Interpretable Mixture Density Estimation by use of Differentiable Tree-module

Ryuichi Kanoh, Tomu Yanabe

In order to develop reliable services using machine learning, it is important to understand the uncertainty of the model outputs. Often the probability distribution that the predic…