A first approach to closeness distributions

TL;DR

This paper introduces an information geometric approach to modeling the similarity between smaller distributions within probabilistic graphical models, offering reinterpretations of existing models.

Contribution

It presents a novel method for integrating similarity information into probabilistic graphical models using information geometry, expanding the understanding of existing models.

Findings

Provides a geometric framework for distribution similarity

Reinterprets existing probabilistic models through this framework

Offers potential for improved modeling of related distributions

Abstract

Probabilistic graphical models allow us to encode a large probability distribution as a composition of smaller ones. It is oftentimes the case that we are interested in incorporating in the model the idea that some of these smaller distributions are likely to be similar to one another. In this paper we provide an information geometric approach on how to incorporate this information, and see that it allows us to reinterpret some already existing models.