Applying a Learning and an Artificial Intelligence-based Framework for Fraud Detection in Anti-money Laundering
Year: 2025
Type: conference-paper
Abstract: As fraud methods in money laundering become increasingly complex, artificial intelligence techniques for identifying and combating these threats are becoming increasingly important and evident. Machine learning and deep learning approaches are subsets of artificial intelligence that can effectively discover hidden patterns and improve predictive performance by using large datasets. These learning-based algorithms can uncover hidden and suspicious patterns through their architectures, thereby assisting financial institutions in anti-money laundering fraud detection. To achieve the final architecture, we designed and evaluated both individual and ensemble models based on machine learning and deep learning. Hence, we selected an end-to-end architecture based on deep learning algorithms. Compared with competing computational and traditional accounting approaches (e.g., Benford’s law and rule-based systems), our artificial intelligence-based method is more effective at detecting money laundering. Using two benchmark datasets, we demonstrate the superiority of the proposed learning and artificial intelligence-based approach for fraud detection in anti-money laundering has been demonstrated. Finally, several experiments validate the effectiveness of the proposed method.
Institution: Graduate University of Advanced Technology
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204158640