Conference on
Mathematics of Machine Learning 2025

September 22 - 25, 2025

Hamburg University of Technology (TU Hamburg)

 

Poster Session 2

September 24, 2025

1. Oleksii Molodchyk (Hamburg University of Technology, Germany)
    Kernel Regression Meets System Identification

2. Johannes Müller (TU Berlin, Germany)
    Optimal Rates of Convergence for Entropy Regularization in Discounted Markov Decision Processes

3. Paul Navas (TU Dortmund, Germany)
    A Hilbert statistical manifold based on the Sinkhorn divergence

4. Elias Nyholm (Chalmers University of Technology, Sweden)
    Unifying Transformers and Convolutional Networks as Equivariant Maps

5. Guanru Pan (Hamburg University of Technology, Germany)
    Data-driven stochastic modeling and prediction with residual disturbance

6. Moritz Piening (TU Berlin, Germany)
    Slicing the Gaussian Mixture Wasserstein Distance

7. Konstantin Riedl (University of Oxford, United Kingdom)
      Convergence Analysis of Nonlinear Parabolic PDE Models with Neural Network Terms Trained with Gradient Descent

8. Frank Röder and Pradeep Banerjee (Hamburg University of Technology, Germany)
      Dynamics-Aligned Latent Imagination in Contextual World Models for Zero-Shot Generalization

9. Tim Roith (DESY, Germany)
      Gullible networks and the mathematics of adversarial attacks

10. Daniel Ruprecht (Hamburg University of Technology, Germany)
      Neural operators as coarse models for parallel-in-time integration methods

11. Melanie Schaller (Institut für Informationsverarbeitung, Germany)
      ModeConv: A Novel Convolution for Distinguishing Anomalous and Normal Structural Behavior

12. Darius Schaub (University Medical Center Hamburg-Eppendorf, Germany)
      2D Tissue reassembly with E(2)-equivariant graph neural cellular automata

13. Lucas Schmitt (Julius-Maximilians-Universität Würzburg, Germany)
      Adversarial Training as a Primal-Dual Problem

14. Alvaro Fernandez (Center for Free-Electron Laser Science, Germany)
      The coordinate is right -  Enhancing spectral approximations via normalizing flows

15. Sebastian Scott (University of Würzburg, Germany)
      On Optimal regularisation parameters via bilevel learning

16. Mariia Seleznova (LMU Munich, Germany)
      Neural Tangent Kernel Alignment as a Lens on Trained Neural Networks

17. Rishi Sonthalia (Boston College, United States)
      Identification of Mean-Field Dynamics using Transformers

18. Martin Spindler (University of Hamburg, Germany)
      DoubleMLDeep: Estimation of Causal Effects with Multimodal Data

19. Maximilian Steffen (Karlsruhe Institute of Technology, Germany)
      Statistical guarantees for stochastic Metropolis-Hastings

20. Florin Suciu (Université Paris Dauphine-PSL, France)
      Motion planning, rough controls, and deep neural networks

21. Mahsa Taheri (University of Hamburg, Germany)
      Regularization can make diffusion models more efficient

22. Alessandro Tamai (SISSA, Italy)
      Testing the Algebraicity Hypothesis

23. Johanna Tengler (University of Twente, Netherlands)
      Manifold limit for the training of shallow graph convolutional neural networks

24. Lucia Testa (University Medical Center Hamburg Eppendorf, Germany)
      Stability of graph convolutional networks under small random topology perturbations

25. Nicolás Valenzuela (University of Chile, Chile)
      Bounds on the approximation error for deep neural networks applied to dispersive models: nonlinear waves

26. Antonis Vasileiou (RWTH Aachen University, Germany)
      Covered Forest: Fine-grained generalization analysis of graph neural networks

27. Viktor Vigren Näslund (Umeå University, Sweden)
      Extending neural ODEs to encoder-decoder architectures via compressing vector fields

28. Lukas Weigand (Helmholtz Imaging, Germany)
      A gradient flow interpretation of Adversarial Attacks

29. Tak Ming Wong (Helmholtz-Zentrum Hereon, Germany)
      Self-supervised and physics-informed generative networks for the phase retrieval problem

30. Zusen Xu (Weierstrass Institute, Germany)
      Wasserstein Fisher-Rao Proximal Sampler

31. Hossein Zakerinia (Institute of Science and Technology Austria, Austria)
      Deep Multi-Task Learning Has Low Amortized Intrinsic Dimensionality

32. Giulio Zucal (Max Planck Institute of Molecular Cell Biology and Genetics, Germany)
      Graph and Hypergraph limits: new tools for Machine Learning

33. Alexander Klemps (Hamburg University of Technology, Germany)
      Generating Arbitrary Temporal Laser Pulse Shapes with Wasserstein Autoencoders

34. Oskar Nordenfors (Umeå University, Sweden)
      Neural Network Ensembles Provably Learn Symmetries

35. Filip Kovačević (Institute of Science and Technology Austria, Austria)
      Spectral Estimators for Multi-Index Models: Precise Asymptotics and Optimal Weak Recovery