Towards Trustworthy Machine Learning in Medicine
Katherine Heller
December 1, 2021, Wednesday, 3:00 PM - 4:00 PM EST
As ML is increasingly used in society, we need methods that we have confidence that we can rely on, particularly in the medical domain. In this talk I discuss 3 pieces of work:

1) Sepsis Watch - I present a Gaussian Process (GP) + Recurrent Neural Network (RNN) model for predicting sepsis infections in Emergency Department patients. I will discuss the benefit of uncertainty given by the GP. I will then discuss the social context in introducing such a system into a hospital setting.

2) Underspecification and the credibility implications of hyperparameter choices in ML models -- I will discuss medical imaging applications and how using the uncertainty of model performance conditioned on choice of hyperparameters can help identify situations in which methods may not generalize well outside the training domain.

3) Inclusive Mobile Health Technology - I will present our MS Mosaic iphone app, which attempts to understand someone's MS between clinic visits, and summarize their health for themselves and their provider. I will present a first look at medical predictions in this space, as well as some of our aims in including a more representative set of voices in our design and analysis.
Katherine Heller is a Research Scientist at Google. She has worked on developing and integrating multiple machine learning systems into hospitals and clinical care including: a sepsis detection system which has been integrated into the Duke University Hospital Emergency Departments, a system for detecting the likelihood of complications resulting from surgery, and a nationally released mobile study on Multiple Sclerosis. She is interested in machine learning and ethics in a medical context, and the inclusion of all people in the development of medical technology. Before joining Google, she was at Duke University in Statistical Science, Neurobiology, Neurology, Computer Science, and Electrical and Computer Engineering. She was the recipient of an NSF CAREER award and a first round BRAIN initiative award.