View Uppdrag

Uppdrag

Rubrik

Organisation/Företag

Plats

Uppdrag

Beskrivning

Theoretical and technological advances in machine learning have enabled in the last few years a revolution in many scientific and engineering domains. At the forefront of this revolution are neural networks, which have successfully been brought to bear on problems ranging from computer vision to natural language processing, from power systems to traffic networks. However, neural networks are a purely data-driven approach, as they do not account for the specific characteristics of the system generating the data. In particular, any insight on the system that we may have (for example the physics of a mechanical system) is disregarded in favor of the data only. To remedy this, the field of physics-informed learning (PIL) was developed, combining data-driven and physics-based approaches. This project then focuses on leveraging the PIL paradigm to design stabilizing controllers, especially for partially unmodeled systems. 


Objective: The roadmap of the project is as follows: 1) guided literature review, 2) design of physics-informed learning techniques to learn stabilizing controllers, 3) exploration of applications and numerical evaluations. The project has the potential to lead to conference and journal publications in top venues.

We are looking for 1-2 students to join this project. Requirements are: some background on machine learning, good programming skills (Python).

The start date is flexible, and the project will take around 20 weeks.

The project will be conducted in the Division of Decision and Control Systems, with flexible remote work, under the supervision of Asst. Prof. Matthieu Barreau and Dr. Nicola Bastianello.

Det finns inga poster att visa.

Du har inte behörighet att visa de här posterna.

Fel när begäran slutfördes.

Läser in ...

Sista ansökningsdatum

Publiceringsdatum

Omfattning

Applikationskanal

E-post för ansökan

Appens URL

Ansökningshandlingar

CV, academic transcripts, motivation letter (optional)