60
4

Sparse Gaussian Graphical Models with Discrete Optimization: Computational and Statistical Perspectives

Abstract

We consider the problem of learning a sparse graph underlying an undirected Gaussian graphical model, a key problem in statistical machine learning. Given nn samples from a multivariate Gaussian distribution with pp variables, the goal is to estimate the p×pp \times p inverse covariance matrix (aka precision matrix), assuming it is sparse (i.e., has a few nonzero entries). We propose GraphL0BnB, a new estimator based on an 0\ell_0-penalized version of the pseudolikelihood function, while most earlier approaches are based on the 1\ell_1-relaxation. Our estimator can be formulated as a convex mixed integer program (MIP) which can be difficult to compute at scale using off-the-shelf commercial solvers. To solve the MIP, we propose a custom nonlinear branch-and-bound (BnB) framework that solves node relaxations with tailored first-order methods. As a by-product of our BnB framework, we propose large-scale solvers for obtaining good primal solutions that are of independent interest. We derive novel statistical guarantees (estimation and variable selection) for our estimator and discuss how our approach improves upon existing estimators. Our numerical experiments on real/synthetic datasets suggest that our method can solve, to near-optimality, problem instances with p=104p = 10^4 -- corresponding to a symmetric matrix of size p×pp \times p with p2/2p^2/2 binary variables. We demonstrate the usefulness of GraphL0BnB versus various state-of-the-art approaches on a range of datasets.

View on arXiv
Comments on this paper