Paper ID: 2304.03641

A Block Coordinate Descent Method for Nonsmooth Composite Optimization under Orthogonality Constraints

Ganzhao Yuan

Nonsmooth composite optimization with orthogonality constraints is crucial in statistical learning and data science, but it presents challenges due to its nonsmooth objective and computationally expensive, non-convex constraints. In this paper, we propose a new approach called \textbf{OBCD}, which leverages Block Coordinate Descent (BCD) to address these challenges. \textbf{OBCD} is a feasible method with a small computational footprint. In each iteration, it updates $k$ rows of the solution matrix, where $k \geq 2$, while globally solving a small nonsmooth optimization problem under orthogonality constraints. We prove that \textbf{OBCD} converges to block-$k$ stationary points, which offer stronger optimality than standard critical points. Notably, \textbf{OBCD} is the first greedy descent method with monotonicity for this problem class. Under the Kurdyka-Lojasiewicz (KL) inequality, we establish strong limit-point convergence. We also extend \textbf{OBCD} with breakpoint searching methods for subproblem solving and greedy strategies for working set selection. Comprehensive experiments demonstrate the superior performance of our approach across various tasks.

Submitted: Apr 7, 2023