Representation learning with a transformer by contrastive learning for money laundering detection

TL;DR

This paper presents a novel transformer-based contrastive learning approach for money laundering detection that effectively learns representations from structured time series data, enabling improved fraud detection with controlled false positives.

Contribution

It introduces a new procedure combining contrastive learning with transformers and a two-thresholds approach for money laundering detection, outperforming rule-based and LSTM methods.

Findings

Transformer produces general representations capturing laundering patterns.

Higher detection accuracy for fraudsters and nonfraudsters.

False positive rate effectively controlled using BH procedure.

Abstract

The present work tackles the money laundering detection problem. A new procedure is introduced which exploits structured time series of both qualitative and quantitative data by means of a transformer neural network. The first step of this procedure aims at learning representations of time series through contrastive learning (without any labels). The second step leverages these representations to generate a money laundering scoring of all observations. A two-thresholds approach is then introduced, which ensures a controlled false-positive rate by means of the Benjamini-Hochberg (BH) procedure. Experiments confirm that the transformer is able to produce general representations that succeed in exploiting money laundering patterns with minimal supervision from domain experts. It also illustrates the higher ability of the new procedure for detecting nonfraudsters as well as fraudsters, while keeping the false positive rate under control. This greatly contrasts with rule-based procedures or the ones based on LSTM architectures.