Guarantee optimal recovery from COVID-19 crisis thanks to Machine Learning

Published on 02/07/2020


With the emergence and spread of the new SARS-Cov-2 coronavirus, many countries adopted strict health measures to protect their citizens. In a world in limbo, certain daily habits had to stop, as well as the economy slowing down. As the virus recedes, governments take steps to take steps in deconfinement phases. However, what measures should be taken to minimise the risk of a new pandemic while ensuring a return of economic activity?

"The objective of the REBORN project is to use a hybrid technological approach associating models and artificial intelligence in order to develop various scenarios of what could happen when taking this or that choice. In short, it is a matter of providing a decision-making aid which will enable political decision-makers to deal with such a pandemic by knowing the potential impact of their measures," explains Jean-Sébastien Sottet, LIST researcher involved in the REBORN project.

Through this research, in which LIST actively contributes, researchers use a large amount of data linked to the pandemic, as well as computer models, and integrate them into the Machine Learning system. The latter has the characteristic of learning by itself improving all the time as is fed with so-called learning data.

For example, the rapid deconfinement of the population in the United States has led to a dramatic increase in the number of people affected by COVID-19, further undermining the resilience of the health system and the economy. The creation of multiple scenarios through the REBORN project shows how to avoid this type of critical situation. "Our research may well be interested in Luxembourg health and economic impacts regarding the reopening of free movement of goods, or even road traffic in general." Illustrates Jean-Sébastien.


"As a partner of the REBORN project, we are working closely with the SnT at the University of Luxembourg, which is the coordinator of this research. Our goal is to go beyond dealing with the COVID-19 crisis by developing innovative tools and methods that policy makers can use to stem various health crises. Therefore, the project can be compared to the concept of Digital-Twin, which consists in replicating an existing situation with a digitalised system to, in this context, stem the crisis in the country.

The LIST team therefore, uses its expertise and experience to develop models capable of integrating with Machine Learning managed by SnT. In particular, these models will make it possible to specify or provide cause and effect relationships, and an additional level of detail on predictive scenarios. Therefore we will be able to focus on the impacts on certain business sectors, and even consequences between sectors”.


“I am responsible for the work carried out by LIST within the framework of the REBORN project. My colleague Sébastien Pineau and I are in close contact with coordinators in order to define our roles. At LIST, I work with my entire team on the development of modelling as part of this innovative hybrid technological approach that we want to implement. More specifically, I focus on the technical part of the project, while Sébastien Pineau takes care of relations with our multiple partners for data collection”.


Jean-Sébastien Sottet obtained his thesis on engineering focused on models applied to the design of human-machine interfaces in 2008 at the University of Grenoble. Subsequently, he became interested in various applications of computer systems modelling: reverse engineering during a post-doctorate at the École des Mines de Nantes and INRIA (FR), then support and software development at OneTree Technologies (SME, Luxembourg). In 2011, he joined CRP Henri Tudor (which became LIST in 2015 following a merger) where he worked on multiple aspects related to the modelling of interactive systems, and more broadly of enterprise information systems. He is also interested in the more fundamental aspects of modelling, and in particular its link with linguistics. In 2015, his research activities focused on information ecosystems modelling applied to risk management and regulation (eg GDPR). This ecosystem modelling associated with data processing and artificial intelligence leads to the notion of Digital Twin, fundamental for predicting and managing crisis situations.



Share this page:


Dr Jean-Sebastien SOTTET
Dr Jean-Sebastien SOTTET
Send an e-mail