FURI | Fall 2023

Utilizing Neural Networks with Molecular Dynamics to Design High-Entropy Functional Alloys for Thermoelectric Technology

The standard approach of discovering thermoelectric materials is costly and very complex. In this current study, Graph Neural Networks (GNNs) will be utilized to select elements and determine the compositions. GNNs use message passing to organize graphs so machine learning algorithms can use them. With the knowledge that machine learning can extract, the key characteristics from these databases can be used to predict behavior of potential compounds.