Physics-inspired machine learning of localized intensive properties

article Physics-inspired machine learning of localized intensive properties published yes Ke Chen author c636c5ca-e8b8-11ed-b2d4-cc2c37613a8d Christian Kunkel author Bingqing Cheng author cbe3cda4-d82c-11eb-8dc7-8ff94289fcc90000-0002-3584-9632 Karsten Reuter author Johannes T. Margraf author BiCh department Machine learning (ML) has been widely applied to chemical property prediction, most prominently for the energies and forces in molecules and materials. The strong interest in predicting energies in particular has led to a ‘local energy’-based paradigm for modern atomistic ML models, which ensures size-extensivity and a linear scaling of computational cost with system size. However, many electronic properties (such as excitation energies or ionization energies) do not necessarily scale linearly with system size and may even be spatially localized. Using size-extensive models in these cases can lead to large errors. In this work, we explore different strategies for learning intensive and localized properties, using HOMO energies in organic molecules as a representative test case. In particular, we analyze the pooling functions that atomistic neural networks use to predict molecular properties, and suggest an orbital weighted average (OWA) approach that enables the accurate prediction of orbital energies and locations. https://research-explorer.ista.ac.at/download/12879/12883/2023_ChemialScience_Chen.pdf application/pdfno https://creativecommons.org/licenses/by/3.0/ Royal Society of Chemistry2023 eng Chemical Science 2041-6520 2041-6539 00097150810000110.1039/d3sc00841j Chen K, Kunkel C, Cheng B, Reuter K, Margraf JT. Physics-inspired machine learning of localized intensive properties. Chemical Science. 2023. doi:<a href="https://doi.org/10.1039/d3sc00841j">10.1039/d3sc00841j</a> Chen, Ke, et al. “Physics-Inspired Machine Learning of Localized Intensive Properties.” Chemical Science, Royal Society of Chemistry, 2023, doi:<a href="https://doi.org/10.1039/d3sc00841j">10.1039/d3sc00841j</a>. K. Chen, C. Kunkel, B. Cheng, K. Reuter, and J. T. Margraf, “Physics-inspired machine learning of localized intensive properties,” Chemical Science. Royal Society of Chemistry, 2023. Chen, K., Kunkel, C., Cheng, B., Reuter, K., & Margraf, J. T. (2023). Physics-inspired machine learning of localized intensive properties. Chemical Science. Royal Society of Chemistry. <a href="https://doi.org/10.1039/d3sc00841j">https://doi.org/10.1039/d3sc00841j</a> Chen, Ke, Christian Kunkel, Bingqing Cheng, Karsten Reuter, and Johannes T. Margraf. “Physics-Inspired Machine Learning of Localized Intensive Properties.” Chemical Science. Royal Society of Chemistry, 2023. <a href="https://doi.org/10.1039/d3sc00841j">https://doi.org/10.1039/d3sc00841j</a>. K. Chen, C. Kunkel, B. Cheng, K. Reuter, J.T. Margraf, Chemical Science (2023). Chen K, Kunkel C, Cheng B, Reuter K, Margraf JT. 2023. Physics-inspired machine learning of localized intensive properties. Chemical Science. 128792023-04-30T22:01:06Z2023-08-01T14:18:10Z