Thesis subject: Machine Learning for Neutron Reflectometry

The aim of the PhD project is to provide machine learning (ML) based neutron reflectometry (NR) analysis as an automatized workflow for the reflectometry instruments at the Institut Laue-Langevin (ILL). The current ML models are optimized mainly for (monochromatic) X-ray reflectometry. We aim to generalize this approach to a wide range of samples and time-of-flight NR, coupled with automatic data reduction, to fully automatize the workflow at the ILL reflectometers. The integration of the ML module in the data acquisition and data processing ecosystem will enable real-time data analysis and will open the possibility for closed-loop experiments, where analysis results are used for feedback control in an ongoing experiment. As sample systems polymer thin films and protein layers will be used to test the developed analysis tools.

You will join the Large Scale Structures (LSS) group at the ILL, Grenoble, France and work mainly with the two NR machines D17 and FIGARO for the implementation of the ML tools. A total of 2-6 months will be spent at Tübingen University in Germany, mainly at the beginning of the project, in order to adapt the existing ML tools to the needs of this project.