Sharp Global Guarantees for Nonconvex Low-Rank Matrix Recovery in the Overparameterized Regime

04/21/2021
by   Richard Y. Zhang, et al.
0

We prove that it is possible for nonconvex low-rank matrix recovery to contain no spurious local minima when the rank of the unknown ground truth r^⋆<r is strictly less than the search rank r, and yet for the claim to be false when r^⋆=r. Under the restricted isometry property (RIP), we prove, for the general overparameterized regime with r^⋆≤ r, that an RIP constant of δ<1/(1+√(r^⋆/r)) is sufficient for the inexistence of spurious local minima, and that δ<1/(1+1/√(r-r^⋆+1)) is necessary due to existence of counterexamples. Without an explicit control over r^⋆≤ r, an RIP constant of δ<1/2 is both necessary and sufficient for the exact recovery of a rank-r ground truth. But if the ground truth is known a priori to have r^⋆=1, then the sharp RIP threshold for exact recovery is improved to δ<1/(1+1/√(r)).

READ FULL TEXT

page 1

page 2

page 3

page 4

research
01/07/2019

Sharp Restricted Isometry Bounds for the Inexistence of Spurious Local Minima in Nonconvex Matrix Recovery

Nonconvex matrix recovery is known to contain no spurious local minima u...
research
05/18/2021

Sharp Restricted Isometry Property Bounds for Low-rank Matrix Recovery Problems with Corrupted Measurements

In this paper, we study a general low-rank matrix recovery problem with ...
research
09/24/2018

Nonconvex Robust Low-rank Matrix Recovery

In this paper we study the problem of recovering a low-rank matrix from ...
research
03/07/2022

Flat minima generalize for low-rank matrix recovery

Empirical evidence suggests that for a variety of overparameterized nonl...
research
05/25/2018

How Much Restricted Isometry is Needed In Nonconvex Matrix Recovery?

When the linear measurements of an instance of low-rank matrix recovery ...
research
03/08/2022

Noisy Low-rank Matrix Optimization: Geometry of Local Minima and Convergence Rate

This paper is concerned with low-rank matrix optimization, which has fou...
research
06/12/2020

How Many Samples is a Good Initial Point Worth?

Given a sufficiently large amount of labeled data, the non-convex low-ra...

Please sign up or login with your details

Forgot password? Click here to reset