An efficient GPU-Parallel Coordinate Descent Algorithm for Sparse Precision Matrix Estimation via Scaled Lasso

TL;DR

This paper introduces a GPU-parallel coordinate descent algorithm for sparse precision matrix estimation via scaled lasso, significantly improving computational efficiency and demonstrating superior estimation performance in high sparsity scenarios.

Contribution

The paper develops a GPU-parallel coordinate descent algorithm for SPMESL, making it computationally feasible and faster than existing methods like LARS, with extensive numerical validation.

Findings

The proposed algorithm is much faster than LARS for SPMESL.

SPMESL achieves the lowest false discovery rate across various cases.

It performs best in high sparsity levels.

Abstract

The sparse precision matrix plays an essential role in the Gaussian graphical model since a zero off-diagonal element indicates conditional independence of the corresponding two variables given others. In the Gaussian graphical model, many methods have been proposed, and their theoretical properties are given as well. Among these, the sparse precision matrix estimation via scaled lasso (SPMESL) has an attractive feature in which the penalty level is automatically set to achieve the optimal convergence rate under the sparsity and invertibility conditions. Conversely, other methods need to be used in searching for the optimal tuning parameter. Despite such an advantage, the SPMESL has not been widely used due to its expensive computational cost. In this paper, we develop a GPU-parallel coordinate descent (CD) algorithm for the SPMESL and numerically show that the proposed algorithm is much faster than the least angle regression (LARS) tailored to the SPMESL. Several comprehensive numerical studies are conducted to investigate the scalability of the proposed algorithm and the estimation performance of the SPMESL. The results show that the SPMESL has the lowest false discovery rate for all cases and the best performance in the case where the level of the sparsity of the columns is high.