Machine learning of optical properties of materials – predicting spectra from images and images from spectra

Abstract

As the materials science community seeks to capitalize on recent advancements in computer science, the sparsity of well-labelled experimental data and limited throughput by which it can be generated have inhibited deployment of machine learning algorithms to date. Several successful examples in computational chemistry have inspired further adoption of machine learning algorithms, and in the present work we present autoencoding algorithms for measured optical properties of metal oxides, which can serve as an exemplar for the breadth and depth of data required for modern algorithms to learn the underlying structure of experimental materials science data. Our set of 178 994 distinct materials samples spans 78 distinct composition spaces, includes 45 elements, and contains more than 80 000 unique quinary oxide and 67 000 unique quaternary oxide compositions, making it the largest and most diverse experimental materials set utilized in machine learning studies. The extensive dataset enabled training and validation of 3 distinct models for mapping between sample images and absorption spectra, including a conditional variational autoencoder that generates images of hypothetical materials with tailored absorption properties. The absorption patterns auto-generated from sample images capture the salient features of ground truth spectra, and band gap energies extracted from these autogenerated patterns are quite accurate with a mean absolute error of 180 meV, which is the approximate uncertainty from traditional extraction of the band gap energy from measurements of the full transmission and reflection spectra. Optical properties of materials are not only ubiquitous in materials applications but also emblematic of the confluence of underlying physical phenomena yielding the type of complex data relationships that merit and benefit from neural network-type modelling. † Electronic supplementary information (ESI) available: Details of model structure, additional model results, and le containing all model weights. See

A Critical Review of Machine Learning of Energy Materials

“…One possible future direction is to combine various sources of computational data by data fusion and multifidelity modeling. On the experimental side, high‐throughput measurements of materials optical properties are providing a large quantity of data that may be used for ML purposes . We are likely to see more efforts on experimental automation and high‐throughput works in the near future …”

A data ecosystem to support machine learning in materials science

“…Now, with MDF, the dataset can be partitioned via user queries, immediately enabling new applications and data mixing opportunities. Second, we published several models into DLHub (as servables) based on, and extending, the models of Stein et al [26], including a first model that resembles the optical image VAE described in the paper, a second optical image autoencoder (AE) model, and a third model that instead uses color clustering techniques to predict the material bandgap. We then used the dataset, as available in MDF Discover, to streamline the process of retrieving data (Figure 4a) and running servables within DLHub on the retrieved data ( Figure 4b).…”

Deep Learning for Deep Chemistry: Optimizing the Prediction of Chemical Patterns

“…However, the establishment and understanding of holistic, or macro insights on the major research trends in Chemistry subfields, are critical tasks. The challenge relies on how the analysis of these sub-fields, with thousands published works, reveals the most prominent applications supported by ML approaches (Butler et al, 2018;Chmiela et al, 2018;Chuang and Keiser, 2018a;Coley et al, 2018a;Gao et al, 2018;Lo et al, 2018;Panteleev et al, 2018;Xia and Kais, 2018;Ceriotti, 2019;Chan et al, 2019;Christensen et al, 2019;Gallidabino et al, 2019;Häse et al, 2019;Iype and Urolagin, 2019;Mezei and Von Lilienfeld, 2019;Schleder et al, 2019;Stein et al, 2019a;Wang et al, 2019).…”