Helmholtz Association funds materials science AI project led by CASUS
Many materials are tested on computers long before they are produced in the laboratory. Yet even on high-performance computers, some of the most powerful methods in materials science remain limited to comparatively small systems far from realistic scales. The MLGREEN project, initiated by CASUS Department Lead Dr. Attila Cangi, plans to use artificial intelligence (AI) to reduce the computational cost of the Korringa-Kohn-Rostoker (KKR) Green function method, making this highly accurate but demanding simulation approach applicable to much larger and more realistic materials systems. The MLGREEN proposal was submitted by two Helmholtz centers: the Helmholtz-Zentrum Dresden-Rossendorf (HZDR) with its Center for Advanced Systems Understanding (CASUS) and the Forschungszentrum Jülich (FZJ). In the competitive process run by the Helmholtz Association’s Artificial Intelligence Cooperation Unit Helmholtz AI only 11 out of 45 proposals received funding.
Many material properties depend on electron behavior, including electrical conductivity, magnetism, certain optical properties, and superconductivity. So what material scientists, chemists and physicists want to know is, broadly speaking, where are the electrons, how do they move, and what energy do they have? The KKR Green function is a mathematical tool that describes how electrons move through a material while interacting with atomic nuclei that determine the material’s internal structure. The function captures all possible paths and scattering events in a compact way. It is therefore particularly useful for studying complex systems where regular atomic order is interrupted, for example in alloys, deliberately introduced impurities or defects of materials, material surfaces, or magnetic materials. The name of this function derives from the British mathematician George Green (1793–1841), who introduced the mathematical ideas behind these functions, and Jan Korringa, Walter Kohn, and Norman Rostoker, who developed the method around 1950.
Additional information:
Dr. Attila Cangi
Research Team Leader
Center for Advanced Systems Understanding (CASUS) at HZDR
Media contact:
Dr. Martin Laqua
Officer Communications, Press and Public Relations
Center for Advanced Systems Understanding (CASUS) at HZDR
The Helmholtz AI Project Call supports innovative research projects that develop new artificial intelligence methods or advance their application in scientific research. They also strengthen collaboration both within the Helmholtz Association and with partners from academia and industry. Source: Helmholtz AI
The KKR Green method is powerful, but it is not used widely in science and industry. “The method is mathematically complex and is therefore harder to learn and implement than more standard methods,” says Dr. Attila Cangi, Head of the CASUS Research Team “Machine Learning for Materials Design”. “However, commonly used methods such as plane-wave density functional theory are not well suited to challenging and technologically relevant systems with disorders, defects, or interfaces – and to magnetic materials in general,” he adds.
Cutting down on computational effort
Another severe obstacle is computational cost. The application of the KKR Green method is – despite high-performance computers – restricted to systems of only a few hundred atoms, precluding the simulation of realistic nanoscale systems where long-range interactions and complex geometries govern functionality. “Computing the KKR Green function scales cubically. For solving materials problems that are ten times larger, the computational effort increases by a factor of 1,000 with the conventional KKR Green method. We want to overcome this bottleneck,” says Cangi.
The MLGREEN project, short for “Machine Learning Green functions for magnetic materials and spintronics”, directly addresses the scaling challenge. Using neural networks, the aim is to reduce the computational cost to linear scaling with system size. If successful, the project will 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 – that is, devices that exploit not only the charge but also the spin of electrons.
In addition to Dr. Attila Cangi, the principal investigators of MLGREEN are Prof. Dr. Stefan Blügel and Dr. Daniel Wortmann both from the Peter Grünberg Institute at FZJ.
The Helmholtz Artificial Intelligence Cooperation Unit (Helmholtz AI) strengthens the application and development of applied artificial intelligence and machine learning. Within the Helmholtz AI competition, a panel of experts selected in particular those research projects that promise a high gain in insight. However, such projects are also considered especially risky. There is a good chance that unsolvable problems will arise and the outlined project objective will not be achieved. CASUS is especially successful in acquiring funds through this competitive program. MLGREEN is already the third selected CASUS proposal since this funding instrument was installed in 2019. MLGREEN will receive 150,000 euros from the Helmholtz Association’s Impulse and Networking Fund, with the same amount to be contributed by the participating Helmholtz Centers HZDR and FZJ.
________________________________________________________
About the Center for Advanced Systems Understanding
CASUS was founded 2019 in Görlitz/Germany and pursues data-intensive interdisciplinary systems research in such diverse disciplines as earth systems research, systems biology or materials research. The goal of CASUS is to create digital images of complex systems of unprecedented fidelity to reality with innovative methods from mathematics, theoretical systems research, simulations as well as data and computer science to give answers to urgent societal questions. The founding partners of CASUS are the Helmholtz-Zentrum Dresden-Rossendorf (HZDR), the Helmholtz Centre for Environmental Research in Leipzig (UFZ), the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden (MPI-CBG), the Technical University of Dresden (TUD) and the University of Wrocław (UWr). CASUS, managed as an institute of the HZDR, is funded by the German Federal Ministry of Research, Technology and Space (BMFTR) and the Saxon State Ministry for Science, Culture and Tourism (SMWK).
Additional information:
Dr. Attila Cangi
Research Team Leader
Center for Advanced Systems Understanding (CASUS) at HZDR
Media contact:
Dr. Martin Laqua
Officer Communications, Press and Public Relations Center for Advanced Systems Understanding (CASUS) at HZDR