The Berlin Mathematics Research Center MATH+ sets out to advance mathematics itself and its interdisciplinary power with the aim of achieving progress on grand challenges in a wide variety of application fields.
MATH+ will start with nine major units for project-oriented research: four Application Areas and five Emerging Fields. In both, mathematicians from a wide range of different disciplines collaborate – with each other and with leading researchers from diverse application fields as well as representatives from industry. These Research Units are complemented by Transfer Units designed for translational research. For further details please visit www.mathplus.de.
Jobs description:
In the frame of the Cluster of Excellence, MATH+ project EF3-4 “Physics-regularized learning” is looking for a research assistant. This methodologically oriented project is devoted to the development and numerical analysis of learning algorithms for (typically large) data sets exploiting a priori knowledge in terms of an auxiliary physical model formulated as partial differential equation (PDEs). Within the advertised position the applicant will work on the development and numerical analysis of neural network discretizations for learning problems and PDEs and on the coupling of both. To this end the project will combine techniques from PDE numerics, nonlinear approximation theory, and statistical mechanics. The position also includes the implementation of the developed methods and their application to problems from microscopy. The project is headed by Prof. Dr. R. Kornhuber (FU Berlin), Prof. Dr. C. Gräser (FU Berlin), and Prof. Dr. C. Schütte (FU Berlin / Zuse Institute Berlin).
Requirements:
MSc. in mathematics or equivalent.
Desirable:
The successful applicant has pre-knowledge on either PDE numerics, applied statistical mechanics, or stochastic differential equations and some programing experience. Pre-knowledge on machine learning and neural networks is helpful but not required.