Machine learning predictions of Knoop hardness in lithium disilicate glass‐ceramics

Abstract

Predicting the effects of ceramic microstructures on macroscopic properties, such as the Knoop hardness, has long been a difficult task. This is particularly true in glass–ceramics, where multiple unique crystalline phases can overlap with a background glassy phase. The combination of crystalline and glassy phases makes it difficult to quantify the percent crystallinity and to predict properties that are the result of the chemical composition and microstructure. To overcome this difficulty and take the first step to build a system for characterizing glass‐ceramics, we predict the Knoop hardness based on scanning electron microscopy images using two computational techniques. The first technique is a computer vision algorithm that allows for physical insights into the system because the features used in a predictive model are extracted from the images. The second technique is machine learning with convolutional neural networks that are trained through transfer learning, allowing for more accurate predictions than the first method but with the downside of being a black box. Discussion of the relative merits of the models is included.

Machine learning‐based accelerated design of fluorphlogopite glass ceramic chemistries with targeted hardness

“…The application of machine learning in glass ceramics is still being explored, as it is a relatively new topic. There has been recent reports on the development of glass ceramic systems using machine learning [34][35][36] ; however, due to the lack of data reliabil-ity and limited data accessibility, there is a difficulty in bridging the gap between physics-based models and data-driven models. 11 In this work, the reason behind selecting the features has been explained based on physics and chemistry, thereby narrowing the gap between datadriven models and empirical models.…”