Multi-Modal ML Pre-Print Published
Our new pre-print on multi-modal machine learning (ML) to describe local order in materials for high-radiation environments is now out! This study was spearheaded by Arman Ter-Petrosyan and Mike Holden, who both have brilliant scientific careers ahead of them. We are also grateful for the collaboration with the groups of Khalid Hattar, Ryan Comes, and Yingge Du.
Our study uses ML to analyze the chemistry and structure of interfaces at the atomic scale. We show that multi-modal methods, which combine multiple data imaging and spectroscopic sources, can provide new insights that wouldn't be available with the individual data streams alone. We discuss the potential and limitations of these models to provide deep descriptors of order in extreme environments.