Motif-aware temporal GCN for fraud detection in signed cryptocurrency trust networks

TL;DR

This paper introduces a motif-aware temporal GCN model that leverages local topological motifs and balance theory to improve fraud detection in evolving signed cryptocurrency trust networks, outperforming existing methods.

Contribution

It proposes a novel temporal GCN incorporating motif matrices and balance theory for dynamic signed networks, enhancing fraud detection accuracy.

Findings

Outperforms existing models on bitcoin-alpha and bitcoin-otc datasets.

Effectively captures local topological structures and trust balance.

Improves fraud detection in evolving cryptocurrency networks.

Abstract

Graph convolutional networks (GCNs) is a class of artificial neural networks for processing data that can be represented as graphs. Since financial transactions can naturally be constructed as graphs, GCNs are widely applied in the financial industry, especially for financial fraud detection. In this paper, we focus on fraud detection on cryptocurrency truct networks. In the literature, most works focus on static networks. Whereas in this study, we consider the evolving nature of cryptocurrency networks, and use local structural as well as the balance theory to guide the training process. More specifically, we compute motif matrices to capture the local topological information, then use them in the GCN aggregation process. The generated embedding at each snapshot is a weighted average of embeddings within a time window, where the weights are learnable parameters. Since the trust networks is signed on each edge, balance theory is used to guide the training process. Experimental results on bitcoin-alpha and bitcoin-otc datasets show that the proposed model outperforms those in the literature.