Adaptive Conformal Prediction for Image Regression Models with Application to an Inertial Confinement Fusion Emulator
My rating
The paper introduces adaptive conformal prediction for image-based regression models deployed in scientific machine learning. Its application targets an inertial confinement fusion emulator and addresses the quantification of predictive uncertainty for black-box model outputs. The work also considers assessing reliability across different regions of the input space when model internals are difficult to interpret.
Tags: