A. René Geist

I’m a postdoc working with Georg Martius at the University of Tübingen (until recently we’ve been located at the MPI-IS). My research focuses on data-driven dynamics simulators. Simulators help us to predict the motion of physical systems such as robots, folding proteins, weather, and energy networks over time. To unlock new applications in these domains, simulators have to be increasingly accurate and reliable. In my research, I look at how we can improve simulator performance by combining physics with machine learning.

I’ve been a postdoc at the RWTH Aachen with Sebastian Trimpe working on physics-informed neuralODEs. During my PhD supervised by Sebastian Trimpe at the Max Planck Institute for Intelligent Systems. I developed probabilistic models to learn implicitly-constrained robot dynamics. In collaboration with Jonathan Fiene, we developed the “Wheelbot” being a robot platform for teaching control theory and robotics.

During my Master’s in Theoretical Mechanical Engineering at TU Hamburg, I worked on end-to-end vehicle control as a research intern at the Vehicle Dynamics and Control laboratory of the University of California, Berkeley and as a research student on autonomous field exploration via AUV swarms working with Eugen Solowjow.

news

Jul 03, 2024	Our work “Learning deformable linear object dynamics from a single trajectory” is available on arxiv. The paper proposes a physics-informed NeuralODE to learn accurate deformable object dynamics by learning local errors in force space via neural networks.
Jun 30, 2024	“Pseudo-rigid body networks: learning interpretable deformable object dynamics from partial observations” has been accepted for presentation at IROS 2024.
May 02, 2024	“Learning with 3D rotations, a hitchhiker’s guide to SO(3)” has been accepted to ICML 2024!
May 15, 2023	I’ve joined Georg Martius’ Autonomous Learning group for a postdoc
Aug 19, 2022	“The Wheelbot: A Jumping Reaction Wheel Unicycle” has been nominated for an award at IROS 2022

selected publications

Learning deformable linear object dynamics from a single trajectory

Shamil Mamedov , A. René Geist, Ruan Viljoen , and 2 more authors

2024

Bib PDF

@misc{mamedov2024learningdeformablelinearobject,
  el = {true},
  title = {Learning deformable linear object dynamics from a single trajectory},
  author = {Mamedov, Shamil and Geist, A. René and Viljoen, Ruan and Trimpe, Sebastian and Swevers, Jan},
  year = {2024},
  eprint = {2407.03476},
  archiveprefix = {arXiv},
  primaryclass = {cs.RO},
  url = {https://arxiv.org/abs/2407.03476},
}

The Wheelbot: A jumping reaction wheel unicycle

A. René Geist, Jonathan Fiene , Naomi Tashiro , and 2 more authors

IEEE Robotics and Automation Letters, 2022

Bib PDF Video Website

@article{geist2022wheelbot,
  des = {true},
  author = {Geist, A. Ren\'e and Fiene, Jonathan and Tashiro, Naomi and Jia, Zheng and Trimpe, Sebastian},
  journal = {IEEE Robotics and Automation Letters},
  title = {The Wheelbot: A jumping reaction wheel unicycle},
  year = {2022},
  number = {4},
  pages = {9683-9690},
  volume = {7},
  doi = {10.1109/LRA.2022.3192654},
  url = {https://ieeexplore.ieee.org/abstract/document/9834094},
}

Learning constrained dynamics with Gauss principle adhering Gaussian processes

A. René Geist, and Sebastian Trimpe

In Proceedings of the 2nd Conference on Learning for Dynamics and Control , Jun 2020

Bib PDF

@inproceedings{geist2020gauss,
  dyn = {true},
  author = {Geist, A. Ren\'e and Trimpe, Sebastian},
  booktitle = {Proceedings of the 2nd Conference on Learning for Dynamics and Control},
  title = {Learning constrained dynamics with Gauss principle adhering Gaussian processes},
  year = {2020},
  month = jun,
  pages = {225--234},
  publisher = {PMLR},
  series = {Proceedings of Machine Learning Research (PMLR)},
  volume = {120},
  editors = {Bayen, Alexandre M. and Jadbabaie, Ali and Pappas, George and Parrilo, Pablo A. and Recht, Benjamin and Tomlin, Claire and Zeilinger, Melanie},
  url = {https://proceedings.mlr.press/v120/geist20a.html},
}

Learning with 3D rotations, a hitchhiker’s guide to SO(3)

Andreas René Geist , Jonas Frey , Mikel Zhobro , and 2 more authors

21–27 jul 2024

Bib PDF PDF

@misc{geist2024learning,
  bas = {true},
  title = {Learning with 3{D} rotations, a hitchhiker’s guide to {SO}(3)},
  author = {Geist, Andreas Ren\'{e} and Frey, Jonas and Zhobro, Mikel and Levina, Anna and Martius, Georg},
  booktitle = {Proceedings of the 41st International Conference on Machine Learning},
  pages = {15331--15350},
  year = {2024},
  editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix},
  volume = {235},
  series = {Proceedings of Machine Learning Research},
  month = {21--27 Jul},
  publisher = {PMLR},
  url = {https://proceedings.mlr.press/v235/geist24a.html},
}