Machine Learning for Materials Design

We develop the Materials Learning Algorithms (MALA), a physics-informed machine learning framework that accelerates conventional density functional theory simulations. Using neural networks, MALA efficiently computes the electronic structure of matter, enabling accurate determination of energies and forces that are critical for atomistic simulations. MALA is a scalable method that overcomes the limitations of density functional theory simulations, paving the way for electronic structure calculations at unprecedented length and time scales.

We use a combination of first-principles calculations and machine learning models to generate interatomic potentials for high-performance molecular-spin dynamics simulations. This allows us to simulate atomistic and spin dynamics simultaneously, enabling simulations of structural stability, transport phenomena, and magneto-structural phase transitions in materials. This approach shows promise in advancing next-generation magnetic materials and ultrafast magnetic storage technologies.

We apply state-of-the-art machine learning techniques to advance first-principles simulations, paving the way for rapid and targeted materials discovery. We employ physics-informed neural networks for inverting fundamental quantum mechanical equations, neural operators for modeling electron dynamics, and generative models for materials discovery.