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

MM: Fachverband Metall- und Materialphysik

MM 31: Data Driven Materials Science: Big Data and Work Flows – Machine Learning

MM 31.4: Vortrag

Mittwoch, 29. März 2023, 16:30–16:45, SCH A 251

Exploring materials dataspaces by combining supervised and unsupervised machine learning — •Andreas Leitherer¹, Angelo Ziletti¹, Christian H. Liebscher², Timofey Frolov³, and Luca M. Ghiringhelli^{1, 4} — ¹NOMAD Laboratory at the FHI of the Max-Planck-Gesellschaft and IRIS-Adlershof of the Humboldt-Universität (HU) zu Berlin — ²Max-Planck-Institut für Eisenforschung — ³Lawrence Livermore National Laboratory — ⁴Physics Department and IRIS-Adlershof of HU zu Berlin

To enable meaningful applications of AI to materials science, much of current efforts is concentrated on the creation of characterized datasets. In this talk, we discuss a rarely addressed topic - the development of automatic tools to explore available materials-science data. In particular, we go beyond purely supervised learning by combining unsupervised analysis with a recently developed crystal-structure recognition method [1]. This neural-network (NN) model automatically learns data representations that contain information on structurally diverse geometries. Using clustering, physically meaningful subgroups can be identified in the NN latent space, which are shown, e.g., to correspond to distinct, experimentally verified grain-boundary phases [2]. Moreover, dimension-reduction analysis allows us to create low-dimensional, interpretable materials charts that visualize complex structural data from both theoretical and experimental origin.

[1] A. Leitherer, A. Ziletti and L. M. Ghiringhelli. Nat. Commun. 12, 6234 (2021)

[2] T. Meiners et al. Nature 579, 375-378 (2020)