High-throughput neutron scattering (SANS) instruments such as KWS-1 and KWS-2 measure 2D-SANS patterns at high rates. The obtained large number of data sets with often complex SANS-patterns are challenging to analyze. A detailed analysis is need to extract the maximum possible information about the investigated systems. This information leads to a fundamental understanding of materials properties and materials processing and function, often with in situ and operando experiments.

The project will develop machine-learning based algorithms and software that automatically recognizes and classifies large numbers of scattering patterns of the most common nano- and meso-scale structures. The project will deal with isotropic and anisotropic SANS- and also SAXS-patterns. After classification, the software will further automatically model the 1D- and 2D-data using GPUs for rapid calculations. The project will be conducted with the partner XENOCS, a company that produces state-of-the-art X-ray equipment where a large number of X-ray scattering patterns exist and can be analyzed.

Students will become familiar with the software development, GPU computing and machine learning tools for advanced data analysis, and develop a modern portfolio of skills in both academic and industrial settings.