Namu Kroupa

PhD thesis: TBC

In recent years, machine learning algorithms have shown a strong ability to accurately predict data in the fields of materials science and molecular modelling. The main research focus is to understand precisely why certain methods, such as neural networks, are suited to this task, how architectural changes influence this ability and how this understanding can be translated into even better algorithms. this entails the exploration of the parameter space of the neural networks. As part of this, sampling algorithms will be advanced which explore the configuration space of molecules and materials, enabling property prediction from first principles. As the field of cosmology has been strongly developing these algorithms from the viewpoint of Bayesian inference, a further focal point will be working towards convergence of the two fields.

Who or what inspired you to pursue your research interests?

A key inspiration was my first summer project with my former Director of Studies, Dr Chris Lester (Peterhouse), an avid physicist and excellent advisor for my studies. Since taking the computer science lectures in my first-year physics course, my undeveloped interest grew in both directions. When I talked to him about it, he immediately offered me a magical project drawn from a dusty drawer spurring me to continue research at ever higher levels in a second project, where I first encountered machine learning in the context of physics, and my Master’s project. This fascinating work motivates me every day.