We develop and apply machine learning/AI methods with the objective of extracting physically and chemically relevant structural and dynamic information from neutron scattering data. Our approach incorporates a diverse range of techniques, including classical machine learning approaches such as Gaussian processes as well as diffusion and large language models. Our research supports JCNS scientists and users at all steps of their scientific workflow: from data acquisition through data reduction to data analysis.

Our second activity area encompasses the development and maintenance of data reduction software for the time-of-flight (ToF), SANS (small-angle neutron scattering), and neutron diffraction instruments. In addition to software and method development, we are engaged in the dissemination of knowledge regarding artificial intelligence and machine learning methods to facility scientists and users.