Computing the matrix geometric mean: Riemannian versus Euclidean conditioning, implementation techniques, and a Riemannian BFGS method.

DOI10.1002/NLA.2321MaRDI QIDQ5140448zbMATH OpenOpenAlexWikidataFDO

Authors Xinru Yuan, Wen Huang, P.-A. Absil, K. A. Gallivan

Publication date 15 December 2020

Published in Numerical Linear Algebra with Applications (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1002/nla.2321

zbMATH Keywords

Karcher meanlimited-memory Riemannian BFGS methodRiemannian center of massRiemannian quasi-Newton methodssymmetric positive definite matrices

Mathematics Subject Classification ID

Positive matrices and their generalizations; cones of matrices (15B48) Numerical linear algebra (65F99)

Recommendations

• A survey and comparison of contemporary algorithms for computing the matrix geometric mean
• Riemannian metric and geometric mean for positive semidefinite matrices of fixed rank
• A Riemannian submersion-based approach to the Wasserstein barycenter of positive definite matrices
• A Differential Geometric Approach to the Geometric Mean of Symmetric Positive-Definite Matrices
• The Riemannian Barzilai-Borwein method with nonmonotone line search and the matrix geometric mean computation

Cited in

(13)

• Algorithms for approximating means of semi-infinite quasi-Toeplitz matrices
• Computing Riemannian center of mass on Hadamard manifolds
• A survey and comparison of contemporary algorithms for computing the matrix geometric mean
• A limited-memory Riemannian symmetric rank-one trust-region method with a restart strategy
• Geometric means of quasi-Toeplitz matrices
• Averaging symmetric positive-definite matrices
• The Riemannian Barzilai-Borwein method with nonmonotone line search and the matrix geometric mean computation
• Approximated logarithmic maps on Riemannian manifolds and their applications
• Conic geometric optimization on the manifold of positive definite matrices
• Riemannian optimization via Frank-Wolfe methods
• Convergence properties of sequences related to the Ando-Li-Mathias construction and to the weighted cheap mean
• Federated learning on Riemannian manifolds with differential privacy
• Characterizations of locally Lipschitz geodesic pseudolinear functions on Hadamard manifolds

Describes a project that uses

Uses Software

• BLAS
• ROPTLIB
• Matlab
• Matrix Means Toolbox
• LAPACK

This page was built for publication: Computing the matrix geometric mean: Riemannian versus Euclidean conditioning, implementation techniques, and a Riemannian BFGS method.

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5140448)