Despite their incredible versatility and experimental success, Artificial Neural Networks are subject to unusual sensitivities which have been recently discovered. This research focuses on expanding mathematical theory to explain and classify such examples in order to make machine-learning models more robust and transparent.

Working with the college of medicine, our team has received and is analyzing data from more than 1600 covid patients from the Banner hospital system. The goal of this research is to predict patient trajectories and need for escalated treatment protocols from initial patient data and vital signs.

The goal of this research is to understand how time-varying series of data can be clustered and compared by extending the geometric concept of "depth"