Muhammad Zakwan

Muhammad Zakwan
Muhammad Zakwan
PhD Student
Realizing mathematical constructions to tangible real-world applications for a better future.

Muhammad Zakwan is a doctoral assistant in the Dependable Control and Decision group (DECODE) at École Polytechnique Fédérale de Lausanne (EPFL). He is also a member of the National Centre of Competence in Research (NCCR) Automation. Zakwan holds an Electrical and Electronics  Engineering degree from Bilkent University (Turkey) and a Bachelor of Science in Electrical Engineering from the Pakistan Institute of Engineering and Applied Sciences (Pakistan). His research interests include LPV systems, hybrid dynamical systems, time-delay systems, distributed systems, and machine learning.

Publications scientifiques

Published
Stable Linear Subspace Identification: A Machine Learning Approach
ArXiv
Published
Universal Approximation Property of Hamiltonian Deep Neural Networks
IEEE Control System Letters
Vol 7 Pages 2689-2694
Published
Physically Consistent Neural ODEs for Learning Multi-Physics Systems
Proceedings of IFAC World congress 2023
Published
Robust Classification Using Contractive Hamiltonian Neural ODEs
IEEE Control Systems Letters
Vol 7 Pages 145-150
Published
Distributed Neural Network Control with Dependability Guarantees: a Compositional Port-Hamiltonian Approach
Proceedings of The 4th Annual Learning for Dynamics and Control Conference
Vol 168 Pages 571-583
Published
Neural Energy Casimir Control for Port-Hamiltonian Systems
arXiv

Projets de recherche comme Chercheur

Titre
Chercheurs principaux

ODE networks for data-based control

Résumé

In spite of several success stories, standard deep networks  can lead to unstable forward propagation or suffer from an ill-posed learning process, which might result in vulnerability to adversarial attacks. In this project we will study deep network architectures based on ODEs for mitigating these problems. Furthermore, we will investigate the use of these networks for parametrizing regulators that can be tuned from available data. The new methods will be tested for the control of inverters, smart loads and microgrid clusters, which are relevant for the energy systems considered in the NCCR.

ODE networks for data-based control

In spite of several success stories, standard deep networks  can lead to unstable forward propagation or suffer from an ill-posed learning process, which might result in vulnerability to adversarial attacks. In this project we will study deep network architectures based on ODEs for mitigating these problems. Furthermore, we will investigate the use of these networks for parametrizing regulators that can be tuned from available data. The new methods will be tested for the control of inverters, smart loads and microgrid clusters, which are relevant for the energy systems considered in the NCCR.

120
bba0c33c-021f-4dbb-848c-3b2f57e14a0b