Materials discovery by DFT and machine learning

Michael Sluydts^a,b, Michiel Larmuseau^a,b, Stefaan Cottenier ^a,b

^aDepartment of Electrical Energy, Metals, Mechanical Constructions and Systems, Ghent University, Belgium

^bCenter for Molecular Modeling, Ghent University, Belgium.

*stefaan.cottenier@ugent.be

Millenia ago, the only way to discover materials not yet known to mankind, was by observing attentively the natural world surrounding us. Later on, mankind learned to search for new materials in a laboratory, and made materials that nature does not spontaneously grow for us. By now, we have firmly entered the era in which new materials are also discovered by computers, and only afterwards confirmed by lab experiments. In this talk, I will discuss a strategy that combines machine learning with DFT to perform this task.