Area of Interest

In machine learning, we have two main areas of interest. The first one is the use of geometric and topological methods to improve the effectiveness of training machine learning models. Currently many models are have to learn known symmetries in data. Geometric machine learning is concerned with learning under symmetry constraints and geometric methods can achieve better performance with less data and shorter training time.

A second focus area is decision-making under uncertainty. We study uncertainty propagation in machine learning and simulation methods. A particular interest are multi-modal distributions and how to summarize them in meaningful ways. These arises for example in the context of weather prediction, with applications to wind and solar energy.

For more information, please contact Nello Blaser.

In optimization, we are interested in designing efficient and robust solutions for large-scale problems that exist in energy section. 

For more information, please contact Ahmad Hemmati.

For more information, please contact Magne Haveraaen.