The Weierstrass Institute (WIAS) in Berlin, Germany, is looking for a postdoc candidate passionate about conducting research in the intersection of mathematical optimization, machine learning, optimal control with a focus on robustness under distribution shift. The position is associated with a third-party funded research project led by Dr. Jia-Jie Zhu (WIAS Berlin) and Dr. Michael Hintermüller (WIAS/ Humboldt-Universität zu Berlin). The initial funding period will run for two years, starting January 1st, 2022. The applicants should have completed their Ph.D. degrees by the starting date of the project.Subject and candidate qualification
Motivated by numerous partial differential equation (PDE) related practical applications, a pressing challenge for data-driven optimization and control systems is the ubiquitous distribution shift, which implies higher demand for the robustness of the learning-for-control design. The project, funded by the Excellence Cluster MATH+: The Berlin Mathematics Research Center, aims to address data-driven robust control and optimization of dynamical systems under data distribution shifts, using principled tools from applied mathematics and statistical machine learning as well as reinforcement learning. We invite candidates whose scholar profiles are mainly theoretical and exhibit proven excellence in research. We specifically prefer two types of mathematical research experiences:
(1) either in principled statistical machine learning/reinforcement learning theory (related to dynamical systems, time series, Markov decision process (MDP), principled control theory).
(2) and/or in applied mathematics, in optimization, numerical analysis, optimal control, dynamical systems (PDEs and S(P)DEs), data-driven modeling of dynamics, theory of model predictive control (MPC).
To apply or learn more
Follow this link. Please direct all inquiries to the email address doc.ber...@gmail.com, with the subject containing [inquiry-postdoc-mathplus]. For more information about WIAS Berlin, see https://www.wias-berlin.de/.