Predicting the inhibition efficiencies of magnesium dissolution modulators using sparse machine learning models

• Published in: npj computational materials 2021-12, Vol.7 (1), p.1-9, Article 193
• Main Authors: Schiessler, Elisabeth J., Würger, Tim, Lamaka, Sviatlana V., Meißner, Robert H., Cyron, Christian J., Zheludkevich, Mikhail L., Feiler, Christian, Aydin, Roland C.
• Format: Article
• Language: English
• Subjects: 639/301/1034/1037; 639/638/563; Additives; Artificial neural networks; Characterization and Evaluation of Materials; Chemical compounds; Chemistry and Materials Science; Computational Intelligence; Corrosion; Learning algorithms; Machine learning; Magnesium; Magnesium base alloys; Materials Science; Mathematical and Computational Engineering; Mathematical and Computational Physics; Mathematical Modeling and Industrial Mathematics; Modulators; Neural networks; Organic chemistry; Theoretical; Variance analysis
• ISSN: 2057-3960; 2057-3960
• DOI: 10.1038/s41524-021-00658-7

The degradation behaviour of magnesium and its alloys can be tuned by small organic molecules. However, an automatic identification of effective organic additives within the vast chemical space of potential compounds needs sophisticated tools. Herein, we propose two systematic approaches of sparse feature selection for identifying molecular descriptors that are most relevant for the corrosion inhibition efficiency of chemical compounds. One is based on the classical statistical tool of analysis of variance, the other one based on random forests. We demonstrate how both can—when combined with deep neural networks—help to predict the corrosion inhibition efficiencies of chemical compounds for the magnesium alloy ZE41. In particular, we demonstrate that this framework outperforms predictions relying on a random selection of molecular descriptors. Finally, we point out how autoencoders could be used in the future to enable even more accurate automated predictions of corrosion inhibition efficiencies.