Stellenangebot

Postdoc Position at WIAS Berlin, intersection of robust optimization-learning-control

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). Those qualifications
are demonstrated, among others, by high-quality publications in credible
venues such as NeurIPS/ICLR/AISTATS/ICML/CoLT/JMLR/L4DC.

(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). Those
qualifications are demonstrated, among others by relevant publications
in credible venues such as SIAM OPT/CON/COAP/COCV/MathProg.

To learn more, follow this link: https://jj-zhu.github.io/postdoc_1/. Please
direct all inquiries to the email address doc.berlin.wias@gmail.com, with
the subject containing [inquiry-postdoc-mathplus].

Zurück zur Übersicht