Publication | Open Access
Global Optimality of Local Search for Low Rank Matrix Recovery
158
Citations
24
References
2016
Year
Mathematical ProgrammingNumerical AnalysisLow-rank ApproximationSparse RepresentationEngineeringRandom InitializationGlobal OptimalityLocal Search (Optimization)Matrix FactorizationRegularization (Mathematics)Compressive SensingInverse ProblemsComputer ScienceCombinatorial OptimizationApproximation TheorySpurious Local MinimaLocal MinimaIterated Local Search
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}.
| Year | Citations | |
|---|---|---|
Page 1
Page 1