Michael Muehlebach

I am leading the independent research group Learning and Dynamical Systems at the Max Planck Institute for Intelligent Systems in Tuebingen, Germany. 

I studied mechanical engineering at ETH Zurich and specialized in robotics, systems, and control during my Master's degree. I received the B.Sc. and the M.Sc. in 2010 and 2013, respectively, before joining the Institute for Dynamic Systems and Control for my Ph.D. I graduated under the supervision of Prof. R. D'Andrea in 2018 and went on to join the group of Prof. Michael I. Jordan at the University of California, Berkeley as a postdoctoral researcher.

I am interested in a wide variety of subjects, including machine learning, dynamics, control, and optimization. During my Ph.D. I worked on approximations of the constrained linear quadratic regulator problem with applications to model predictive control (see here). I also designed control, estimation, and learning algorithms for a balancing robot and a flying machine. As a postdoctoral researcher at Berkeley, I analyzed first-order optimization algorithms from a dynamical system's point of view (see here).

I received the Outstanding D-MAVT Bachelor Award, the Willi-Studer prize for the best Master's degree, and the ETH Medal and the HILTI prize for my doctoral thesis. I am a Branco Weiss Fellow since 2018, was awarded the Emmy Noether Fellowship in 2020, and an Amazon Fellowship in 2024.

I am actively looking for talented and motivated PhD or Master's students. More information can be found on the group website.

Contact

Adress: Max Planck Ring 4, 72076 Tuebingen, Germany

E-Mail: michaelm@tuebingen.mpg.de

Publications

Preprints

M. Muehlebach, Z. He, and M. I. Jordan, "The Sample Complexity of Online Reinforcement Learning: A Multi-Model Perspective", https://arxiv.org/abs/2501.15910, 2025

J. Zughaibi, B. J. Nelson, and M. Muehlebach, "Dynamic Electromagnetic Navigation", https://arxiv.org/abs/2402.06012, 2025

M. Bal, V. Cevher, and M. Muehlebach, "Adversarial Training for Defense Against Label Poisoning Attacks", International Conference on Learning Representations, 2025, https://openreview.net/forum?id=UlpkHciYQP (accepted)

K.-R. Kladny, B. Schölkopf, and M. Muehlebach, "Conformal Generative Modeling with Improved Sample Efficiency Through Sequential Greedy Filtering", International Conference on Learning Representations, 2025, https://arxiv.org/abs/2410.01660 (accepted)

M. Cummins, D. Er, and M. Muehlebach, "Controlling Participation in Federated Learning with Feedback", https://arxiv.org/abs/2411.19242, 2024

D. Er, S. Trimpe, and M. Muehlebach, "Distributed Event-based Learning via ADMM", https://arxiv.org/abs/2405.10618, 2024

Z. He, S. Bolognani, M. Muehlebach, and F. Dörfler, "Grey-Box Nonlinear Feedback Optimization", https://arxiv.org/abs/2404.04355, 2024

L. Zhang, N. He, and M. Muehlebach, "Primal Methods for Variational Inequality Problems with Functional Constraints", https://arxiv.org/abs/2403.12859 , 2024

M. Muehlebach, and M. I. Jordan, "Accelerated First-Order Optimization under Nonlinear Constraints", https://arxiv.org/abs/2302.00316, 2024


Journal Publications

A. Ibrahim, M. Muehlebach, and C. De Bacco, "Optimal transport with constraints: from mirror descent to classical mechanics", Physical Review Letters, 2024, https://arxiv.org/abs/2309.04727

K.-R. Kladny, J. von Kügelgen, B. Schölkopf, and M. Muehlebach, "Deep Backtracking Counterfactuals for Causally Compliant Explanations", Transactions on Machine Learning Research, 2024, https://arxiv.org/abs/2310.07665 

F. Dörfler, Z. He, G. Belgioioso, S. Bolognani, J. Lygeros, and M. Muehlebach, "Towards a Systems Theory of Algorithms", IEEE Control Systems Letters, 2024, https://arxiv.org/abs/2401.14029 

H. Ma, D. Büchler, B. Schölkopf, and M. Muehlebach, "Reinforcement Learning with Model-Based Feedforward Inputs for Robotic Table Tennis", Autonomous Robots, 2023, link

M. Hofer, M. Muehlebach, and R. D'Andrea, "The One-Wheel Cubli: A 3D inverted pendulum that can balance with a single reaction wheel", Mechatronics, 2023, link, video 

M. Muehlebach and M. I. Jordan, "On Constraints in First-Order Optimization: A View from Non-Smooth Dynamical Systems", Journal of Machine Learning Research, 2022, https://arxiv.org/abs/2107.08225

M. Muehlebach and M. I. Jordan, "Optimization with Momentum: Dynamical, Control-Theoretic, and Symplectic Perspectives", Journal of Machine Learning Research, 2021, https://arxiv.org/abs/2002.12493

C. Sferrazza, M. Muehlebach, and R. D'Andrea, "Learning-based Parametrized Model Predictive Control for Trajectory Tracking", Optimal Control: Applications and Methods, 2020, https://onlinelibrary.wiley.com/doi/full/10.1002/oca.2656

M. Muehlebach and R. D'Andrea, "A Method for Reducing the Complexity of Model Predictive Control in Robotics Applications", IEEE Robotics and Automation Letters, 2019, https://arxiv.org/abs/1903.07648

M. Muehlebach and R. D'Andrea, "Accelerometer-Based Tilt Determination for Rigid Bodies with a Non-Accelerated Pivot Point", IEEE Transactions on Control Systems Technology, 2018

M. Muehlebach and S. Trimpe, "Distributed Event-Based State Estimation for Networked Systems: An LMI-Approach", IEEE Transactions on Automatic Control, 2017

M. Muehlebach and R. D'Andrea, "The Flying Platform - A Testbed for Ducted Fan Actuation and Control Design", Mechatronics, 2017

M. Muehlebach and R. D'Andrea, "Nonlinear Analysis and Control of a Reaction Wheel-based 3-D Inverted Pendulum", IEEE Transactions on Control Systems Technology, 2016

M. Muehlebach, T. Heimsch, and Ch. Glocker, "Variational Integrators - A Continuous Time Approach", International Journal for Numerical Methods in Engineering, 2016

H. Maes, G. Vandersteen, M. Muehlebach, and C. Ionescu, "A Fan-based Low-frequent Forced Oscillation Technique Apparatus", IEEE Transactions on Instrumentation and Measurements, 2014


Conference Publications

W. Zhao, Y. Zhao, J. Pajarinen, and M. Muehlebach, "Bi-level Motion Imitation for Humanoid Robots", Conference on Robot Learning, 2024, https://arxiv.org/abs/2410.01968 

P. Fischer, H. Willms, M. Schneider, D. Thorwarth, M. Muehlebach, and C. Baumgartner, "Subgroup-Specific Risk-Controlled Dose Estimation in Radiotherapy", International Conference on Medical Image Computing and Computer-Assisted Intervention, 2024, https://arxiv.org/abs/2407.08432 

H. Ma, M. Zeilinger, and M. Muehlebach, "Online Optimization of Closed-Loop Control Systems", Workshop on Foundations of Reinforcement Learning and Control, International Conference on Machine Learning, 2024, https://arxiv.org/abs/2404.05318 

O. Eberhard, C. Vernade, and M. Muehlebach, "A Pontryagin Perspective on Reinforcement Learning", Workshop on Foundations of Reinforcement Learning and Control, International Conference on Machine Learning, 2024, https://arxiv.org/abs/2405.18100 

D. Er and M. Muehlebach, "Event-Based Federated Q-Learning",  Workshop on Foundations of Reinforcement Learning and Control, International Conference on Machine Learning, 2024, link 

Z. He, M. Muehlebach, S. Bolognani, and F. Dörfler, "Online Performance Optimization of Nonlinear Systems: A Gray-Box Approach", Workshop on Foundations of Reinforcement Learning and Control, International Conference on Machine Learning, 2024, link 

A. Wundram, P. Fischer, M. Muehlebach, L. Koch, and C. Baumgartner, "Conformal Performance Range Prediction for Segmentation Output Quality Control", International Workshop on Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, 2024, https://arxiv.org/abs/2407.13307  - Best Paper Award!

K.-R. Kladny, J. von Kügelgen, B. Schölkopf, and M. Muehlebach, "Backtracking Counterfactuals for Deep Structural Causal Models", Causal Inference Workshop at the Conference on Uncertainty in Artificial Intelligence, 2024, https://openreview.net/forum?id=5dCVMpqcKp 

M. Muehlebach and M. I. Jordan, "Smooth and Non-Smooth Dynamics Perspectives on Accelerated Optimization", European Nonlinear Dynamics Conference, 2024

S. Guist, J. Schneider, H. Ma, L. Chen, V. Berenz, J. Martus, H. Ott, F. Grüninger, M. Muehlebach, J. Fiene, B. Schölkopf, and D. Büchler, "Safe and Accurate at Speed with Tendons: A Robot Arm for Exploring Dynamic Motion", Robotics: Science and Systems, 2024, https://arxiv.org/abs/2307.02654 

P. Kolev, G. Martius, M. Muehlebach, "Online Learning under Adversarial Nonlinear Constraints", Conference on Neural Information Processing Systems, 2023, https://arxiv.org/abs/2306.03655 

P. Tobuschat, H. Ma, D. Büchler, B. Schölkopf, M. Muehlebach, "Data-Efficient Online Learning of Ball Placement in Robot Table Tennis", IEEE/RSJ International Conference on Intelligent Robots and Systems, 2023, https://arxiv.org/abs/2308.14562 

K. Kladny, J. von Kügelgen, B. Schölkopf, and M. Muehlebach, "Causal Effect Estimation from Observational and Interventional Data Through Matrix Weighted Linear Estimators", Conference on Uncertainty Quantification in Artificial Intelligence, 2023, https://arxiv.org/abs/2306.06002 

M. Muehlebach, "Adaptive Decision-Making with Constraints and Dependent Losses: Performance Guarantees and Applications to Online and Nonlinear Identification", IFAC World Congress, 2023, https://arxiv.org/abs/2304.03321 

G. Tong and M. Muehlebach, "A Dynamical Systems Perspective on Discrete Optimization", Proceedings of Machine Learning Research, 2023, https://arxiv.org/abs/2305.08536 

J. Achterhold, P. Tobuschat, H. Ma, D. Buechler, M. Muehlebach, J. Stueckler, "Black-Box vs. Gray-Box: A Case Study on Learning Table Tennis Ball Trajectory Prediction with Spin and Impacts", Proceedings of Machine Learning Research, 2023, https://arxiv.org/abs/2305.15189 

S. Schechtman, D. Tiapkin, M. Muehlebach, and E. Moulines, "Orthogonal Directions Constrained Gradient Method: from non-linear equality constraints to Stiefel manifold", Conference on Learning Theory, 2023, https://arxiv.org/abs/2303.09261 

A. Das, B. Schölkopf, and M. Muehlebach, "Sampling Without Replacement Leads to Faster Rates in Finite-sum Minimax Optimization", Conference on Neural Information Processing Systems, 2022, https://arxiv.org/abs/2206.02953 

H. Ma, D. Büchler, B. Schölkopf, and M. Muehlebach, "A Learning-based Iterative Control Framework for Controlling a Robot Arm with Pneumatic Artificial Muscles", Robotics: Science and Systems, 2022, http://www.roboticsproceedings.org/rss18/p029.html 

D. Schechtman, D. Tiapkin, E. Moulines, M. I. Jordan, and M. Muehlebach, "First-order Constrained Optimization: Non-smooth Dynamical System Viewpoint", IFAC Workshop on Control Applications of Optimization, 2022

N. S. Wadia, M. I. Jordan, and M. Muehlebach, "Optimization with Adaptive Step Size Selection from a Dynamical Systems Perspective", OPT2021 Workshop, Conference on Neural Information Processing Systems, 2021, https://opt-ml.org/papers/2021/paper28.pdf

M. Muehlebach and M. I. Jordan, "Continuous-time Lower Bounds for Gradient-based Algorithms", Proceedings of the International Conference on Machine Learning, 2020, https://arxiv.org/abs/2002.03546

M. Muehlebach and M. I. Jordan, "A Dynamical Systems Perspective on Nesterov Acceleration", Proceedings of the International Conference on Machine Learning, 2019, https://arxiv.org/abs/1905.07436

N. B. Erichson, M. Muehlebach, and M. Mahoney, "Physics-informed Autoencoders for Lyapunov-stable Fluid Flow Prediction", Machine Learning and the Physical Sciences Workshop, Conference on Neural Information Processing Systems, 2019,  https://arxiv.org/abs/1905.10866 

M. Muehlebach and R. D'Andrea,  "Basis Functions Design for the Approximation of Constrained Linear Quadratic Regulator Problems Encountered in Model Predictive Control", Proceedings of the International Conference on Decision and Control, 2017

C. Sferrazza, M. Muehlebach, and R. D'Andrea, "Trajectory Tracking and Iterative Learning on an Unmanned Aerial Vehicle using Parametrized Model Predictive Control", Proceedings of the International Conference on Decision and Control, 2017

M. Muehlebach, C. Sferrazza, and R. D'Andrea, "Implementation of a Parametrized Infinite-Horizon Model Predictive Control Scheme with Stability Guarantees", Proceedings of the International Conference on Robotics and Automation, 2017

M. Muehlebach and R. D'Andrea, "Approximation of Continuous-Time Infinite-Horizon Optimal Control Problems Arising in Model Predictive Control", Proceedings of the International Conference on Decision and Control, 2016

M. Muehlebach and R. D'Andrea, "Parametrized Infinite-horizon Model Predictive Control for Linear Time-invariant Systems with Input and State Constraints", Proceedings of the American Control Conference, 2016

M. Hofer, M. Muehlebach, and R. D'Andrea, "Application of an Approximate Model Predictive Control Scheme on an Unmanned Aerial Vehicle", Proceedings of the Conference on Robotics and Automation, 2016

M. Muehlebach and S. Trimpe, "LMI-based Synthesis for Distributed Event-based State Estimation", Proceedings of the American Control Conference, 2015

M. Muehlebach and S. Trimpe, "Guaranteed H2 Performance in Distributed Event-based State Estimation", Proceedings of the Conference on Event-based Control, Communication, and Signal Processing, 2015

M. Muehlebach, Gajamohan M., and R. D'Andrea, "Nonlinear Analysis and Control of a Reaction Wheel-based 3D Inverted Pendulum", Proceedings of the International Conference on Decision and Control, 2013

M. Gajamohan, M. Muehlebach, T. Widmer, and R. D'Andrea, "The Cubli: A Reaction Wheel-based 3D Inverted Pendulum", Proceedings of the European Control Conference, 2013


Technical Reports

M. Muehlebach, "The Silver Ratio and its Relation to Controllability", 2019, https://arxiv.org/abs/1908.07109 

M. Muehlebach and R. D'Andrea, "On the Approximation of Constrained Linear Quadratic Regulator Problems and their Application to Model Predictive Control", 2018, https://doi.org/10.3929/ethz-b-000292793 

Videos from past projects

We developed a data-efficient learning method for controlling a robot arm and engaging in playful activities such as ping-pong. The robot arm is actuated with pneumatic artificial muscles. This is joint work with H. Ma, D. Büchler, and B. Schölkopf. More details can be found here: http://www.roboticsproceedings.org/rss18/p029.html .

The One-Wheel Cubli is a three-dimensional pendulum system, that can balance on its pivot using a single reaction wheel. This is an extremely challenging task that requires stabilizing about ten degrees of freedom, many of which are unstable or marginally stable, with a single control input. After more than five years of research, M. Hofer, R. D'Andrea and I finally managed to realize the project in hardware.

The Flying Platform was designed to study ducted fan actuation. It was also used for benchmarking novel control strategies that account for actuation limits. Control algorithms explicitly accounting for these limitation can provide larger stability margins and other performance enhancements.

I supervised Julien Kohler's Master thesis, where we designed control, estimation, and learning algorithms for aggressive quadrotor maneuvers.

The Cubli is a balancing robot that can balance on its corner and jump up. I investigated the dynamics, and implemented and tested a nonlinear controller. I also designed the learning algorithm that enables the system to adapt to a changing environment.