PLACE FINANCIÈRE & MARCHÉS — Banques

New regulations: how to follow changes?

“How to develop a lean and mean compliance machine”


Retour au dossier

Deputy Head of SnT’s Software Verification and Validation Research Group, Prof. Domenico Bianculli, explains how his research group is using artificial intelligence to help businesses and public authorities in Luxembourg develop leaner compliance processes.

Software is everywhere nowadays, and businesses are investing large sums of money to develop and test their software systems. But according to Prof. Domenico Bianculli, Chief Scientist at the University of Luxembourg’s Interdisciplinary Centre for Security, Reliability and Trust (SnT), many companies still rely on manual processes to verify and validate systems or adopt highly sophisticated techniques that make sense in theory but fail to work in practice.

“Companies will often try to solve a problem by bringing in more engineers to test the software system or find solutions that cannot operate at scale within a business. That doesn’t solve the issues of automated validation and verification in an effective way,” says Prof. Bianculli, who is the Deputy Head of the Software Verification and Validation (SVV) Research Group at SnT.

We try to identify the problem together and understand the context in which the solution will operate for a specific domain.
Prof. Domenico Bianculli

Prof. Domenico Bianculli,  Chief Scientist in Software Engineering,  SnT

Instead, the native Italian believes every software system operates in a unique environment and requires a tailored solution. Unlike other research centres, the team applies context-driven research, which involves defining and solving a problem together with the main stakeholder. This collaborative approach to problem solving is a defining feature of SnT. “We try to identify the problem together and understand the context in which the solution will operate for a specific domain,” says Prof. Bianculli.

The SVV Research Group was established in 2012 with the goal of solving real-world engineering challenges in collaboration with industrial and public sector partners. The group’s main research areas include automated software testing, requirement engineering, verification, security analysis and testing, and regulatory compliance.

Aiding compliance through AI

Artificial intelligence is at the heart of the group’s activities. SVV researchers are currently working with the CSSF, Luxembourg’s regulator, to develop a tool that will automate compliance checks and flag any critical issues, such as an omission in a document, where human discretion is needed. “The end result is a much leaner compliance process,” says Prof. Bianculli.

If a design has to comply with regulations, one must take into account that the regulation might change. This change could have an impact on the system at any level.
Prof. Domenico Bianculli

Prof. Domenico Bianculli,  Chief Scientist in Software Engineering,  SnT

The Chief Scientist says developing a software that both meets end-user needs and fulfils the product requirements is becoming extremely important for businesses. Automating verification and validation activities can help achieve this.

Future-proofing systems

In the business world, it’s important to prepare for the future. According to Prof. Bianculli flexibility must be considered when designing a system, especially in areas like financial regulation, that evolve rapidly and frequently. “If a design has to comply with regulations, one must take into account that the regulation might change. This change could have an impact on the system at any level,” says Prof. Bianculli. He cites the example of GDPR, which required service providers to introduce new functionalities so that customers could access and export their personal data. A system designed to already support the dual operation of receiving and exporting data is a big step forward to compliance with the new regulations, according to Prof. Bianculli. “My advice is to design a system that enables one to be prepared for the evolution of that system and the environment in which it will operate,” he says.

Partnering with the SVV Lab

Like many of the other research groups at SnT, the SVV team uses an interdisciplinary approach to research. More specifically, they mainly rely on artificial intelligence, metaheuristic search, and model-driven engineering. Their ultimate goal is to build scalable and effective solutions that have a real impact on businesses and their software development. “Our job is more than just finding a solution to a problem – it is about trying to revolutionise the way our partners work,” says Prof. Bianculli.

Interested in partnering with the SVV Lab or another SnT Research Group? Contact the SnT Partnership Programme at [email protected] .