Difference between revisions of "Kilian Röhner, M.Sc."

From Sccswiki
Jump to navigation Jump to search
Line 60: Line 60:
 
Some topic ideas:
 
Some topic ideas:
 
* Parallelization of model refinement techniques for sparse grid classification
 
* Parallelization of model refinement techniques for sparse grid classification
* Block- and Rectilinear Adaptivity for SGDE
+
* Block- and Rectilinear Adaptivity for Sparse Grid density Estimation
  
 
=== Running ===
 
=== Running ===

Revision as of 11:43, 7 June 2019

Kilian Roehner.jpg
Address:
TU München
Institut für Informatik
Boltzmannstr. 3
85748 Garching b. München
Office:
MI 02.05.041
Email:
Roehnermail.png
Phone:
(089) 289 18 638
Fax:
(089) 289 18 607
Office hours:
by arrangement

Research interests

  • Online Data Mining
  • Machine Learning with Big Data
  • Sparse Grids Methods

Talks

Teaching

Student Projects

Open

If you are interested in doing a student project, please contact me via e-mail and provide some ideas, what the project should be about.

Some topic ideas:

  • Parallelization of model refinement techniques for sparse grid classification
  • Block- and Rectilinear Adaptivity for Sparse Grid density Estimation

Running

  • K. Glas: Exploiting Component Grid Symmetries for Sparse Grid Density Estimation with the Combination Technique
    Bachelor's thesis, Fakultät für Informatik, Technische Universität München, since June 2019
  • V. Bautista Anguiano: Visualization of High Dimensional Models within the SG++ Datamining Pipeline
    Guided Research, Fakultät für Informatik, Technische Universität München, since April 2019
  • S. Weber: Exploiting the data hierarchy with geometry aware sparse grid for image classification
    Bachelor's thesis, Fakultät für Informatik, Technische Universität München, since April 2019
  • D. Boschko: Generalization and Parallelization of Sherman-Morrison System Matrix Updates for Sparse Grid Density Estimation
    IDP, Fakultät für Mathematik, Technische Universität München, since January 2018

Finished

About me