Efficient coordinate-wise leading eigenvector computation

Abstract

We develop and analyze efficient "coordinate-wise" methods for finding the leading eigenvector, where each step involves only a vector-vector product. We establish global convergence with overall runtime guarantees that are at least as good as Lanczos's method and dominate it for slowly decaying spectrum. Our methods are based on combining a shift-and-invert approach with coordinate-wise algorithms for linear regression.

Cited by in corpus (3)

• Coordinate descent full configuration interaction
• Stochastic Variance-Reduced Heavy Ball Power Iteration
• Tensor Canonical Correlation Analysis with Convergence and Statistical Guarantees