Corpus ID: 15673678

Low-rank Solutions of Linear Matrix Equations via Procrustes Flow

@inproceedings{Tu2016LowrankSO,
  title={Low-rank Solutions of Linear Matrix Equations via Procrustes Flow},
  author={Stephen Tu and Ross Boczar and Max Simchowitz and M. Soltanolkotabi and B. Recht},
  booktitle={ICML},
  year={2016}
}
  • Stephen Tu, Ross Boczar, +2 authors B. Recht
  • Published in ICML 2016
  • Mathematics, Computer Science
  • In this paper we study the problem of recovering a low-rank matrix from linear measurements. Our algorithm, which we call Procrustes Flow, starts from an initial estimate obtained by a thresholding scheme followed by gradient descent on a non-convex objective. We show that as long as the measurements obey a standard restricted isometry property, our algorithm converges to the unknown matrix at a geometric rate. In the case of Gaussian measurements, such convergence occurs for a $n_1 \times n_2… CONTINUE READING

    Topics from this paper.

    Nonconvex Robust Low-rank Matrix Recovery
    31
    Guarantees of Riemannian Optimization for Low Rank Matrix Recovery
    63
    Fast Low-Rank Matrix Estimation for Ill-Conditioned Matrices
    Fast and Sample Efficient Inductive Matrix Completion via Multi-Phase Procrustes Flow
    10
    Beyond Procrustes: Balancing-free Gradient Descent for Asymmetric Low-Rank Matrix Sensing
    2
    Convergence Analysis for Rectangular Matrix Completion Using Burer-Monteiro Factorization and Gradient Descent
    89

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 113 REFERENCES
    Global Convergence of Stochastic Gradient Descent for Some Non-convex Matrix Problems
    114
    Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm Minimization
    2851
    Low-Rank Optimization on the Cone of Positive Semidefinite Matrices
    205
    Guaranteed Rank Minimization via Singular Value Projection
    431