Dr. Alexander Kumpf

Visualization of scientific data with a focus on meteorological ensemble data. Using machine learning approaches such as clustering or correlation analysis, feature detection and tracking, relevant properties of the underlying data are visualized. To achieve interactive performance, the methods are implemented in C++ or Python. Deep neural network approaches (mainly Endocder-Decoder and Generative Adversarial architectures) are used for rendering completion and context aware data interpolation.

Visualization of scientific data | Performant interactive systemsClustering | Feature detection | Tracking of structures Correlation analysis | Numerical weather prediction Deep neural networks for data enhancement Supervision of student workers, Master’s and Bachelor’s theses

Optimization, statistics, finance and insurance Master thesis: “An algorithmic study of the reverse Loomis-Whitney inequality”

Marchov Chain Monte Carlo, Computer Vision

Focus on Analysis and Optimization, Economics, Marketing and Accounting