Pith. sign in

REVIEW 3 cited by

Nonconvex Optimization Meets Low-Rank Matrix Factorization: An Overview

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1809.09573 v3 pith:X3PL5J6K submitted 2018-09-25 cs.LG cs.ITeess.SPmath.ITmath.OCmath.STstat.MLstat.TH

classification cs.LGcs.ITeess.SPmath.ITmath.OCmath.STstat.MLstat.TH
keywords matrixoptimizationalgorithmsnonconvexbeenfactorizationanalysisefficient
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Substantial progress has been made recently on developing provably accurate and efficient algorithms for low-rank matrix factorization via nonconvex optimization. While conventional wisdom often takes a dim view of nonconvex optimization algorithms due to their susceptibility to spurious local minima, simple iterative methods such as gradient descent have been remarkably successful in practice. The theoretical footings, however, had been largely lacking until recently. In this tutorial-style overview, we highlight the important role of statistical models in enabling efficient nonconvex optimization with performance guarantees. We review two contrasting approaches: (1) two-stage algorithms, which consist of a tailored initialization step followed by successive refinement; and (2) global landscape analysis and initialization-free algorithms. Several canonical matrix factorization problems are discussed, including but not limited to matrix sensing, phase retrieval, matrix completion, blind deconvolution, robust principal component analysis, phase synchronization, and joint alignment. Special care is taken to illustrate the key technical insights underlying their analyses. This article serves as a testament that the integrated consideration of optimization and statistics leads to fruitful research findings.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Boundary Defense against Cyber Threat for Power System Operation

    eess.SP 2019-08 conditional novelty 7.0 of 10

    A robust state estimation method guarantees that cyber attack damage stays inside the attacked region whenever per-line vulnerability indices on the boundary are below one.

  2. On The Geometric Analysis of A Quartic-quadratic Optimization Problem under A Spherical Constraint

    math.OC 2019-08 conditional novelty 7.0 of 10

    For the quartic-quadratic sphere problem, the paper characterizes all local minima in the diagonal case, proves strict-saddle properties for extreme beta, and establishes a Kurdyka-Lojasiewicz exponent of 1/4.

  3. A Nonconvex Approach for Exact and Efficient Multichannel Sparse Blind Deconvolution

    eess.SP 2019-08 conditional novelty 6.0 of 10

    For multichannel sparse blind deconvolution, Huber-loss Riemannian gradient descent with random initialization plus an LP-rounding step provably recovers the kernel and sparse signals up to a signed shift, with sample...

Pith tools