Kolloquium

Abstract

Machine learning can accelerate the screening, design, and discovery of new molecules and materials in multiple ways, e.g. by virtually predicting properties of molecules and materials, by extracting hidden relations from large amounts of simulated or experimental data, or even by interfacing machine learning algorithms for autonomous decision-making directly with automated high-throughput experiments. In this talk, I will focus on our research activities on graph neural networks for property prediction [1,2] and understanding of structure-property relations [3], as well as on the use of graph neural networks for accelerated atomistic simulations [4,5]. Application areas range from superconductors, metal-organic frameworks, and organic semiconductors to photochemical reactions. Furthermore, I will go beyond virtual design and simulations, and give a brief outlook on the use of machine learning methods for data analysis and decision-making processes in automated materials science and chemistry labs [6,7].

[1] Reiser et al., Communications Materials 3 (1) (2022), https://www.nature.com/articles/s43246-022-00315-6
[2] KGCNN library, https://github.com/aimat-lab/gcnn_keras
[3] Teufel et al., NeurIPS Workshop (2022), https://arxiv.org/abs/2211.13236
[4] Friederich et al., Nature Materials 20 (6) (2021), https://www.nature.com/articles/s41563-020-0777-6
[5] Li et al., Chemical Science 12 (2021), https://pubs.rsc.org/en/content/articlehtml/2021/sc/d0sc05610c
[6] Luo et al., Angewandte Chemie International Edition 61 (19) (2022), https://onlinelibrary.wiley.com/doi/full/10.1002/anie.202200242
[7] Velasco et al., Advanced Materials 33 (43) (2021), https://onlinelibrary.wiley.com/doi/full/10.1002/adma.202102301

Bio

After his B.S. and M.Sc. in physics and a Ph.D. in physics on multiscale modeling of organic semiconductors, Pascal Friederich received a Marie-Sklodowska-Curie Postdoctoral Fellowship at Harvard University and the University of Toronto where he worked on machine learning methods for chemistry. In 2020, Pascal Friederich joined the Informatics Department of the Karlsruhe Institute of Technology as a tenure-track professor, leading the AI for Materials Science (AiMat, https://aimat.science) research group. The AiMat research group focuses on developing and applying machine learning methods for property prediction, simulation, understanding, and design of molecules and materials, as well as on interfacing machine learning methods with automated materials experiments. In 2022, Pascal Friederich received the Heinz-Maier-Leibnitz Prize from the German Research Foundation.