Machine learning modelling of materials: Max Veit

Materials make up everything from concrete and steel to solar cells and battery cathodes, even including things like petroleum, cell membranes, and water.

In order to predict how these materials will behave in real-world applications, scientists are employing a unique blend of physics, chemistry, and computer science to simulate materials at the atomistic scale.

A recent development in this field is the application of machine learning to simulate materials with quantum-mechanical accuracy at a small fraction of the traditional cost. We'll see how this method works, as well as the problems it can encounter in many types of materials. I'll also take you through the work I'm doing to overcome some of those problems and extend fast,accurate atomistic simulation to a new class of materials.


Max is a graduate student at Churchill College studying the CDT in Computational Methods for Materials Science at The Centre for Scientific Computing.

Department profile

Subscribe to the CHUtalks playlist

Presented at Churchill College, 20 October 2015.