Inference and uncertainty quantification for noisy matrix completion
Top Cited Papers
Open Access
- 30 October 2019
- journal article
- research article
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 116 (46) , 22931-22937
Abstract
Significance: Matrix completion finds numerous applications in data science, ranging from information retrieval to medical imaging. While substantial progress has been made in designing estimation algorithms, it remains unknown how to perform optimal statistical inference on the unknown matrix given the obtained estimates—a task at the core of modern decision making. We propose procedures to debias the popular convex and nonconvex estimators and derive distributional characterizations for the resulting debiased estimators. This distributional theory enables valid inference on the unknown matrix. Our procedures 1) yield optimal construction of confidence intervals for missing entries and 2) achieve optimal estimation accuracy in a sharp manner.Keywords
All Related Versions
Funding Information
- DOD | United States Navy | Office of Naval Research (N00014-19-1-2120)
- DOD | USAF | AFMC | Air Force Office of Scientific Research (FA9550-19-1-0030)
- DOD | United States Army | RDECOM | Army Research Office (W911NF-18-1-0303)
- National Science Foundation (CCF-1907661)
- National Science Foundation (IIS-1900140)
- National Science Foundation (DMS-1662139)
- National Science Foundation (DMS-1712591)
- HHS | National Institutes of Health (R01-GM072611-12)
This publication has 25 references indexed in Scilit:
- On asymptotically optimal confidence regions and tests for high-dimensional modelsThe Annals of Statistics, 2014
- A significance test for the lassoThe Annals of Statistics, 2014
- Large Covariance Estimation by Thresholding Principal Orthogonal ComplementsJournal of the Royal Statistical Society Series B: Statistical Methodology, 2013
- Confidence Intervals for Low Dimensional Parameters in High Dimensional Linear ModelsJournal of the Royal Statistical Society Series B: Statistical Methodology, 2013
- Nuclear-norm penalization and optimal rates for noisy low-rank matrix completionThe Annals of Statistics, 2011
- Angular synchronization by eigenvectors and semidefinite programmingApplied and Computational Harmonic Analysis, 2011
- Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm MinimizationSIAM Review, 2010
- Exact Matrix Completion via Convex OptimizationFoundations of Computational Mathematics, 2009
- Theory of semidefinite programming for Sensor Network LocalizationMathematical Programming, 2006
- Rank, Trace-Norm and Max-Norm2005