Uncertainty-aware post-training rejectors for deciding when an AI system should defer to a human expert.
Feb 1, 2026
Data-split and calibration choices for inductive conformal predictors, with a focus on uncertainty quality and practical reliability.
Jan 1, 2024