Dr. Jean Rabault
Affiliation: Department of Mathematics, University of Oslo
Application of Deep Reinforcement Learning (DRL) to control non-linear, high-dimensional, time-dependent systems such as those found in Active Flow Control (AFC).
Recent works on the topic demonstrated several milestones, such as the first application of DRL to perform AFC of the Karman Vortex alley, the first demonstration of DRL potential for robust AFC over a range of flow conditions, and the first demonstration of AFC taking advantage of invariants of physical systems for reducing the cost of learning.
The next milestones, which are the aim of the COMPSCI-COFUND project presently submitted, will consist in applying DRL to industrial problems.
- French origin, settled in Norway, French / English / Norwegian / Swedish speaker.
- 2010 - 2015: BSc. and MSc. from Ecole Polytechnique, France (majors: physics, mechanics).
- 2014 - 2015: Double Degree student / Sivilingenjor from KTH, Sweden (major: fluid mechanics).
- 2015 - 2018: PhD from the Univ. of Oslo, wave-ice interaction in the polar regions, side projects on Machine Learning and Fluid Mechanics.
- falls 2017 and 2018: Guest lecturer, University Center in Svalbard.
- 2018 - present: PostDoc, Univ. of Oslo, dynamics of polar ice, and deep reinforcement learning applications to fluid mechanics.
- 2019 - present: Associate Researcher, Mines ParisTech.
Heavy focus on Open-Source, reproducible science:
- development of open source instruments for measurements of ice dynamics in the Arctic
- systematic open source and container release policy for works in Machine Learning.
- See list of publications at Google Scholar (Statistics from Google Scholar: citations=270; h-index=10)
- 21 peer-reviewed papers published between 2015 and 2020, a 10 as first author, including 2 in Journal of Fluid Mechanics, 1 in Journal of Geophysical Research, 2 in Physics of Fluids, 1 in Physical Review Letters.
- 12 presentations at international conferences, including 4 invited talks.
Supervisor for the following CompSci project
Mathematics and Statistics