English

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

Statistical Mechanics 2020-08-12 v3 Computational Physics Data Analysis, Statistics and Probability

Abstract

We show that the predicted probability distributions for any NN-parameter statistical model taking the form of an exponential family can be explicitly and analytically embedded isometrically in a N+NN{+}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, nn 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.

Keywords

Cite

@article{arxiv.1912.06039,
  title  = {Visualizing probabilistic models in Minkowski space with intensive symmetrized Kullback-Leibler embedding},
  author = {Han Kheng Teoh and Katherine N. Quinn and Jaron Kent-Dobias and Colin B. Clement and Qingyang Xu and James P. Sethna},
  journal= {arXiv preprint arXiv:1912.06039},
  year   = {2020}
}

Comments

18 pages, 9 figures

R2 v1 2026-06-23T12:44:15.474Z