In our events, we are committed to Aalto University’s principles for a safer space.
Host: Prof. Patrick Rinke (Computational Electronic Structure Theory)
Title: The predictive accuracy of machine learning force fields under active learning conditions
Abstract:
Upcoming automated laboratories for accelerated materials discovery are expected to have an important simulation component. Machine Learning Force Fields (MLFF) with an active learning mode provides an excellent simulations workhorse. In such labs, seed crystal structures are generated by combinatorial synthesis and then passed to simulations to rapidly explore the broader thermodynamic phase space from the known stable point. It is essential to assess the predictive accuracy of the physical observables simulated by these methods, in order to minimize the required number of (expensive time/equipment) experiments, and to prevent the simulations steering the materials exploration into the wrong direction. In this talk, I will introduce our implementation of on-the-fly learning in a commonly used Density Functional Theory software package. I will showcase some successful application examples from the (in)organic halide perovskites, where we have simulated, using (non)-equilibrium molecular dynamics, crystal phase transitions, thermal conductivity, and other physical observables that would have been impossible using fully ab-initio methods. These soft materials are illustrative of a much larger class of ‘Dynamic Solids’, and are characterized by strong anharmonic lattice dynamics and wide variety of rare-events. I will conclude with some very recent results on the predictive accuracy of superionic phase transitions in AX2 fluorites. In this study we emulate the situation where the MLFF is part of an automated materials discovery pipeline.
Menno Bokdam is an assistant professor in physics and materials science at the University of Twente, the Netherlands. His research focuses on understanding the atomic and electronic structure of complex crystalline materials under realistic temperature and pressure conditions. He develops actively learned machine-learning interatomic potentials trained on first-principles electronic structure calculations. This enables large-scale atomistic simulations of materials with strongly anharmonic behavior, such as molecular rotations, ion rattling, and ionic diffusion. His work combines artificial intelligence and computational materials science to predict and explain material properties including phonon spectra, thermal conductivity, thermal expansion, lattice dynamics, and ion transport.
CEST is developing electronic structure and machine learning methods and applying them to computational materials science problems. CEST moved their operations to the Technical University of Munich in 2024.
