Prof. Dr. Dirk Lorenz
Leader of AG Inverse Problems and Imaging(Photo Lukas Klose/Universität Bremen)
Office hours: During lecture time I hold weekly office hours (hybrid and in-person). You can find the dates on my profile page in Stud.IP - just navigate to my profile page, click on "Appointments", choose some date and show up either online (link in Stud.IP as well) or in my office.
CV
30.08.1978 | born in Bremen. |
1997 - 2002 | study of Mathematics, Physics and Philosophy at the University of Bremen |
08.2002 | Diploma in Mathematics (minor subject Physics), title "Methoden der Multiskalenglättung" |
2002 - 2004 | researcher in the group "Technomathematik" at the ZeTeM |
2004 - 2006 | postdoc in the graduate program SCiE |
02.2005 | PhD in Mathematics, supervisor: Prof. Peter Maass, 2nd referee: Prof. Otmar Scherzer, title "Wavelet Shrinkage in Signal & Image Processing - An Investigation of Relations and Equivalences" |
03.2006 - 08.2006 | postdoc at the Electrical Engineering department at Technion, Haifa funded by the research training network HASSIP. Host: Prof. Y.Y. Zeevi |
2006 - 03.2009 | postdoc in the group "Technomathematik" at the ZeTeM |
04.2009 - 09.2023 | Professor at the Institute for Analysis and Algebra, TU Braunschweig |
10.2023 - heute | Professor at the Center for Industrial Mathematics, Uni Bremen |
Research Areas
- Inverse problems
- Signal and image processing
- Optimization
- Machine Learning
Projects
- Sparsity and Compressed Sensing in Inverse Probleme (01.06.2008 - 31.05.2011)
- BMBF-INVERS: Dekonvolution vs. Shrinkage: Mathematische Methoden für eine verbesserte Peakdetektion (01.10.2007 - 30.06.2010)
- BMBF-INVERS: Regularisierung inverser Faltungsgleichungen in Besov-Skalen (01.10.2007 - 30.06.2010)
- Parameteroptimierung für die High-Content-Analyse (01.12.2005 - 30.09.2007)
- DFG-SPP 1114: Wavelet-shrinkage in der Bildverarbeitung – Eine Untersuchung von Zusammenhängen und Äquivalenzen (01.10.2002 - 30.09.2004)
Leader of Projects
- Automated data-driven damage detection (01.10.2023 - 30.09.2026)
- Training Data Driven Experts in Optimization (01.06.2020 - 01.12.2024)
- Mathematics for Machine Learning for Graph-Based Data with Integrated Domain Knowledge (01.04.2020 - 31.12.2023)
- Sparsity and Compressed Sensing in Inverse Probleme (01.06.2008 - 31.05.2011)
Courses (Selection)
- Mathematical Methods for Data Analysis and Image Processing (Wintersemester 2023/2024)
- Convex Analysis (Wintersemester 2023/2024)
- Seminar zur Regularisierung inverser Probleme (Wintersemester 2008/2009)
Theses (Selection)
- Die theoretische Analyse und Anwendung differentieller Methoden zur Flussberechnung (Kanglin Chen)
- Image Inpainting (Inna Korabova)
Publications (Selection)
- D. Lorenz, J. Marquardt, E. Naldi.
The degenerate variable metric proximal point algorithm and adaptive stepsizes for primal-dual Douglas-Rachford.
Erscheint in Optimizationonline at: https://arxiv.org/abs/2302.13128
- C. Brauer, N. Breustedt, T. de Wolff, D. Lorenz.
Learning variational models with unrolling and bilevel optimization.
Erscheint in Analysis and Applicationsonline at: https://arxiv.org/abs/2209.12651
- D. Lorenz, F. Schneppe, L. Tondji.
Linearly convergent adjoint free solution of least squares problems by random descent.
Inverse Problems, , 2023.DOI: 10.1088/1361-6420/ad08ed
online at: https://arxiv.org/abs/2306.01946 - L. Tondji, D. Lorenz, I. Necoara.
An accelerated randomized Bregman-Kaczmarz method for strongly convex linearly constraint optimization.
2023 European Control Conference (ECC).
- L. Tondji, I. Necoara, D. Lorenz.
Acceleration and restart for the randomized Bregman-Kaczmarz method.
Zur Veröffentlichung eingereicht.online at: https://arxiv.org/abs/2310.17338