pith. sign in

arxiv: 1605.07221 · v2 · pith:IAH4WXC5new · submitted 2016-05-23 · 📊 stat.ML · cs.LG· math.OC

Global Optimality of Local Search for Low Rank Matrix Recovery

classification 📊 stat.ML cs.LGmath.OC
keywords globallocalmatrixmeasurementsminimarecoveryboundclose
0
0 comments X
read the original abstract

We show that there are no spurious local minima in the non-convex factorized parametrization of low-rank matrix recovery from incoherent linear measurements. With noisy measurements we show all local minima are very close to a global optimum. Together with a curvature bound at saddle points, this yields a polynomial time global convergence guarantee for stochastic gradient descent {\em from random initialization}.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.