Detecting fraud better with predictive models, a challenge for society

LOGOS ITS has entered into a partnership with SnT, the Interdisciplinary Centre for Security, Reliability and Trust of the University of Luxembourg. Together, they aim to establish and hone predictive models to fight against fraud, financing of terrorism, and money laundering.

francois-gourdonLOGOS ITS has entered into a partnership with SnT, the Interdisciplinary Centre for Security, Reliability and Trust of the University of Luxembourg. Together, they aim to establish and hone predictive models to fight against fraud, financing of terrorism, and money laundering.

By Sébastien Lambotte

LOGOS ITS, the publisher of the software solution IDETECT, a big data analytics in real time platform, used by financial stakeholders to fight more efficiently against the financing of terrorism and money laundering, has just concluded a partnership with SnT.

The Interdisciplinary Center for Security, Reliability and Trust of the University of Luxembourg conducts research into partnership programms with economic stakeholders. “These partnerships enable us to guide research better so as to generate a positive impact on the economic environment,” according to Professor Björn Ottersten, director of SnT. Co-financed with the parties in place, these 3- or 4-year research programmes manage to combine economic needs with scientific interests.”

A challenge for society

olivier-merlanLOGOS ITS has found research teams in SnT capable of helping to improve the technology that supports the predictive analytical models of IDETECT solution. “The stakeholders are bound by legal obligations in terms of compliance. The aim is to prevent transactions used to finance terrorism or money laundering from going through traditional and regulated financial networks. But beyond obligations, there is a real challenge for society. A fraudulent transaction that is not detected can wreak havoc in certain regions of the world,” says Olivier Merlan, former Europol specialist and deputy general manager of Logos IT Services. “Today, the technologies within iDETECT are ways to improve the fight against such practices by relying on predictive analytical models. They help to detect fraudulent behaviour. The final purpose is to provide compliance officers with quality information, in real time, facilitating decision making” continues François Gourdon, managing director of LOGOS ITS.

Advantages of machine learning

At stake is being able to capitalise optimally on the available data volumes. “Technology now enables us to work on large quantities of data in real time,” says Radu State, Head of the Service and Data Management in Distributed Systems (SEDAN) research group in SnT. A sufficiently fine risk behaviour detection model is needed to avoid false positives, alerts or exclusions which have no reason to take place, and which can cause serious losses or require extensive verifications. But we cannot accept false negatives, i.e. fraudulent transactions get through.”

This partnership should thus make it possible to integrate new, finer and auto-learning models in iDETECT. This type of model uses data as a learning base.

More data means more various detected cases, and the more that the models will be self-adapt and improve its performance.

It’s a cerebral and high technology challenge. SnT and LOGOS ITS teams will work together with the aim of providing to the global financial industry real tools with very high added value, that help them to face the great challenges of the coming years.

Thus, “this partnership will also generate new knowledge that will benefit the entire scientific community in the fight against fraud, cyber-criminality, financing of terrorism and money laundering”, say Radu State and Olivier Merlan.

« Fin-Tech » is a perfect acronym for this fantastic 100% Luxembourgish partnership combining the worlds of finance and high technology .