The project MLGREEN (Machine Learning Green Functions for Magnetic Materials and Spintronics) has been selected for funding by Helmholtz AI.
MLGREEN explores how machine learning can help researchers design better materials. The project focuses on speeding up the KKR Green function method, a powerful approach for studying complex materials, especially magnetic materials and systems with defects or disorder.
At the moment, this method becomes very computationally expensive when the systems under study grow larger. With cubic scaling, making a problem 10 times larger can increase the computational cost by a factor of 1000. The goal of MLGREEN is to use machine learning to reduce this to linear scaling. In that case, a system that is 10 times larger would require only about 10 times more computational cost.
This could make it possible to study much larger and more realistic materials systems than before. If successful, the project could help accelerate the discovery of novel materials for applications such as permanent magnets that do not rely on problematic rare-earth elements and spintronic devices.
The project is led by CASUS at HZDR together with partners at Forschungszentrum Jülich.
Further information is available in the two press releases:
