Dresden 2020 – wissenschaftliches Programm

Die DPG-Frühjahrstagung in Dresden musste abgesagt werden! Lesen Sie mehr ...

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

MM: Fachverband Metall- und Materialphysik

MM 62: Topical Session: Data Driven Materials Science - Machine Learning Applications (joint session MM/CPP)

MM 62.3: Vortrag

Donnerstag, 19. März 2020, 18:00–18:15, BAR 205

An equation for membrane permeability: Insight from compressed sensing — •Arghya Dutta and Tristan Bereau — Max Planck Institute for Polymer Research, Mainz, Germany

Using a material’s structure and readily available properties to predict a difficult-to-measure but important property is crucial in natural sciences and engineering. Data mining—the process of discovering associations, correlations, and anomalies in data—can significantly facilitate the search for these generalized structure-property relationships by providing relevant descriptors. To give an example, the efficacy of a targeted drug depends on whether or not it can go through a cell membrane. This capacity is quantified by permeability which measures the drug’s flux across the membrane. However, calculating permeability is computationally expensive. In this presentation, I will discuss results from our ongoing search for a simple and interpretable equation, which is a function of key physical descriptors, for permeability using compressed sensing methods. This simplified description of permeability will allow simulation-free prediction, and, potentially, assist in rapid screening of candidate drug molecules.