Epistemic Uncertainty

Epistemic Uncertainty in Conformal Scores: A Unified Approach
WE introduce EPICSCORE, a model-agnostic method that enhances conformal prediction by integrating epistemic uncertainty. Compatible with any Bayesian model and maintaining distribution-free guarantees, EPICSCORE adapts prediction intervals based on data availability, achieving both finite-sample marginal and asymptotic conditional coverage.
L. M. C. Cabezas, V. S. Santos, T. R. Ramos, Rafael Izbicki
May, 2025 Proceedings of Machine Learning Research (UAI Track; Oral Presentation)
PDF Cite Code
Epistemic Uncertainty in Conformal Scores: A Unified Approach