This is Jessica. Conformal prediction, referring to a class of distribution-free approaches to quantifying predictive uncertainty, has attracted interest for medical AI applications. Reasons include because prediction sets seem to align with the kinds of differential diagnoses doctors already use, … Continue reading