Manuel Marschall

Short CV Notes

Research interests

Association to the Mathematical Topic "Numerical Methods for PDEs with Stochastic Data".

  • Adaptive spectral methods for PDE with stochastic data
  • Bayesian inversion
  • Low-rank tensor approximations
  • Stochastic Galerkin methods
  • FEM a posteriori error estimators
  • Random surface scattering
  • Random domain mapping method

Contact details

E-mail Manuel.Marschall-please remove this
Phone +49 (0) 30 20372 503
Fax +49 (0) 30 20372 412

Current projects

  • Bayesian inversion for high-dimensional parametric partial differential equations.
    Joint project with R. Schneider (TU Berlin) and M. Eigel (WIAS).

Notes and tutorials

The numerical examples are created using a Docker environment. A simplified installation guide is given here. For the installation, the image[22.2.2019] is used, implementing libraries such as:

  • FEniCS for the solution of partial differential equations
  • Xerus to handle computational intensive tensor arithmetic

The following Jupyter-Notebooks are created during the Daedalus workshop at TU-Berlin and maintained afterwards. You can download the *.ipynb files here.

1. Finite Element Method with FEniCSNotebook
2. Introduction to Random fields Part INotebook
3. Random fields Part II and KL expansionNotebook
4. Stochastic GalerkinNotebook
5. Introduction to a tensor library: XerusNotebook
6. Tensor reconstruction and low-rank SGFEMNotebook
7. Inverse problems and the Bayesian approachNotebook

Short CV

Since September 2016Phd candidate, research group of Prof. Hömberg, WIAS, Berlin
August 2016M.Sc. in Wirtschaftsmathematik at TU-Berlin. „Bayesian Inversion Using Hierarchical Tensor Approximation“