The Federal Ministry of Research, Technology and Space (BMFTR) has approved funding for three consortia to establish high-performance AI computing infrastructures for AI research within the Leibniz Association.
AI Infrastructure (Diagram) | Copyright: WIAS (Design); stock.adobe.com
One of these initiatives, focusing on materials science and process engineering, is being coordinated by the Weierstrass Institute for Applied Analysis and Stochastics (WIAS).
"The goal of the project is to establish a sustainable, high-performance research infrastructure for developing and applying modern artificial intelligence methods," stated Prof. Dr. Michael Hintermüller, consortium spokesperson and WIAS director. The project focuses on key technologies of Germany's High-Tech Agenda, including microelectronics, quantum technologies, nuclear fusion, climate-neutral energy technologies and mobility, and biotechnology. It covers the entire spectrum of innovation, from basic and high-tech research to commercial implementation. In addition to strengthening institutional cooperation within the Leibniz Association, the project also aims to attract outstanding researchers from around the world.
The unique integration of fundamental mathematical research and experimental research groups is based on existing collaborations, particularly within the Leibniz Research Network "Mathematical Modeling and Simulation (MMS)", as well as complementary scientific and technological expertise. It will greatly improve the efficiency, reliability, and trustworthiness of AI methods. In the field of scientific machine learning, this will lead to the development of physically informed neural networks that can adapt to specific problems, for example. Improvements in numerical solution algorithms for high-dimensional learning problems will contribute to competitive methods in modern material development. Similarly, measurement data and neural networks will be used to simulate, control, or optimize complex process engineering procedures in a data-driven manner. Other areas of application include, for example, quantum chemical methods for characterizing the electronic structure of materials; cross-scale methods for simulating plasma processes and mechanical stresses on materials; and thermodynamic modeling to determine phase transitions. Large language models will also be incorporated into the research methodology.
The following Leibniz Institutes are contributing their complementary scientific and technological expertise to the project under the coordination of WIAS: the Leibniz Institute for Plasma Science and Technology (INP) in Greifswald, the Leibniz Institute of Surface Engineering (IOM) in Leipzig, the Leibniz Institute of Polymer Research (IPF) and the Leibniz Institute for Solid State and Materials Research (IFW), both based in Dresden, Kaiserslautern's Leibniz-Institut für Verbundwerkstoffe (IVW), the Leibniz Institute for Materials Engineering (IWT) in Bremen, Rostock's Leibniz Institute for Catalysis (LIKAT), Schloss Dagstuhl - Leibniz Center for Informatics (LZI), and the Berlin-based institutes Ferdinand-Braun-Institut - Leibniz-Institut für Höchstfrequenztechnik (FBH), the Leibniz-Institut für Kristallzüchtung (IKZ), and the Paul-Drude-Institut für Festkörperelektronik (PDI).
