On the Computability of Continuous Maximum Entropy Distributions: Adjoint Orbits of Lie Groups

11/03/2020
by   Jonathan Leake, et al.
0

Given a point A in the convex hull of a given adjoint orbit 𝒪(F) of a compact Lie group G, we give a polynomial time algorithm to compute the probability density supported on 𝒪(F) whose expectation is A and that minimizes the Kullback-Leibler divergence to the G-invariant measure on 𝒪(F). This significantly extends the recent work of the authors (STOC 2020) who presented such a result for the manifold of rank k-projections which is a specific adjoint orbit of the unitary group U(n). Our result relies on the ellipsoid method-based framework proposed in prior work; however, to apply it to the general setting of compact Lie groups, we need tools from Lie theory. For instance, properties of the adjoint representation are used to find the defining equalities of the minimal affine space containing the convex hull of 𝒪(F), and to establish a bound on the optimal dual solution. Also, the Harish-Chandra integral formula is used to obtain an evaluation oracle for the dual objective function. While the Harish-Chandra integral formula allows us to write certain integrals over the adjoint orbit of a Lie group as a sum of a small number of determinants, it is only defined for elements of a chosen Cartan subalgebra of the Lie algebra 𝔤 of G. We show how it can be applied to our setting with the help of Kostant's convexity theorem. Further, the convex hull of an adjoint orbit is a type of orbitope, and the orbitopes studied in this paper are known to be spectrahedral. Thus our main result can be viewed as extending the maximum entropy framework to a class of spectrahedra.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
04/16/2020

On the computability of continuous maximum entropy distributions with applications

We initiate a study of the following problem: Given a continuous domain ...
research
08/03/2022

On the Identity Problem and the Group Problem in nilpotent groups

Let 𝒢 be a finite set of matrices in a unipotent matrix group G over ℚ, ...
research
11/10/2020

A Polynomial-Time Algorithm and Applications for Matrix Sampling from Harish-Chandra–Itzykson-Zuber Densities

Given two n × n Hermitian matrices Y and Λ, the Harish-Chandra–Itzykson–...
research
11/06/2017

Computing Maximum Entropy Distributions Everywhere

We study the problem of computing the maximum entropy distribution with ...
research
06/07/2021

Bridge Simulation and Metric Estimation on Lie Groups

We present a simulation scheme for simulating Brownian bridges on comple...
research
03/07/2019

Reparameterizing Distributions on Lie Groups

Reparameterizable densities are an important way to learn probability di...
research
06/28/2019

Computing Haar Measures

According to Haar's Theorem, every compact group G admits a unique left-...

Please sign up or login with your details

Forgot password? Click here to reset