Delator: Automatic Detection of Money Laundering Evidence on Transaction Graphs via Neural Networks

• Henrique S. Assumpção UFMG
• Fabrício Souza UFMG
• Leandro Lacerda Campos UFMG / InterMind
• Vinícius T. de Castro Pires UFMG / InterMind
• Paulo M. Laurentys de Almeida InterMind
• Fabrício Murai UFMG

DOI: https://doi.org/10.5753/brasnam.2022.223137

Abstract

Money laundering is one of the most relevant criminal activities today, due to its potential to cause massive financial losses to governments, banks, etc. We propose DELATOR, a new CAAT (computer-assisted audit technology) to detect money laundering activities based on neural network models that encode bank transfers as a large-scale temporal graph. In collaboration with a Brazilian bank, we design and apply an evaluation strategy to quantify DELATOR's performance on historic data comprising millions of clients. DELATOR outperforms an off-the-shelf solution from Amazon AWS by 18.9% with respect to AUC. We conducted real experiments that led to discovery of 8 new suspicious among 100 analyzed cases, which would have been reported to the authorities under the current criteria.

Keywords: graph neural networks, money laundering detection, multi-task learning, dynamic graphs

References

Assumpção, H. S., Souza, F., Lacerda Campos, L., Castro Pires, V. T., Laurentys de Almeida, P. M., and Murai, F. (2022). Delator: Automatic detection of money laundering evidence on transaction graphs via neural networks. arXiv preprint arXiv:2205.10293.

BACEN (2020a). Circular nº 3.978, de 23 de janeiro de 2020.

BACEN (2020b). Circular nº 4.001, de 29 de janeiro de 2020.

FATF (2012-2021). International standards on combating money laundering and the financing of terrorism & proliferation.

Fernández, A., Garcia, S., Herrera, F., and Chawla, N. V. (2018). Smote for learning from imbalanced data: progress and challenges, marking the 15-year anniversary. JAIR, 61:863–905.

Gopinathan, K. M., Biafore, L. S., Ferguson, W. M., Lazarus, M. A., Pathria, A. K., and Jost, A. (1998). Fraud detection using predictive modeling. US Patent 5,819,226.

Halbouni, S. S., Obeid, N., and Garbou, A. (2016). Corporate governance and information technology in fraud prevention and detection. Managerial Auditing Journal.

Hamilton, W., Ying, Z., and Leskovec, J. (2017). Inductive Representation Learning on Large Graphs. In NeurIPS, volume 30.

Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., and Liu, T.-Y. (2017). Lightgbm: A highly efficient gradient boosting decision tree. In NeurIPS, volume 30.

Liang, C., Liu, Z., Liu, B., Zhou, J., Li, X., Yang, S., and Qi, Y. (2019). Uncovering insurance fraud conspiracy with network learning. In SIGIR, pages 1181–1184.

Liu, Y., Ao, X., Qin, Z., Chi, J., Feng, J., Yang, H., and He, Q. (2021). Pick and choose: A gnn-based imbalanced learning approach for fraud detection. In WWW, pages 3168– 3177.

Othman, R., Aris, N. A., Mardziyah, A., Zainan, N., and Amin, N. M. (2015). Fraud detection and prevention methods in the malaysian public sector: Accountants’ and internal auditors’ perceptions. Procedia Economics and Finance, 28:59–67.

Pareja, A., Domeniconi, G., Chen, J., Ma, T., Suzumura, T., Kanezashi, H., Kaler, T., Schardl, T. B., and Leiserson, C. E. (2020). EvolveGCN: Evolving graph convolutional networks for dynamic graphs. In AAAI.

Pereira, R. and Murai, F. (2021). Quão efetivas são redes neurais baseadas em grafos na detecção de fraude para dados em rede? In BraSNAM, pages 205–210.

Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., and Gulin, A. (2018). Catboost: Unbiased boosting with categorical features. In NeurIPS, page 6639–6649.

Radford, A., Narasimhan, K., Salimans, T., and Sutskever, I. (2018). Improving language understanding by generative pre-training.

Wang, D., Lin, J., Cui, P., Jia, Q., Wang, Z., Fang, Y., Yu, Q., Zhou, J., Yang, S., and Qi, Y. (2019). A semi-supervised graph attentive network for financial fraud detection. In ICDM, pages 598–607.

Wang, J., Zhang, S., Xiao, Y., and Song, R. (2021). A review on graph neural network methods in financial applications.

Weber, M., Domeniconi, G., Chen, J., Weidele, D. K. I., Bellei, C., Robinson, T., and Leiserson, C. E. (2019). Anti-money laundering in bitcoin: Experimenting with graph convolutional networks for financial forensics. arXiv preprint arXiv:1908.02591.

Widuri, R. and Gautama, Y. (2020). Computer-assisted audit techniques (caats) for financial fraud detection: A qualitative approach. In ICIMTech, pages 771–776.

Zhao, T., Zhang, X., and Wang, S. (2021). Graphsmote: Imbalanced node classification on graphs with graph neural networks. In WSDM, pages 833–841.

Zheng, D., Song, X., Ma, C., Tan, Z., Ye, Z., Dong, J., Xiong, H., Zhang, Z., and Karypis, G. (2020). Dgl-ke: Training knowledge graph embeddings at scale. In SIGIR, pages 739–748.

Zhu, Y.-N., Luo, X., Li, Y.-F., Bu, B., Zhou, K., Zhang, W., and Lu, M. (2020). Heterogeneous mini-graph neural network and its application to fraud invitation detection. In ICDM, pages 891–899.