Predicting Young's modulus of oxide glasses with sparse datasets using machine learning

Abstract

Machine learning (ML) methods are becoming popular tools for predicting and designing novel materials. In particular, neural network (NN) is a promising ML method, which can be used to identify hidden trends in the data. However, these methods rely on a large dataset and often exhibit overfitting when used with a sparse dataset. Further, assessing the uncertainty in predictions for a new dataset or an extrapolation of the present dataset is challenging. Herein, using Gaussian process regression (GPR), we predict Young's modulus for silicate glasses having a sparse dataset. We show that GPR significantly outperforms NN for the sparse dataset while ensuring no overfitting. Further, thanks to the nonparametric nature, GPR provides quantitative bounds for the reliability of predictions while extrapolating. Overall, GPR presents an advanced ML methodology for accelerating the development of novel functional materials such as glasses.