Skip to content

Machine Learning for the Acceleration of Materials Discovery and Design


We are happy to welcome Matthew Evans as a joint Matgenix-UCLouvain BEWARE Fellowship. He will work on Machine-Learning techniques to accelerate existing techniques for the design of innovative materials.

First-principles or “ab initio” simulations make it possible to predict the properties of materials and compounds even before having synthesized them.
Some of these properties are very computationally expensive, making their large-scale application difficult or even impossible and economically unviable.
The objective of the MLxlMD project is to accelerate this type of simulation and to automate the process.
This will make it possible to apply this type of simulations to more realistic systems and to a large range of materials.
The properties targeted by this type of simulation are multiple: ionic conductivity of Lithium ions (materials for batteries), activity and selectivity of new catalytic materials, corrosion resistance, storage hydrogen in materials, chemical reactions, …
Being able to virtually test materials candidates provides a significant competitive advantage over traditional trial-and-error approaches.

Today, the MLxlMD project was officially kicked off in-person in Charleroi. The work is planned over three years during which existing methods will be benchmarked and automatized to accelerate materials research and development powered by machine-learning algorithms. Gian-Marco Rignanese, Guillaume Brunin and David Waroquiers discussed with Matthew Evans about the next steps in this ambitious project.