The paper recommends to monotonize the confidence bounds when it is meaningful. GenericML should have this functionality.
The paper recommends to monotonize the confidence bounds when it is meaningful. GenericML should have this functionality.