] Lo, W. W., et al. (2023). Inspection-L: Self-Supervised GNN Node Embeddings for Money Laundering Detection in Bitcoin. Applied Intelligence, 53, 19406–19417. https://doi.org/10.1007/s10489-023-04504-9 DOI: https://doi.org/10.1007/s10489-023-04504-9

[16] Teng, D. (2025). TRIAD: Deadline- and reliability-aware scheduling for post-earthquake structural assessment workflows. World Journal of Engineering Research, 3(6), 55–66. DOI: https://doi.org/10.61784/wjer4124

[17] Sankar, A., et al. (2020). DySAT: Deep Neural Representation Learning on Dynamic Graphs via Self-Attention Networks. In Proceedings of the WSDM (pp. 519–527). https://doi.org/10.1145/3336191.3371845 DOI: https://doi.org/10.1145/3336191.3371845

[18] You, Y., et al. (2020). Graph Contrastive Learning with Augmentations. Advances in Neural Information Processing Systems, 33, 5812–5823.

[19] Zhu, Y., et al. (2020). Deep Graph Contrastive Representation Learning. arXiv:2006.04131.

[20] Wang, J., Liu, J., Ge, Y., & Jiang, B. (2023). Risk-aware ensemble learning for credit default prediction under economic uncertainty. World Journal of Economics and Business Research, 1(1), 18–32. DOI: https://doi.org/10.61784/wjebr4124

[21] Wang, X., Liu, N., Han, H., & Shi, C. (2021). Self-Supervised Heterogeneous Graph Neural Network with Co-Contrastive Learning. In Proceedings of the KDD (pp. 1726–1736). https://doi.org/10.1145/3447548.3467415 DOI: https://doi.org/10.1145/3447548.3467415

[22] Lin, Y., et al. (2026). Subgraph-Level Temporal Graph Transformer With Laundering-Motif Prototypes for Anti-Money Laundering on Blockchains. IEEE Access. https://doi.org/10.1109/ACCESS.2026.3731160 DOI: https://doi.org/10.1109/ACCESS.2026.3731160

[23] Ren, S., et al. (2025). ARCS: Adaptive Reinforcement Learning Framework for Automated Cybersecurity Incident Response Strategy Optimization. Applied Sciences, 15, 951. https://doi.org/10.3390/app15020951 DOI: https://doi.org/10.3390/app15020951

Downloads

Published

12-11-2026

Issue

Section

Articles

How to Cite

Rivera, S., Edwards, S., & Park, L. (2026). Causal Adaptive Meta-Path Gating for Dynamic Heterogeneous Graph-Based Financial Fraud Detection. International Journal of World Economic Research, 4(1), 15-18. https://doi.org/10.54097/4jgjh479