Abstract
Retractions are a prevalent tool in Riemannian optimization that provides a way to smoothly select a curve on a manifold with given initial position and velocity. We review and propose several retractions on the manifold \({\mathcal {M}}_r\) of rank-\(r\) \(m\times n\) matrices. With the exception of the exponential retraction (for the embedded geometry), which is clearly the least efficient choice, the retractions considered do not differ much in terms of run time and flop count. However, considerable differences are observed according to properties such as domain of definition, boundedness, first/second-order property, and symmetry.
Similar content being viewed by others
Notes
The Matlab code that we used to generate the tables of this paper is available from http://sites.uclouvain.be/absil/2013.04.
References
Absil, P.-A., Amodei, L., Meyer, G.: Two Newton methods on the manifold of fixed-rank matrices endowed with riemannian quotient geometries. Comput. Stat. 29(3–4), 569–590 (2014). doi:10.1007/s00180-013-0441-6
Absil, P.-A., Baker, C.G., Gallivan, K.A.: Trust-region methods on Riemannian manifolds. Found. Comput. Math. 7(3), 303–330 (2007). doi:10.1007/s10208-005-0179-9
Adler, R.L., Dedieu, J.-P., Margulies, J.Y., Martens, M., Shub, Mike: Newton’s method on Riemannian manifolds and a geometric model for the human spine. IMA J. Numer. Anal. 22(3), 359–390 (2002). doi:10.1093/imanum/22.3.359
Absil, P.A., Malick, J.: Projection-like retractions on matrix manifolds. SIAM J. Optim. 22(1), 135–158 (2012). doi:10.1137/100802529
Absil, P.-A., Mahony, R., Sepulchre. R.: Optimization Algorithms on Matrix Manifolds. Princeton University Press, Princeton, NJ (2008). http://press.princeton.edu/titles/8586.html
Boumal, N., Mishra, B., Absil, P.-A., Sepulchre, R.: Manopt, a matlab toolbox for optimization on manifolds. J. Mach. Learn. Res. 15, 1455–1459 (2014). http://jmlr.org/papers/v15/boumal14a.html
Boothby, W.M.: An Introduction to Differentiable Manifolds and Riemannian Geometry, Revised Second Edition. Academic Press, London (2003)
Golub, G.H., Van Loan, C.F.: Matrix Computations. Johns Hopkins Studies in the Mathematical Sciences, 3rd edn. Johns Hopkins University Press, Baltimore, MD (1996)
Koch, O., Lubich, C.: Dynamical low-rank approximation. SIAM J. Matrix Anal. Appl. 29(2), 434–454 (2007). doi:10.1137/050639703
Kressner, Daniel, Steinlechner, Michael, Vandereycken, Bart: Low-rank tensor completion by Riemannian optimization. BIT Numer. Math. 54(2), 447–468 (2014). doi:10.1007/s10543-013-0455-z
Lee, J.M.: Introduction to Smooth Manifolds, Volume 218 of Graduate Texts in Mathematics. Springer, New York (2003)
Lubich, C., Oseledets, I.V.: A projector-splitting integrator for dynamical low-rank approximation (2013) arXiv:1301.1058v2
Luenberger, D.G.: The gradient projection method along geodesics. Manag. Sci. 18, 620–631 (1972)
Mishra, B., Meyer, G., Bach, F., Sepulchre, R.: Low-rank optimization with trace norm penalty. SIAM J. Optim. 23(4), 2124–2149 (2013). doi:10.1137/110859646
Mishra, B., Meyer, G., Bonnabel, S., Sepulchre, R.: Fixed-rank matrix factorizations and Riemannian low-rank optimization (2013). arXiv:1209.0430v2
Mishra, B., Sepulchre, R.: R3MC: a Riemannian three-factor algorithm for low-rank matrix completion, 2014. Accepted for publication in the proceedings of the 53rd IEEE Conference on Decision and Control (2014). arXiv:1306.2672v2
Rosen, J.B.: The gradient projection method for nonlinear programming. II. Nonlinear constraints. J. Soc. Ind. Appl. Math. 9, 514–532 (1961)
Ring, W., Wirth, B.: Optimization methods on Riemannian manifolds and their application to shape space. SIAM J. Optim. 22(2), 596–627 (2012). doi:10.1137/11082885X
Shub, M.: Some remarks on dynamical systems and numerical analysis. In L. Lara-Carrero and J. Lewowicz, (eds), Proceedings of the VII ELAM. Equinoccio, U. Simón Bolívar, Caracas, pp. 69–92 (1986)
Shalit, U., Weinshall, D., Chechik, G.: Online learning in the embedded manifold of low-rank matrices. J. Mach. Learn. Res. 13, 429–458 (2013)
Vandereycken, B.: Low-rank matrix completion by Riemannian optimization. SIAM J. Optim. 23(2), 1214–1236 (2013). doi:10.1137/110845768
Acknowledgments
We are grateful to the anonymous referees and to Bart Vandereycken for several useful comments on the first version of this paper. This work was financially supported by the Belgian FRFC (Fonds de la Recherche Fondamentale Collective). The work of I.O. was supported by Russian Science Foundation Grant 14-11-00659
Author information
Authors and Affiliations
Corresponding author
Additional information
This paper presents research results of the Belgian Network DYSCO (Dynamical Systems, Control, and Optimization), funded by the Interuniversity Attraction Poles Programme initiated by the Belgian Science Policy Office.
Rights and permissions
About this article
Cite this article
Absil, PA., Oseledets, I.V. Low-rank retractions: a survey and new results. Comput Optim Appl 62, 5–29 (2015). https://doi.org/10.1007/s10589-014-9714-4
Received:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10589-014-9714-4