National Institute for Materials Science, Tsukuba (Japan)
https://samurai.nims.go.jp/profiles/lambard_guillaume/publications?locale=en
Tuesday April 11, 2023 - 2:30 p.m.
ENSCR
11 allée de Beaulieu - Amphi B
Inorganic Theoretical Chemistry & Solid State Chemistry and Materials teams
You can attend the conference online: https://zoom.us/j/97085487084?pwd=a2FRUlNOL2ZqSlRQZ1VLOHIwY2dtdz09
Contact: Régis Gautier (3 80 02) – rgautier
ensc-rennes [dot] fr
ABSTRACT
"In this talk, we will introduce the active learning principle assisted by Bayesian optimization and its application to design new materials. We will demonstrate the capabilities of an active learning pipeline through the process of an epoxy resin with improved adhesive strength, the synthesis of a mesoporous PtPdAu alloy for electrochemical oxidation of methanol, the optimal combination of additives for rechargeable Li–O2 batteries, the optimization of the direct extrusion process for Nd-Fe-B magnets with improved magnetic properties, and the synthesis of thermoelectric materials like the binary alloy of GeTe as well as a quaternary alloy of Cu_{2.125}Zn_{0.875}Sn_{1}S_{4} (kesterite) with improved thermoelectric properties. We will demonstrate that an active pipeline coupled with experimental data can deliver novel perspectives on material behaviours regarding their synthesis process. Additionally, we will introduce our dedicated and open graphical user interface (GUI) system that should allow any experimentalist to leverage our active learning pipeline with no previous coding experience."
