Conditional Matrix Flows for Gaussian Graphical Models

Code

• github.com/fabricioarendtorres/flowconductor
  OfficialIn paperpytorch★ 12
• github.com/marcello-negri/cmf
  OfficialIn paperpytorch★ 1

Abstract

Studying conditional independence among many variables with few observations is a challenging task. Gaussian Graphical Models (GGMs) tackle this problem by encouraging sparsity in the precision matrix through l_q regularization with q1. However, most GMMs rely on the l_1 norm because the objective is highly non-convex for sub-l_1 pseudo-norms. In the frequentist formulation, the l_1 norm relaxation provides the solution path as a function of the shrinkage parameter . In the Bayesian formulation, sparsity is instead encouraged through a Laplace prior, but posterior inference for different requires repeated runs of expensive Gibbs samplers. Here we propose a general framework for variational inference with matrix-variate Normalizing Flow in GGMs, which unifies the benefits of frequentist and Bayesian frameworks. As a key improvement on previous work, we train with one flow a continuum of sparse regression models jointly for all regularization parameters and all l_q norms, including non-convex sub-l_1 pseudo-norms. Within one model we thus have access to (i) the evolution of the posterior for any and any l_q (pseudo-) norm, (ii) the marginal log-likelihood for model selection, and (iii) the frequentist solution paths through simulated annealing in the MAP limit.

Tasks

Reproductions