Exponential Varieties

Abstract

Exponential varieties arise from exponential families in statistics. These real algebraic varieties have strong positivity and convexity properties, familiar from toric varieties and their moment maps. Among them are varieties of inverses of symmetric matrices satisfying linear constraints. This class includes Gaussian graphical models. We develop a general theory of exponential varieties. These are derived from hyperbolic polynomials and their integral representations. We compare the multidegrees and ML degrees of the gradient map for hyperbolic polynomials.

References in corpus (2)

• Discrete Polymatroids
• Dimension of gram spectrahedra of univariate polynomials

Cited by in corpus (11)

• Discrete Gaussian distributions via theta functions
• Fast approximations of the Jeffreys divergence between univariate Gaussian mixture models via exponential polynomial densities
• Real Fibered Morphisms and Ulrich Sheaves
• The Chow form of a reciprocal linear space
• Lagrangian and Hamiltonian Mechanics for Probabilities on the Statistical Manifold
• Differential Equations for Gaussian Statistical Models with Rational Maximum Likelihood Estimator
• Spaces of Sums of Powers and Real Rank Boundaries
• Solution manifold and Its Statistical Applications
• Computing Maximum Likelihood Estimates for Gaussian Graphical Models with Macaulay2
• Random walks with local memory
• Mixed volumes of hypersimplices