On the matrix square root via geometric optimization
From MaRDI portal
Abstract: This paper is triggered by the preprint "emph{Computing Matrix Squareroot via Non Convex Local Search}" by Jain et al. ( extit{ extcolor{blue}{arXiv:1507.05854}}), which analyzes gradient-descent for computing the square root of a positive definite matrix. Contrary to claims of~citet{jain2015}, our experiments reveal that Newton-like methods compute matrix square roots rapidly and reliably, even for highly ill-conditioned matrices and without requiring commutativity. We observe that gradient-descent converges very slowly primarily due to tiny step-sizes and ill-conditioning. We derive an alternative first-order method based on geodesic convexity: our method admits a transparent convergence analysis ( page), attains linear rate, and displays reliable convergence even for rank deficient problems. Though superior to gradient-descent, ultimately our method is also outperformed by a well-known scaled Newton method. Nevertheless, the primary value of our work is its conceptual value: it shows that for deriving gradient based methods for the matrix square root, emph{the manifold geometric view of positive definite matrices can be much more advantageous than the Euclidean view}.
Recommendations
- The Matrix Square Root from a New Functional Perspective: Theoretical Results and Computational Issues
- Newton's Method for the Matrix Square Root
- A note on computing the matrix square root
- Computing the Square Root of a Low-Rank Perturbation of the Scaled Identity Matrix
- Newton's method for the matrix nonsingular square root
Cites work
- A general framework for extending means to higher orders
- A survey and comparison of contemporary algorithms for computing the matrix geometric mean
- Backward stability of iterations for computing the polar decomposition
- Computing A^\alpha, \log(A), and Related Matrix Functions by Contour Integrals
- Computing the Karcher mean of symmetric positive definite matrices
- Conic geometric optimization on the manifold of positive definite matrices
- Geometric means
- scientific article; zbMATH DE number 3730780 (Why is no real title available?)
- Invariant metrics, contractions and nonlinear matrix equations
- Ladder networks, fixpoints, and the geometric mean
- Means of positive linear operators
- Newton's Method for the Matrix Square Root
- On Certain Contraction Mappings in a Partially Ordered Vector Space
- Optimizing Halley's Iteration for Computing the Matrix Polar Decomposition
- Palindromic matrix polynomials, matrix functions and integral representations
- Positive definite matrices
- Positive definite matrices and the S-divergence
- The geometric mean of two matrices from a computational viewpoint.
- Unified Framework to Regularized Covariance Estimation in Scaled Gaussian Models
Cited in
(6)- Metrics induced by Jensen-Shannon and related divergences on positive definite matrices
- scientific article; zbMATH DE number 7024474 (Why is no real title available?)
- scientific article; zbMATH DE number 6311330 (Why is no real title available?)
- Computing the Square Root of a Low-Rank Perturbation of the Scaled Identity Matrix
- A new ZNN model for finding discrete time-variant matrix square root: from model design to parameter analysis
- Scaled fixed point algorithm for computing the matrix square root
This page was built for publication: On the matrix square root via geometric optimization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3186698)
