×

The extrinsic geometry of dynamical systems tracking nonlinear matrix projections. (English) Zbl 07099843

Summary: A generalization of the concepts of extrinsic curvature and Weingarten endomorphism is introduced to study a class of nonlinear maps over embedded matrix manifolds. These (nonlinear) oblique projections generalize (nonlinear) orthogonal projections, i.e., applications mapping a point to its closest neighbor on a matrix manifold. Examples of such maps include the truncated SVD, the polar decomposition, and functions mapping symmetric and nonsymmetric matrices to their linear eigenprojectors. This paper specifically investigates how oblique projections provide their image manifolds with a canonical extrinsic differential structure, over which a generalization of the Weingarten identity is available. By diagonalization of the corresponding Weingarten endomorphism, the manifold principal curvatures are explicitly characterized, which then enables us to (i) derive explicit formulas for the differential of oblique projections and (ii) study the global stability of a governing generic ordinary differential equation (ODE) computing their values. This methodology, exploited for the truncated SVD in [Feppon and Lermusiaux, SIAM J. Matrix Anal. Appl., 39 (2018), pp. 510–538], is generalized to non-Euclidean settings and applied to the four other maps mentioned above and their image manifolds: respectively, the Stiefel, the isospectral, and the Grassmann manifolds and the manifold of fixed rank (nonorthogonal) linear projectors. In all cases studied, the oblique projection of a target matrix is surprisingly the unique stable equilibrium point of the above gradient flow. Three numerical applications concerned with ODEs tracking dominant eigenspaces involving possibly multiple eigenvalues finally showcase the results.

MSC:

65C20 Probabilistic models, generic numerical methods in probability and statistics
53B21 Methods of local Riemannian geometry
65F30 Other matrix algorithms (MSC2010)
15A23 Factorization of matrices
53A07 Higher-dimensional and -codimensional surfaces in Euclidean and related \(n\)-spaces
Full Text: DOI

References:

[1] P.-A. Absil, R. Mahony, and R. Sepulchre, Riemannian geometry of Grassmann manifolds with a view on algorithmic computation, Acta Appl. Math., 80 (2004), pp. 199-220. · Zbl 1052.65048
[2] P.-A. Absil, R. Mahony, and R. Sepulchre, Optimization Algorithms on Matrix Manifolds, Princeton University Press, Princeton, NJ, 2009. · Zbl 1147.65043
[3] P.-A. Absil, R. Mahony, and J. Trumpf, An extrinsic look at the Riemannian Hessian, in Geometric Science of Information, Springer, New York, 2013, pp. 361-368. · Zbl 1323.53014
[4] P.-A. Absil and J. Malick, Projection-like retractions on matrix manifolds, SIAM J. Optim., 22 (2012), pp. 135-158. · Zbl 1248.49055
[5] L. Ambrosio, Geometric Evolution Problems, Distance Function and Viscosity Solutions, Springer, New York, 2000. · Zbl 0956.35002
[6] H. Babaee and T. Sapsis, A minimization principle for the description of modes associated with finite-time instabilities, in Proc. A, 472 (2016), 20150779. · Zbl 1371.34064
[7] K. Bergemann, G. Gottwald, and S. Reich, Ensemble propagation and continuous matrix factorization algorithms, Q. J. Roy. Meteorol. Soc., 135 (2009), pp. 1560-1572.
[8] G. E. Bredon, Topology and Geometry, Vol. 139, Springer Science, New York, 2013.
[9] R. W. Brockett, Dynamical systems that sort lists, diagonalize matrices and solve linear programming problems, in Proceedings of the 27th IEEE Conference on Decision and Control, 1988, IEEE, pp. 799-803.
[10] R. W. Brockett, Dynamical systems that learn subspaces, in Mathematical System Theory, Springer, New York, 1991, pp. 579-592. · Zbl 0758.93071
[11] Y.-C. Chen and L. Wheeler, Derivatives of the stretch and rotation tensors, J. Elasticity, 32 (1993), pp. 175-182. · Zbl 0805.73003
[12] M. T. Chu, A list of matrix flows with applications, Fields Inst. Commun., 3 (1994), pp. 87-97. · Zbl 0815.65094
[13] M. T. Chu and K. R. Driessel, Constructing symmetric nonnegative matrices with prescribed eigenvalues by differential equations, SIAM J. Math. Anal., 22 (1991), pp. 1372-1387. · Zbl 0739.58053
[14] J. Dehaene, Continuous-Time Matrix Algorithms Systolic Algorithms and Adaptive Neural Networks, Ph.D. thesis, K. U. Leuven, Belgium, 1995.
[15] J. Dehaene, M. Moonen, and J. Vandewalle, Analysis of a class of continuous-time algorithms for principal component analysis and subspace tracking, IEEE Trans. Circuits Syst. I Fundam. Theory Appl., 46 (1999), pp. 364-372. · Zbl 0952.94001
[16] P. Deift, T. Nanda, and C. Tomei, Ordinary differential equations and the symmetric eigenvalue problem, SIAM J. Numer. Anal., 20 (1983), pp. 1-22. · Zbl 0526.65032
[17] M. C. Delfour and J.-P. Zolésio, Shapes and Geometries: Metrics, Analysis, Differential Calculus, and Optimization, Vol. 22, SIAM, Philadelphia, 2011. · Zbl 1251.49001
[18] L. Dieci and T. Eirola, On smooth decompositions of matrices, SIAM J. Matrix Anal. Appl., 20 (1999), pp. 800-819. · Zbl 0930.15014
[19] C. Ebenbauer, A dynamical system that computes eigenvalues and diagonalizes matrices with a real spectrum, in Proceedings of the 46th IEEE Conference on Decision and Control, IEEE, 2007, pp. 1704-1709.
[20] A. Edelman, T. A. Arias, and S. T. Smith, The geometry of algorithms with orthogonality constraints, SIAM J. Matrix Anal. Appl., 20 (1998), pp. 303-353. · Zbl 0928.65050
[21] F. Feppon, Riemannian Geometry of Matrix Manifolds for Lagrangian Uncertainty Quantification of Stochastic Fluid Flows, Master’s thesis, MIT, Cambridge, MA, 2017.
[22] F. Feppon and P. F. Lermusiaux, A geometric approach to dynamical model order reduction, SIAM J. Matrix Anal. Appl., 39 (2018), pp. 510-538. · Zbl 1395.65111
[23] F. Feppon and P. F. J. Lermusiaux, Dynamically Orthogonal Numerical Schemes for Efficient Stochastic Advection and Lagrangian Transport, SIAM Rev., 60 (2018), pp. 595-625. · Zbl 1427.65006
[24] N. H. Getz and J. E. Marsden, Dynamic inversion and polar decomposition of matrices, in Proceedings of the 34th IEEE Conference on Decision and Control, Vol. 1, IEEE, 1995, pp. 142-147.
[25] G. H. Golub and C. F. Van Loan, Matrix Computations, Johns Hopkins University Press, Baltimore, MD, 1996, pp. 374-426. · Zbl 0865.65009
[26] B. Gong, Y. Shi, F. Sha, and K. Grauman, Geodesic flow kernel for unsupervised domain adaptation, in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, 2012, pp. 2066-2073.
[27] G. Grioli, Una proprieta di minimo nella cinematica delle deformazioni finite, Boll. Unione Mat. Ital, 2 (1940). · Zbl 0024.22704
[28] N. Guglielmi and C. Lubich, Differential equations for roaming pseudospectra: Paths to extremal points and boundary tracking, SIAM J. Numer. Anal., 49 (2011), pp. 1194-1209. · Zbl 1232.15009
[29] G. Haller, Dynamic rotation and stretch tensors from a dynamic polar decomposition, J. Mech. Phys. Solids, 86 (2016), pp. 70-93. · Zbl 1473.74003
[30] U. Helmke, K. Hüper, and J. Trumpf, Newton’s Method on Grassmann Manifolds, preprint, arXiv:0709.2205, 2007.
[31] U. Helmke and J. B. Moore, Optimization and Dynamical Systems, Springer, New York, 2012.
[32] H. Hendriks and Z. Landsman, Asymptotic behavior of sample mean location for manifolds, Statist. Probab. Lett., 26 (1996), pp. 169-178. · Zbl 0848.62017
[33] H. Hendriks and Z. Landsman, Mean location and sample mean location on manifolds: Asymptotics, tests, confidence regions, J. Multivariate Anal., 67 (1998), pp. 227-243. · Zbl 0941.62069
[34] G. Hori, Isospectral gradient flows for non-symmetric eigenvalue problem, Jpn. J. Ind. Appl. Math., 17 (2000), pp. 27-42. · Zbl 1306.37061
[35] W. Huang, Optimization Algorithms on Riemannian Manifolds with Applications, Ph.D. thesis, Florida State University, 2013.
[36] J. Jost, Riemannian Geometry and Geometric Analysis, Springer, New York, 2008. · Zbl 1143.53001
[37] T. Kato, Perturbation Theory for Linear Operators, Springer, New York, 2013.
[38] E. Kieri, C. Lubich, and H. Walach, Discretized dynamical low-rank approximation in the presence of small singular values, SIAM J. Numer. Anal., 54 (2016), pp. 1020-1038. · Zbl 1336.65119
[39] S. Kobayashi and K. Nomizu, Foundations of Differential Geometry, Vol. 2, Interscience, New York, 1969. · Zbl 0175.48504
[40] O. Koch and C. Lubich, Dynamical low-rank approximation, SIAM J. Matrix Anal. Appl., 29 (2007), pp. 434-454. · Zbl 1145.65031
[41] J. N. Kutz, Data-Driven Modeling & Scientific Computation: Methods for Complex Systems & Big Data, Oxford University Press, Oxford, 2013. · Zbl 1280.65002
[42] P. F. J. Lermusiaux, Error Subspace Data Assimilation Methods for Ocean Field Estimation: Theory, Validation and Applications, Ph.D. thesis, Harvard University, 1997.
[43] P. F. J. Lermusiaux, Estimation and study of mesoscale variability in the Strait of Sicily, Dyn. Atmos. Oceans, 29 (1999), pp. 255-303.
[44] P. F. J. Lermusiaux, Evolving the subspace of the three-dimensional multiscale ocean variability: Massachusetts Bay, J. Marine Syst., 29 (2001), pp. 385-422.
[45] P. F. J. Lermusiaux, Adaptive modeling, adaptive data assimilation and adaptive sampling, Phys. D, 230 (2007), pp. 172-196. · Zbl 1112.86004
[46] P. F. J. Lermusiaux, T. Lolla, P. J. Haley, Jr., K. Yigit, M. P. Ueckermann, T. Sondergaard, and W. G. Leslie, Science of Autonomy: Time-Optimal Path Planning and Adaptive Sampling for Swarms of Ocean Vehicles, in Springer Handbook of Ocean Engineering, T. Curtin, ed., Springer, New York, 2016, pp. 481-498.
[47] P. F. J. Lermusiaux, D. N. Subramani, J. Lin, C. S. Kulkarni, A. Gupta, A. Dutt, T. Lolla, P. J. Haley, Jr., W. H. Ali, C. Mirabito, and S. Jana, A future for intelligent autonomous ocean observing systems, J. Marine Res., 75 (2017), pp. 765-813.
[48] A. S. Lewis and J. Malick, Alternating projections on manifolds, Math. Oper. Res., 33 (2008), pp. 216-234. · Zbl 1163.65040
[49] J. Lin and P. F. J. Lermusiaux, Minimum-Correction Second-Moment Matching: Theory, Algorithms and Applications, to be submitted. · Zbl 1467.65034
[50] T. Lolla, P. J. Haley, Jr., and P. F. J. Lermusiaux, Path planning in multiscale ocean flows: Coordination and dynamic obstacles, Ocean Model., 94 (2015), pp. 46-66.
[51] T. Lolla and P. F. J. Lermusiaux, A Gaussian mixture model smoother for continuous nonlinear stochastic dynamical systems: Theory and schemes, Mon. Wea. Rev., 145 (2017), pp. 2743-2761.
[52] C. Lubich and I. V. Oseledets, A projector-splitting integrator for dynamical low-rank approximation, BIT, 54 (2014), pp. 171-188. · Zbl 1314.65095
[53] A. Machado and I. Salavessa, Grassmannian manifolds as subsets of Euclidean spaces, Differential Geom., 131 (1984). · Zbl 0648.53029
[54] B. Mishra, G. Meyer, S. Bonnabel, and R. Sepulchre, Fixed-rank matrix factorizations and Riemannian low-rank optimization, Comput. Statist., 29 (2014), pp. 591-621. · Zbl 1306.65107
[55] E. Musharbash, F. Nobile, and T. Zhou, Error analysis of the dynamically orthogonal approximation of time dependent random PDEs, SIAM J. Sci. Comput., 37 (2015), pp. A776-A810. · Zbl 1320.65149
[56] P. Neff, J. Lankeit, and A. Madeo, On Grioli’s minimum property and its relation to Cauchy’s polar decomposition, Internat. J. Engrg. Sci., 80 (2014), pp. 209-217. · Zbl 1423.74041
[57] B. Peherstorfer and K. Willcox, Dynamic data-driven reduced-order models, Comput. Methods Appl. Mech. Engrg., 291 (2015), pp. 21-41. · Zbl 1425.65205
[58] M. Przybylska, On matrix differential equations and abstract FG algorithm, Linear Algebra Appl., 346 (2002), pp. 155-175. · Zbl 1042.34021
[59] J. Salençon, Mécanique des Milieux Continus: Concepts Généraux, Vol. 1, Editions Ecole Polytechnique, 2005. · Zbl 1347.74001
[60] T. P. Sapsis and P. F. J. Lermusiaux, Dynamically orthogonal field equations for continuous stochastic dynamical systems, Phys. D, 238 (2009), pp. 2347-2360. · Zbl 1180.37119
[61] B. Schmidt, Conditions on a connection to be a metric connection, Comm. Math. Phys., 29 (1973), pp. 55-59. · Zbl 0243.53019
[62] M. Spivak, A Comprehensive Introduction to Differential Geometry, Publish or Perish, Houston, TX, 1973.
[63] D. N. Subramani and P. F. J. Lermusiaux, Energy-optimal path planning by stochastic dynamically orthogonal level-set optimization, Ocean Model., 100 (2016), pp. 57-77.
[64] C. Truesdell and W. Noll, The non-linear field theories of mechanics, Encyclopedia Phys., 3 (1965). · Zbl 0779.73004
[65] B. Vandereycken, Low-rank matrix completion by Riemannian optimization, SIAM J. Optim., 23 (2013), pp. 1214-1236. · Zbl 1277.15021
[66] F. Yger, M. Berar, G. Gasso, and A. Rakotomamonjy, Adaptive Canonical Correlation Analysis Based on Matrix Manifolds, preprint, arXiv:1206.6453, 2012.
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.