Vincent Bürgin

20220320-16242750-square_01.jpg

Hi, I’m Vincent! I’m a PhD student in Machine Learning in Stefanie Jegelka’s group at the Technical University of Munich (TUM). I am currently working on weight space symmetries and loss landscapes properties of neural networks.

I was previously involved in a number of research projects during my M.Sc. Informatics at TUM (in the TUM DI-LAB at the CAMP chair, at the Theoretical Foundations of AI chair, at ImFusion, and at the Lab for AI in Medicine). Before that I completed joint B.Sc.s in Computer Science and Mathematics at the University of Passau.

news

Jul 08, 2026 We presented our paper Beyond Structural Symmetries: Linear Mode Connectivity via Neuron Identifiability at ICML 2026.
Jun 01, 2025 I started my PhD at Stefanie Jegelka’s Foundations of Deep Neural Networks group at TUM!
Oct 15, 2024 I published my post on the relAI blog: "Using topological features to prevent topological errors."

publications

  1. Beyond Structural Symmetries: Linear Mode Connectivity via Neuron Identifiability
    Vincent Bürgin*, Daniel Herbst*, Ya-Wei Eileen Lin, and Stefanie Jegelka
    In ICML 2026 Workshop on Weight-Space Symmetries: from Foundations to Practical Applications, 2026
  2. Topologically faithful multi-class segmentation in medical images
    Alexander H Berger, Nico Stucki, Laurin Lux, Vincent Bürgin, Suprosanna Shit, Anna Banaszak, Daniel Rueckert, Ulrich Bauer, and Johannes C. Paetzold
    Medical Image Computing and Computer Assisted Intervention (MICCAI), 2024
  3. Efficient Betti Matching Enables Topology-Aware 3D Segmentation via Persistent Homology
    Nico Stucki, Vincent Bürgin, Johannes C. Paetzold, and Ulrich Bauer
    arXiv preprint: arXiv:2407.04683, 2024
  4. Robust vertebra identification using simultaneous node and edge predicting Graph Neural Networks
    Vincent Bürgin, Raphael Prevost, and Marijn F. Stollenga
    In International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2023
  5. S3M: scalable statistical shape modeling through unsupervised correspondences
    Lennart Bastian*, Alexander Baumann*, Emily Hoppe, Vincent Bürgin, Ha Young Kim, Mahdi Saleh, Benjamin Busam, and Nassir Navab
    In International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2023
  6. On the Localization of Ultrasound Image Slices Within Point Distribution Models
    Lennart Bastian*, Vincent Bürgin*, Ha Young Kim*, Alexander Baumann, Benjamin Busam, Mahdi Saleh, and Nassir Navab
    In International Workshop on Shape in Medical Imaging, 2023
  7. Remarks on the tail order on moment sequences
    Vincent Bürgin, Jeremias Epperlein, and Fabian Wirth
    Journal of Mathematical Analysis and Applications, 2022