Predicting the Young’s Modulus of Silicate Glasses using High-Throughput Molecular Dynamics Simulations and Machine Learning

TLDR

Machine‑learning predictions of material properties typically require large, consistent datasets, which are often lacking or inconsistent in experimental data. We combine machine learning with high‑throughput molecular dynamics simulations to predict the Young’s modulus of silicate glasses. We compare several machine‑learning algorithms to balance accuracy, simplicity, and interpretability, using data generated from high‑throughput MD simulations. The combined approach yields reliable predictions across the full compositional domain of silicate glasses.

Abstract

The application of machine learning to predict materials’ properties usually requires a large number of consistent data for training. However, experimental datasets of high quality are not always available or self-consistent. Here, as an alternative route, we combine machine learning with high-throughput molecular dynamics simulations to predict the Young’s modulus of silicate glasses. We demonstrate that this combined approach offers good and reliable predictions over the entire compositional domain. By comparing the performances of select machine learning algorithms, we discuss the nature of the balance between accuracy, simplicity, and interpretability in machine learning.