Rebecca Yu

I am a first year Ph.D. student in the Machine Learning Department at Carnegie Mellon University.

Previously, I obtained my B.S. in Biomedical Engineering and Computer Science from Johns Hopkins University, where I worked in the Trayanova Lab mentored by Rheeda Ali and in the Johns Hopkins Center for Computational Biology advised by Mihaela Pertea.

email:

CV  /  Google Scholar  /  Twitter  /  Github

I am broadly interested in understanding (and improving) human-AI decision making while considering the social and ethical implications of machine learning in high impact problem domains.

We developed a novel deep learning solution for analysis of contrast-enhanced cardiac magnetic resonance (CMR) images that produces anatomically accurate myocardium and scar/fibrosis ssegmentations and uses these to calculate clinical features of interest.

Variable atrial anatomy and thin atrial walls, compounded with poor and inconsistent image quality, lead to low generalizability in segmentation neural networks. We developed a two-stage 3D neural network that is able to accurately delineate both the left and right atria in under 5 seconds for multi-modal clinical cardiac imaging.

Cardiac sarcoidosis (CS), an inflammatory disease characterized by formation of granulomas in the heart, is associated with high risk of sudden cardiac death (SCD) from ventricular arrhythmias. Current “one-size-fits-all” guidelines for SCD risk assessment in CS result in insufficient appropriate primary prevention. We present a two-step precision risk prediction technology for patients with CS. In a retrospective study of 45 patients, the technology outperformed clinical metrics, highlighting its potential to transform CS risk stratification.