Hi, I’m Tim!

I’m currently a substitute professor at the Technical University of Munich. I am a member of the COST Action InterCoML and the Munich Center for Machine Learning.

My research lies at the intersection of mathematical analysis, optimization, and machine learning. I work on consensus-based optimization methods, robustness and stability of neural networks, and learning problems in imaging and inverse problems.

Recent Posts

Consensus-based optimization for linearly separable functions

Together with Christian Fiedler we posted a new preprint on Consensus-based optimization for linearly separable functions. We propose a reparametrization strategy that provably finds coordinate transformations in which the objective becomes separable, which...

Recent Publications

C. Fiedler, T. Roith. Consensus-based optimization for linearly separable functions. (2026) [Preprint]
L. Paul, H. Rauhut, M. Burger, S. Kabri, T. Roith. Allure of Craquelure: A Variational-Generative Approach to Crack Detection in Paintings. (2026) [Preprint]
Martin Burger, Samira Kabri, Yury Korolev, Tim Roith and Lukas Weigand. Analysis of mean-field models arising from self-attention dynamics in transformer architectures with layer normalization. Philosophical Transactions of the Royal Society A. (2025) [Print] [Preprint]

Recent Presentations

Young applied mathematicians conference, Siena. Oral Presentation. Resolution-invariant image classification via FNOs. (2023)
Young applied mathematicians conference, Siena. Oral Presentation. Polarized consensus-based dynamics for optimization and sampling. (2023)
Digital Total. Poster. Computational Imaging@DESY. (2023)

Projects

LaTeX Projects

During my time at university, LaTeX was constantly following me around and helping me out. I tried to make everything easier by producing some templates that can be reused by me and others for various projects.

Research

Here’s an overview of some of my research projects and interests.

Teaching

Here is an overview over teaching material that I published online.