Nonconvex Matrix Factorization from Rank-One Measurements
@inproceedings{Li2019NonconvexMF, title={Nonconvex Matrix Factorization from Rank-One Measurements}, author={Yuanxin Li and C. Ma and Y. Chen and Yuejie Chi}, booktitle={AISTATS}, year={2019} }
We consider the problem of recovering low-rank matrices from random rank-one measurements, which spans numerous applications including covariance sketching, phase retrieval, quantum state tomography, and learning shallow polynomial neural networks, among others. Our approach is to directly estimate the low-rank factor by minimizing a nonconvex quadratic loss function via vanilla gradient descent, following a tailored spectral initialization. When the true rank is small, this algorithm is… CONTINUE READING
Figures, Tables, and Topics from this paper
32 Citations
Beyond Procrustes: Balancing-free Gradient Descent for Asymmetric Low-Rank Matrix Sensing
- Computer Science, Engineering
- 2019 53rd Asilomar Conference on Signals, Systems, and Computers
- 2019
- 3
- PDF
Nonconvex Low-Rank Matrix Recovery with Arbitrary Outliers via Median-Truncated Gradient Descent
- Mathematics, Computer Science
- ArXiv
- 2017
- 16
- PDF
Low-Rank Matrix Recovery with Scaled Subgradient Methods: Fast and Robust Convergence Without the Condition Number
- Computer Science, Engineering
- ArXiv
- 2020
- 3
- PDF
Accelerating Ill-Conditioned Low-Rank Matrix Estimation via Scaled Gradient Descent
- Computer Science, Engineering
- ArXiv
- 2020
- 9
- PDF
Noisy Matrix Completion: Understanding Statistical Guarantees for Convex Relaxation via Nonconvex Optimization
- Computer Science, Mathematics
- SIAM J. Optim.
- 2020
- 39
- PDF
Nonconvex Optimization Meets Low-Rank Matrix Factorization: An Overview
- Computer Science, Engineering
- IEEE Transactions on Signal Processing
- 2019
- 146
- PDF
Low-rank matrix recovery with composite optimization: good conditioning and rapid convergence
- Mathematics, Computer Science
- ArXiv
- 2019
- 23
- PDF
Median-Truncated Gradient Descent: A Robust and Scalable Nonconvex Approach for Signal Estimation
- Computer Science
- 2019
- 3
- PDF
Harnessing Structures in Big Data via Guaranteed Low-Rank Matrix Estimation: Recent Theory and Fast Algorithms via Convex and Nonconvex Optimization
- Computer Science, Mathematics
- IEEE Signal Processing Magazine
- 2018
- 68
- PDF
References
SHOWING 1-10 OF 60 REFERENCES
Low-Rank Positive Semidefinite Matrix Recovery From Corrupted Rank-One Measurements
- Mathematics, Computer Science
- IEEE Transactions on Signal Processing
- 2017
- 24
- PDF
Noisy Matrix Completion: Understanding Statistical Guarantees for Convex Relaxation via Nonconvex Optimization
- Computer Science, Mathematics
- SIAM J. Optim.
- 2020
- 39
- PDF
Nonconvex Optimization Meets Low-Rank Matrix Factorization: An Overview
- Computer Science, Engineering
- IEEE Transactions on Signal Processing
- 2019
- 146
- PDF
Fast low-rank estimation by projected gradient descent: General statistical and algorithmic guarantees
- Mathematics, Computer Science
- ArXiv
- 2015
- 217
- PDF
Tight Oracle Inequalities for Low-Rank Matrix Recovery From a Minimal Number of Noisy Random Measurements
- Mathematics, Computer Science
- IEEE Transactions on Information Theory
- 2011
- 460
Kaczmarz Method for Solving Quadratic Equations
- Mathematics, Computer Science
- IEEE Signal Processing Letters
- 2016
- 36
- PDF
Implicit Regularization in Nonconvex Statistical Estimation: Gradient Descent Converges Linearly for Phase Retrieval, Matrix Completion, and Blind Deconvolution
- Mathematics, Computer Science
- Found. Comput. Math.
- 2020
- 136
- PDF
Exact and Stable Covariance Estimation From Quadratic Sampling via Convex Programming
- Mathematics, Computer Science
- IEEE Transactions on Information Theory
- 2015
- 167
- PDF