Humboldt-Universität zu Berlin - Faculty of Mathematics and Natural Sciences - Databases and Information Systems

Hermann Stolte

 
Hermann Stolte
© ECDF/PR/Felix Noak

I am a Ph.D. student at the Helmholtz Einstein International Berlin Research School in Data Science (HEIBRiDS), working in the Database and Information Systems group under co-supervision of Prof. Matthias Weidlich (HU Berlin) and Dr. Elisa Pueschel (DESY).

My research interests include, among others:

  • Exploratory data analysis, e.g. applied {unsupervised} machine learning, data visualization, data quality issues such as sparsity, bias, noise, and uncertainty
  • Testing and debugging data pipelines, e.g. root cause analysis, data provenance
  • Sound and music computing, audio technology, sonification

Currently, I conduct interdisciplinary research in the context of the Helmholtz-funded project Dynamic Scheduling of Gamma-ray Source Observations. In this project, we aim to increase the observational coverage of transient and variable gamma-ray phenomena. To this end, we develop a data-driven method for detecting very-high-energy blazar flares with deep learning. This novel approach will enable a dynamic optimization of observation schedules in near real-time. Moreover, we develop methods to support working with complex data processing pipelines through automated plausibility assessment, focusing on both uncertain pipeline inputs and analysis results.

 

Profiles and Bibliography