Meg Stuart

Hello! I am a PhD candidate in computer science at the University of North Carolina at Chapel Hill advised by Dr. Sridhar Duggirala. I am doing research in formal verification of neural network behavior for medical image analysis. More generally, I am interested in data-driven solutions to problems with strong connections to the real world.  This involves examining the human experience represented in the data we work with and analyzing, not just the standard performance metrics, but rather the human impact of the algorithms we write.  I encourage the practice of thinking critically about datasets: in how representative they are of use-case and patient populations, how appropriate they are as a metric for good health outcomes, and, critically, how no dataset is "objective," each affected by structural and systemic bias, bias of the collectors (i.e. only including data from a specific institution), as well as bias inherit in the collection method (i.e. only using state-of-the-art imaging machines, rather than more common clinical diagnostic machines).  The goal is to work within these limitations to build medical software that is safe and effective for everyone. 

Outside of work, you can find me backpacking, hiking, camping, or kayaking, hanging out with my pets, growing plants, watching nature documentaries (especially the ocean ones), reading, or doing art.

I can be contacted via email at mstuart_at_cs_dot_unc_dot_edu.