Money Mule Detection
Status:
The project starts in October 2026.
Researchers:
Dr. Lucrezia Bruni (Swissquote)
Industry Partner:
Swissquote
Description:
Fraud is a significant problem in the financial sector. If criminals would directly move fraudulently gained money to their own accounts, they could be identified, and thus, they often launder money through so-called mule accounts. However, detecting money mules is not easy as mule accounts can operate normally for a long time before being used for money mule operations, and large-scale fraud attacks are often executed during a short period of time, after which mule accounts are closed.
Most banks and financial institutions tackle fraud using a combination of rule-based detection systems and machine learning (ML) models, after which cases are analysed by human case managers. These systems can also be used to identify mule accounts. However, money mules are often particularly difficult to detect, as newly created accounts are abandoned soon after their first use. Moreover, fraudsters frequently change their attack patterns when previous ones get detected.
The goal of this project is to study money mule detection systems. The aim is to improve machine learning systems that can detect money mules and to develop strategies to adapt to changing attack patterns of money mules