Graph Neural Networks and Structural Information on Ionic Liquids: A Cheminformatics Study on Molecular Physicochemical Property Prediction

• DOI: 10.1021/acs.jpcb.3c05521
• Publication Date: 2023-11-28T19:29:39Z
• Authors: Karol Baran; Adam Kloskowski
• Keywords: Cheminformatics

Ionic liquids (ILs) provide a promising solution in many industrial applications, such as solvents, absorbents, electrolytes, catalysts, lubricants, and others. However, due to the enormous variety of their structures, uncovering or designing those with optimal attributes requires expensive exhaustive simulations experiments. For these reasons, searching for an efficient theoretical tool finding relationship between IL structure properties has been subject research studies. Recently, special attention paid machine learning tools, especially multilayer perceptron convolutional neural networks, among other algorithms field artificial networks. latter, graph networks (GNNs) seem be powerful cheminformatic yet not well enough studied dual molecular systems ILs. In this work, usage GNNs structure-property studies is critically evaluated predicting density, viscosity, surface tension The problem data availability integrity discussed show how deal mislabeled chemical data. Providing more training proven important than ensuring that they are immaculate. Great process different ions give transformations electrostatic information. Clues on should applied predict ILs provided. Differences, regarding handling data, favoring use over classical quantitative models discussed.