Digital currency money laundering identification model based on graph structure and transformer self-supervised learning

Abstract

Digital currency money laundering detection faces two major challenges: scarce labeled illicit transactions and complex fund-flow topologies. To address these issues, this study proposes GT-SSL, a graph-structured Transformer self-supervised learning framework for cryptocurrency anti-money laundering detection. The main contributions are threefold. First, transaction records are converted into directed attributed graphs and serialized through biased restart random walk, enabling Transformer-based modeling of local fund-flow contexts. Second, a structure-aware Transformer encoder is pre-trained through masked feature reconstruction and graph contrastive learning to improve representation learning under limited labels. Third, a two-stage pseudo-label self-training mechanism with community consensus filtering is designed to expand reliable training samples and reduce noisy pseudo-label propagation. Experiments on the Elliptic and AML-Bitcoin datasets show that GT-SSL achieves F1-scores of 95.80 and 93.78%, respectively, outperforming GCAF-AML, GNN-GRU and representative Elliptic benchmark baselines. Compared with recent competitive baselines, GT-SSL reduces the average false positive rate to 2.58% and the average false negative rate to 7.09%. These results demonstrate that integrating graph structure, Transformer-based self-supervision and community-guided pseudo-labeling can improve robust money laundering identification in label-scarce cryptocurrency transaction.

The authors' abstract, as published at the source. Discover Artificial Intelligence, 2026 · DOI ↗