Financial Crime & Fraud Detection Using Graph Computing: Application Considerations & Outlook

Abstract

In recent years, the unprecedented growth in digital payments fueled consequential changes in fraud and financial crimes. In this new landscape, traditional fraud detection approaches such as rule-based engines have largely become ineffective. AI and machine learning solutions using graph computing principles have gained significant interest. Graph neural networks and emerging adaptive solutions provide compelling opportunities for the future of fraud and financial crime detection. However, implementing the graph-based solutions in financial transaction processing systems has brought numerous obstacles and application considerations to light. In this paper, we overview the latest trends in the financial crimes landscape and discuss the implementation difficulties current and emerging graph solutions face. We argue that the application demands and implementation challenges provide key insights in developing effective solutions.

Deep Learning and Explainable Artificial Intelligence Techniques Applied for Detecting Money Laundering–A Critical Review

“…Graph Neural Networks is observed as one of the most commonly used method for detection of suspicious transaction and money laundering networks. Along with the potential of using graph neural networks, there are several challenges [104] such as complexities due to high-speed transaction systems, real-time systems, multi-channel updates, data size, data speed, data variety, and several business applications involved; which makes the practical implementation harder. Hence it is essential to consider the right infrastructure, appropriate tools and considerations for these challenges to improve the performance of future graphbased solutions.…”

“…• Graph mining and Social network analysis -These techniques are an apt solution for finding the patterns, groups and perform the link analysis, which is helpful for detection of money laundering, launderer gangs and identify the relationships to find more leads in the money laundering networks. Though there are challenges in graph mining from large data processing point of view [104], but this is one of the potential area where research should be continued.…”

Intelligent financial fraud detection practices in post-pandemic era

“…67 Furthermore, some applications represent transaction data as graphs, using nodes to represent financial entities and edges to represent money transfer. 121 After extracting features through feature engineering and graph-embedding techniques to preserve topological and structural properties, [122][123][124] machine learning models are built afterward. For example, Savage et al extracted meaningful communities from the network and performed classification to detect money-laundering activities.…”

“…Even so, they have yielded promising performance in financial crime and fraud detection, especially GNNs, which have the potential to improve structural representations and causal reasoning. 121 They broadly follow a recursive message passing schema, in which each node computes its new representation through aggregating feature vectors of its neighbors. 144 For instance, Weber et al applied Graph Convolutional Network (GCN), a typical GNN model, in anti-money laundering.…”

Banka Ödemelerinde Dolandırıcılığın Çizge Madenciliği ve Makine Öğrenimi Algoritmalarıyla Tespiti

“…Kurshan ve ark. [8] dolandırıcılık ve mali suç tespitinde grafik tabanlı çözümlerin karşılaştığı yaygın uygulama hususlarına ve kapsamlı uygulama zorluklarına genel bir bakış açısı sunmuştur. Çalışmada ödeme dolandırıcılığı, kimlik hırsızlığı, mali ve eski dolandırıcılar, hesap devralma, sentetik kimlik ve hesap dolandırıcılığı, kara para aklama ve diğer dolandırıcılık türlerine değinilmiştir.…”