<ul><li>Output uncertainty indicates whether the probabilistic properties reflect objective characteristics of the model output.</li><li>A post-processing parametric calibration method called $ho$-Norm Scaling is introduced to mitigate overconfidence in limited data sets.</li><li>The method expands the calibrator expression to preserve accuracy while reducing excessive amplitude.</li><li>Probability distribution regularization is included to ensure the instance-level uncertainty distribution resembles the distribution before calibration.</li></ul>

Parametric $\rho$-Norm Scaling Calibration

Discover more