*2.2. Difference Functions on Multivariate Normal Distribution Manifold*

In this paper, we need to consider the difference between the probability density functions of different multivariate normal distributions. We select the Wasserstein distance as the difference function. At the same time, we also use Kullback–Leibler divergence, which is a difference function commonly used in classical information theory. We will introduce these difference functions respectively below [13–15].
