The advent of powerful computational methods has been the impetus for the next generation of breakthroughs in science and technology. Machine learning and artificial intelligence methods have transformed the contemporary understanding of problem solving, unveiling answers to previously unanswerable questions. The use of these tools in the field of solid state physics has aided in the development of intelligent architectures to accelerate the design and discovery of new materials with targeted properties for specific applications. These intelligent architectures use a predictive model that can learn the relationship between crystal structure properties and resultant functional properties. Such predictive models have incredible potential to revolutionize power electronics, semiconductors, transistor technology, optoelectronics, photodetectors, quantum computers, solar cells, and more by discovering nearly perfect novel materials for each application. 

The goal of this research is to develop a scalable framework capable of predicting bandgap energy, formation energy, and entropic temperature of alloys. Bandgap energy is a measure of transparency in a semiconductor, formation energy is a measure of stability in semiconductors, and entropic temperature is a measure of an alloy’s effectiveness in  high-entropy applications. This research couples open source software and a database of calculated crystal-system properties with original methods of prediction to produce a new integrated framework. Such a framework requires feature generation from the compound properties which then enables regression to solve the central problem of interpolation. Feature generation methods for this framework includes n-gram histograms from crystal graphs as well as multi-dimensional scaling using earth-mover's distance between compound ordering parameters. Regression methods for this framework include kernel ridge regression, random forests, and neural networks. Once developed, this framework is anticipated to be applicable for a range of technological applications including thermoelectric devices. z