Berlin 2024 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

CPP: Fachverband Chemische Physik und Polymerphysik

CPP 34: Emerging Topics in Chemical and Polymer Physics, New Instruments and Methods I

CPP 34.2: Vortrag

Donnerstag, 21. März 2024, 10:00–10:15, H 0106

Deep learning based reflectometry data analysis including prior knowledge — •Alexander Hinderhofer, Valentin Munteanu, Vladimir Starostin, Linus Pithan, Alexander Gerlach, and Frank Schreiber — Institut für Angewandte Physik, Universität Tübingen, Auf der Morgenstelle 10, 72076 Tübingen, Germany

Current machine-learning solutions for automatized analysis of X-ray (XRR) and neutron reflectivity (NR) data is constrained by the range and number of considered parameters, making the approach inflexible for applying it to different material and layer configurations. To overcome this, we present an approach that utilizes prior knowledge to regularize the training process over larger parameter spaces. We demonstrate the effectiveness of our method in various scenarios, including multilayer structures with box model parametrization and a physics-inspired special parametrization of the scattering length density profile for a multilayer structure. In contrast to previous methods, our approach scales favorably when increasing the complexity of the inverse problem, working properly even for a several layer multilayer model and an N-layer periodic multilayer model with up to 20 open parameters. We will also discuss autonomous experiments enabled by machine-learning-based online data analysis in synchrotron beamline environments. [1]

[1] L. Pithan et al. J. Synchrotron Rad. 30 (2023) 1064

Keywords: thin films; machine learning; X-ray scattering; data analysis