Publications and Presentations
A list of my publications, preprints and presentations.
Publications
1.
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]
2.
Leon Bungert, Franca Hoffmann, Dohyeon Kim, Tim Roith. MirrorCBO: A consensus-based optimization method in the spirit of mirror descent.
Mathematical Models and Methods in Applied Sciences (M3AS). (2025)
[Print]
[Preprint]
3.
T.J. Heeringa, T. Roith, C. Brune, M. Burger. Learning a Sparse Representation of Barron Functions with the Inverse Scale Space Flow.
Journal of Machine Learning Research (JMLR). (2025)
[Preprint]
4.
L. Weigand, T. Roith, M. Burger. Adversarial flows: A gradient flow characterization of adversarial attacks.
European Journal of Applied Mathematics. (2025)
[Print]
5.
T. Roith. Consistency, Robustness and Sparsity for Learning Algorithms.
PhD Thesis, Friedrich-Alexander-Universität Erlangen-Nürnberg. (2024)
[Print]
6.
R. Bailo, A. Barbaro, S. N Gomes, K. Riedl, T. Roith, C. Totzeck, U. Vaes. CBX: Python and Julia packages for consensus-based interacting particle methods.
Journal of Open Source Software (JOSS). (2024)
[Print]
[Preprint]
7.
S. Kabri, T. Roith, D. Tenbrinck, M. Burger. Resolution-invariant image classification based on Fourier neural operators.
Scale Space and Variational Methods in Computer Vision (SSVM 2023). (2023)
[Print]
[Preprint]
8.
L. Bungert, J.Calder, T. Roith. Uniform Convergence Rates for Lipschitz Learning on Graphs.
IMA Journal of Numerical Analysis. (2022)
[Print]
[Preprint]
9.
L. Bungert, T. Roith, P. Wacker. Polarized consensus-based dynamics for optimization and sampling.
Mathematical Programming. (2022)
[Print]
[Preprint]
10.
L. Bungert, R. Raab, T. Roith, L. Schwinn, D. Tenbrinck. CLIP: Cheap Lipschitz Training of Neural Networks.
Scale Space and Variational Methods in Computer Vision (SSVM 2021). (2021)
[Print]
[Preprint]
11.
L. Bungert, T. Roith, D. Tenbrinck, M. Burger. A Bregman Learning Framework for Sparse Neural Networks.
Journal of Machine Learning Research (JMLR). (2021)
[Print]
[Preprint]
12.
T. Roith, L. Bungert. Continuum Limit of Lipschitz Learning on Graphs.
Foundations of Computational Mathematics. (2020)
[Print]
[Preprint]
Preprints
1.
C. Fiedler, T. Roith. Consensus-based optimization for linearly separable functions.
(2026)
[Preprint]
2.
L. Paul, H. Rauhut, M. Burger, S. Kabri, T. Roith. Allure of Craquelure: A Variational-Generative Approach to Crack Detection in Paintings.
(2026)
[Preprint]
3.
A. Alcalde, L. Bungert, K. Riedl, T. Roith. Quantifying concentration phenomena of mean-field transformers in the low-temperature regime.
(2026)
[Preprint]
4.
S. Welker, L. Kuger, T. Roith, B. Feng, M. Burger, T. Gerkmann, H. Chapman. Position-Blind Ptychography: Viability of image reconstruction via data-driven variational inference.
(2025)
[Preprint]
5.
T. Roith, L. Bungert, P. Wacker. Consensus-based optimization for closed-box adversarial attacks and a connection to evolution strategies.
(2025)
[Preprint]
6.
L. Bungert, T. Roith, D. Tenbrinck, M. Burger. Neural Architecture Search via Bregman Iterations.
(2021)
[Preprint]
Presentations
1.
Workshop on the Mathematics of Transformers, University of Würzburg.
Oral Presentation.
Mean-field models for self-attention dynamics in transformers.
(2026)
2.
Joint Seminar of TUM, LMU and KU Ingolstadt.
Oral Presentation.
Mean-field models for self-attention dynamics in transformers.
(2026)
3.
Gradient Flows Face-to-Face 6 meets Transformers, Ulm University.
Oral Presentation.
Mean-field models for self-attention dynamics in transformers.
(2026)
4.
11th International Conference on Curves and Surfaces, Saint-Malo.
Oral Presentation.
MirrorCBO: A consensus-based optimization method in the spirit of mirror descent.
(2026)
5.
BIRS Workshop: Interacting Particle Systems, Hangzhou.
Online Talk.
Mean-field models for self-attention dynamics in transformers.
(2026)
6.
CMX Lunch Seminar, California Institute of Technology.
Oral Presentation.
The Mathematics of Adversarial Robustness.
(2024)
7.
Young applied mathematicians conference, Siena.
Oral Presentation.
Resolution-invariant image classification via FNOs.
(2023)
8.
Young applied mathematicians conference, Siena.
Oral Presentation.
Polarized consensus-based dynamics for optimization and sampling.
(2023)
9.
Digital Total.
Poster.
Computational Imaging@DESY.
(2023)
10.
BIRS Workshop: Leveraging Model- and Data-Driven Methods in Medical Imaging, Kelowna.
Oral Presentation.
A Bregman Learning Framework for Sparse Neural Networks.
(2023)
11.
MCQMC - International Conference on Monte Carlo and Quasi-Monte Carlo Methods.
Oral Presentation.
Kernelized Consensus Based Optimization.
(2022)
12.
SIAM - IS22 - Minisymposium Recent Advances on Stable Neural Networks.
Oral Presentation.
Stable Machine Learning via Lipschitz Methods.
(2022)
[YouTube]
13.
HCM - Workshop Synergies between Data Science and PDE Analysis.
Oral Presentation.
Uniform Convergence Rates for Lipschitz Learning.
(2022)
14.
GAMM - 92nd Annual Meeting.
Oral Presentation.
Uniform Convergence Rates for Lipschitz Learning.
(2022)
15.
ECOM - East Coast Optimization Meeting.
Oral Presentation.
A Bregman Learning Framework for Sparse Neural Networks.
(2022)
[Recording via GMU.edu]
16.
Conference on Calculus of Variation.
Oral Presentation.
Uniform Convergence Rates for Lipschitz Learning.
(2022)
17.
WWU Münster: Winterschool on Analysis and Applied Mathematics 2021.
Poster Session.
Continuum Limit of Lipschitz Learning on Graphs (Poster).
(2021)
[Poster]
18.
SSVM: International Conference on Scale Space and Variational Methods in Computer Vision.
Oral Presentation.
CLIP: Cheap Lipschitz Training of Neural Networks.
(2021)
[Slides (via unicloud.unicaen)]
19.
IMA Workshop: Theory and Algorithms in Graph-Based Learning.
Oral Presentation.
L-Infinity Variational Problems on Graphs: Applications and Continuum Limits.
(2020)
[YouTube, jointly with Leon Bungert]
