We have introduced a dually flat Riemannian structure in the manifolds of the exponential family and the mixture family based on the convexity of the cumulant generating function (free energy) and the negative entropy, respectively. The KL-divergence is derived from these convex functions.
Convex Function Fisher Information Coarse Graining Fisher Information Matrix Riemannian Metrics
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
This is a preview of subscription content, log in to check access