DIREC project

Fraud Detection in payments

Project impact

Financial fraud is becoming more organized, automated, and harder to detect with traditional tools. Most fraud detection systems look at transactions on at a time – one login, one payment, or one device – which creates blind spots that professional fraud groups can exploit. By coordinating thousands of small and seemingly unrelated actions across many accounts, attackers hide their activity in the connections between users rather than in single events. 

The Fraud Detection in Payments project addresses this challenge by viewing digital financial activity as a network of connected relationships. This graph‑based approach makes it possible to move from reactive monitoring to proactive cyber defense. It helps protect critical Nordic payment infrastructure, supports citizens’ financial security, and strengthens technological independence through a Danish‑owned solution aligned with European standards for transparency, AI ethics, and data privacy. 

PROJECT DATA

Project name

Fraud detection in payments

Project period
2026-2027
Funding
DKK 400.000

Scientific mission

The project improves fraud detection by combining advanced graph analytics with real‑world performance requirements. Researchers work with Vipps MobilePay and Kynapse to develop a software‑based defense layer that can detect organized fraud in real time and at a large scale. The research focuses on two complementary methods: fast, in‑memory graph algorithms that can stop coordinated attacks within milliseconds, and deeper batch analysis of historical data to uncover hidden fraud networks and complex money-laundering patterns. 

Key scientific challenges include incremental clustering in very large graphs, density‑based clustering in non‑Euclidean spaces, and reliable, transparent graph neural networks that are resistant to manipulation. By integrating this research into an industrial platform, the project connects theory with practice and turns advanced graph science into a practical, low‑latency defense system for modern payment platforms. 

Project Participants

Ira Assent
Ira Assent – Professor – Aarhus University
Davide Mottin
Davide Mottin – Associate professor – Aarhus University
Anders Sandholm
Anders Sandholm – Founder and CEO – Kynapse Technologies
Eivind Bøhn
Eivind Bøhn – Risk Intelligence – Vipps MobilePay

Partners

Aarhus University logoVipps MobilePay logoKynapse Technologies logo