Chemical discovery assisted by machine learning

Julia Westermayr, Rhyan Barrett

Chemical reactions are fundamental to drive the transformation of matter and are pivotal across diverse domains like medicine, materials science, and energy generation. In this talk, we will explore the potential of machine learning algorithms to explore the discovery of chemical reactions in their ground and excited states. Specifically, we will illustrate the proficiency of deep neural networks in accelerating the prediction of excited-state properties, thereby enhancing our understanding of the photochemical processes [1,2,3]. Additionally, we will showcase the efficiency of reinforcement learning in expediting exploration through the vast expanse of chemical (structure) space [4].

1. Julia Westermayr and Philipp Marquetand, Chem. Rev. 2020, 121 (16), 9873-9926.
2. Julia Westermayr, Michael Gastegger, Dóra Vörös, Lisa Panzenboeck, Florian Joerg, Leticia González, and Philipp Marquetand, Nat. Chem. 2022, 14 (8), 914-919.
3. Julia Westermayr, Joe Gilkes, Rhyan Barrett, Nat. Comput. Sci 2023, 3 (2), 139-148.
4. Rhyan Barrett and Julia Westermayr, J. Phys. Chem. Lett. 2024, 15, 349-356.