Regensburg 2022 – wissenschaftliches Programm

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

MM: Fachverband Metall- und Materialphysik

MM 22: Data Driven Materials Science: Experimental Data Treatment and Machine Learning

MM 22.1: Topical Talk

Mittwoch, 7. September 2022, 10:15–10:45, H46

Ingredients for effective computer-augmented experimental materials science — •Christoph T. Koch, Markus Kühbach, Sherjeel Shabih, Benedikt Haas, and Sandor Bockhauser — Humboldt-Universität zu Berlin, Department of Physics & IRIS Adlershof, Berlin, Germany

Experimentally exploring the properties and uses of materials and improving them for particular purposes has been a major driving force for advancing the way people live over the last millennia. Experimental materials characterization techniques have now reached the level of detail that makes them converge with ab-initio computations based on fundamental building blocks: atoms and the electrons they share. During the last decades computers have surpassed the capacity of humans in the extraction of patterns in large amounts of data. It is thus a very natural consequence to involve their strengths also in further accelerating experimental materials science. In this talk we will use modern transmission electron microscopy as an example to illustrate current and future ways of how the process of linking the properties of materials to their fundamental structure can be supported computationally and by the availability of FAIR experimental and theoretical data sets. Along the way the contributions of the NFDI-project FAIRmat to this process will be highlighted, illustrating the importance of defining well-documented metadata catalogues, as well as providing community-specific online data processing capabilities.