From first-principles to ML-accelerated molecular dynamics: Quantitative predictive modeling of disordered materials for memory and energy applications

This talk will cover recent advancements in the atomistic modeling of disordered materials, focusing on chalcogenides, alkali-containing antiperovskites, and polyanionic glasses - materials of interest for memory devices and energy storage. The presentation will detail the use of first-principles molecular dynamics, whether in the Car-Parrinello or Born-Oppenheimer schemes, alongside the fast-pace evolving field of machine learning-accelerated MD techniques. Particular attention will be given to the quantitative structural insights these methods provide, and their relation to other key properties such as bonding and dynamical, offering predictive
capabilities essential for advancing material applications in technology.