no code implementations • 8 Jul 2021 • Mei Wang, Jianwen Su, Zhihua Lin
In this method, each sample in imprecision range space has a certain probability to be the real value, participating in the loss calculation.
