A Two sided Calibration Theorem

Theorem 2. Suppose that the predictive distribution Q has the sufficient ability to approximate the true unknown distribution P, given data is i.i.d. Eqn. ( 13) holds by minimizing the MMD loss L B.1 Baselines MC-Dropout (MCD) [ 12 ]: A variant of standard dropout, named as Monte-Carlo Dropout. Epistemic uncertainties can be quantified with a Monte-Carlo sampling sample by using dropout during the test phase in the network without changing NNs model itself. For all experiments, the dropout probability was set at 0.3. The conventional MSE loss is used in this method.