Interpreting the optical properties of oxide glasses with machine learning and Shapely additive explanations

Due to their excellent optical properties, glasses are used for various applications ranging from smartphone screens to telescopes. Developing compositions with tailored Abbe number (V-d) and refractive index at 587.6 nm (n(d)), two crucial optical properties, is a major challenge. To this extent, machine learning (ML) approaches have been successfully used to develop composition-property models. However, these models are essentially black boxes in nature and suffer from the lack of interpretability. In this paper, we demonstrate the use of ML models to predict the composition-dependent variations of V-d and n(d). Further, using Shapely additive explanations (SHAP), we interpret the ML models to identify the contribution of each of the input components toward target prediction. We observe that glass formers such as SiO2, B2O3, and P2O5 and intermediates such as TiO2, PbO, and Bi2O3 play a significant role in controlling the optical properties. Interestingly, components contributing toward increasing the n(d) are found to decrease the V-d and vice versa. Finally, we develop the Abbe diagram, using the ML models, allowing accelerated discovery of new glasses for optical properties beyond the experimental pareto front. Overall, employing explainable ML, we predict and interpret the compositional control on the optical properties of oxide glasses.

Automated summary

This paper demonstrates the use of machine learning models to predict and interpret the optical properties of oxide glasses, showing the application of ML models in compositional control and providing explanations of the model results through Shapely additive explanations (SHAP).