Skip to main content

Visualizing probabilistic models in Minkowski space with intensive symmetrized Kullback-Leibler embedding

Cornell Affiliated Author(s)

Author

H.K. Teoh
K.N. Quinn
J. Kent-Dobias
C.B. Clement
Q. Xu
J.P. Sethna

Abstract

We show that the predicted probability distributions for any N-parameter statistical model taking the form of an exponential family can be explicitly and analytically embedded isometrically in a N+N-dimensional Minkowski space. That is, the model predictions can be visualized as control parameters are varied, preserving the natural distance between probability distributions. All pairwise distances between model instances are given by the symmetrized Kullback-Leibler divergence. We give formulas for these intensive symmetrized Kullback-Leibler (isKL) coordinate embeddings, and illustrate the resulting visualizations with the Bernoulli (coin-toss) problem, the ideal gas, n-sided die, the nonlinear least-squares fit, and the Gaussian fit. We highlight how isKL can be used to determine the minimum number of parameters needed to describe probabilistic data, and conclude by visualizing the prediction space of the two-dimensional Ising model, where we examine the manifold behavior near its critical point. © 2020 authors. Published by the American Physical Society.

Date Published

Journal

Physical Review Research

Volume

2

Issue

3

URL

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85115906742&doi=10.1103%2fPhysRevResearch.2.033221&partnerID=40&md5=ddde1faa8dd82a48bdbe70e6b427dc15

DOI

10.1103/PhysRevResearch.2.033221

Group (Lab)

James Sethna Group

Funding Source

1719490

Download citation