Dr. Pascal Fernsel

Research Areas

• Matrix Factorizations for Machine Learning
• Clustering
• Inverse Problems
• Imaging mass spectrometry

Projects

1. Nicht-negative Matrix-Faktorisierung mit A-priori-Wissen (since 01.12.2014)

Courses (Selection)complete list

1. Advanced Topics in Image Processing – The Beauty of Variational Calculus (Wintersemester 2024/2025)
2. Inverse Problems in Imaging (Wintersemester 2023/2024)
3. Deep Learning for Inverse Problems (Sommersemester 2023)
4. Nonlinear Inverse Problems (Sommersemester 2023)

Theses (Selection)complete list

1. Hauptkomponentenanalyse zur Untersuchung seismologischer Daten der Neumayer-Station III (Ribana Werner)

Publications (Selection)complete list

1. J. Gödeke, P. Fernsel.
   New universal operator approximation theorem for encoder-decoder architectures (Preprint).
   Zur Veröffentlichung eingereicht.

   online at: https://arxiv.org/abs/2503.24092

2. P. Fernsel, Z. Kereta, A. Denker.
   Convergence Properties of Score-Based Models using Graduated Optimisation for Linear Inverse Problems.
   2024 IEEE 34th International Workshop on Machine Learning for Signal Processing (MLSP), 22.09.-25.09.2024, London, UK.
   IEEE, pp. 1-6, 2024.

   DOI: 10.1109/MLSP58920.2024.10734770

3. P. Fernsel.
   Nonnegative Matrix Factorization: Theory, Algorithms and Applications.
   Dissertationsschrift, Universität Bremen, 2022.

   DOI: 10.26092/elib/1493
   online at: https://doi.org/10.26092/elib/1493

4. S. Arridge, P. Fernsel, A. Hauptmann.
   Joint Reconstruction and Low-Rank Decomposition for Dynamic Inverse Problems.
   Inverse Problems and Imaging, 16(3):483-523, 2022.

   DOI: 10.3934/ipi.2021059

5. P. Fernsel.
   Spatially Coherent Clustering Based on Orthogonal Nonnegative Matrix Factorization.
   MDPI Journal of Imaging, 7(10), 2021.

   DOI: 10.3390/jimaging7100194
   online at: https://www.mdpi.com/2313-433X/7/10/194