Magnetism by machine learning

In order to predict new magnetic materials with desired properties one needs to be able to scan large number of systems and if possible the whole set of possibilities in a given family of materials. To reach this goal one need to evaluate on-the-fly the magnetic interactions of any given system. The accurate calculation of the latter however presently requires time-consuming procedures.

In the field of magnetic materials, the use of deep learning methods is however quite uncommon and essentially focused on the determination of transition temperature, or phase diagrams and not on the determination of the magnetic interactions.

In this project, we propose to explore this new field by elaborating a machine learning methodology to predict the magnetic properties of metal-organic-frameworks (MOF).

The work will be done in the Condensed Matter Theory group at the Néel Institut in Grenoble. The Néel Institut is one of the best known laboratory in condensed Matter Physics. It is located in Grenoble, France in the Alps Mountains.

Contact : [email protected]