Efficient Graph Laplacian Estimation by Proximal Newton
This work addresses graph learning for multivariate statistical modeling, offering incremental improvements in efficiency and accuracy for this domain-specific task.
The paper tackled the problem of learning sparse dependency graphs from data using a Laplacian-constrained Gaussian Markov Random Field, proposing a method that improves computational efficiency and accuracy. Numerical experiments showed advantages in computational complexity and graph learning accuracy compared to existing methods.
The Laplacian-constrained Gaussian Markov Random Field (LGMRF) is a common multivariate statistical model for learning a weighted sparse dependency graph from given data. This graph learning problem can be formulated as a maximum likelihood estimation (MLE) of the precision matrix, subject to Laplacian structural constraints, with a sparsity-inducing penalty term. This paper aims to solve this learning problem accurately and efficiently. First, since the commonly used $\ell_1$-norm penalty is inappropriate in this setting and may lead to a complete graph, we employ the nonconvex minimax concave penalty (MCP), which promotes sparse solutions with lower estimation bias. Second, as opposed to existing first-order methods for this problem, we develop a second-order proximal Newton approach to obtain an efficient solver, utilizing several algorithmic features, such as using Conjugate Gradients, preconditioning, and splitting to active/free sets. Numerical experiments demonstrate the advantages of the proposed method in terms of both computational complexity and graph learning accuracy compared to existing methods.