Epistemic Errors of Imperfect Multitask Learners When Distributions Shift

ArXi:2505.23496v3 Announce Type: replace Uncertainty-aware machine learners, such as Bayesian neural networks, output a quantification of uncertainty instead of a point prediction. We provide uncertainty-aware learners with a principled framework to characterize, and identify ways to eliminate, errors that arise from reducible (epistemic) uncertainty. We