Accurate and Reliable Uncertainty Estimates for Deterministic Predictions Extensions to Under and Overpredictions

ArXi:2604.08755v1 Announce Type: cross Computational models high-stakes decisions across engineering and science, and practitioners increasingly seek probabilistic predictions to quantify uncertainty in such models. Existing approaches generate predictions either by sampling input parameter distributions or by augmenting deterministic outputs with uncertainty representations, including distribution-free and distributional methods.