zum Inhalt springen

Scientific machine learning

One of our most recent research fields is scientific machine learning (SciML). Scientific machine learning is a new, rapidly developing field of research in which techniques of scientific computing and machine learning are combined and further developed. This results in hybrid methods which are applied for the discretization of partial differential equations, the development of fast and robust solvers and new modeling techniques.

In particular, we work on:

  • Domain decomposition methods using machine learning
  • Flow predictions using convolutional neural networks (CNNs)
  • Parameter identification using physics-informed networks (PINNs)
Prediction of the geometric location of coarse space constraints. Image from Heinlein, Lanser, Klawonn, Weber. "Machine Learning in Adaptive Domain Decomposition Methods - Predicting the Geometric Location of Constraints." SIAM J. Sci. Comput., Vol. 41, No. 6, A3887-A3912, 2019
Predicting the flow in an aneurysm using CNNs.