Non-convex Rank/Sparsity Regularization and Local Minima

Research output: Chapter in Book/Report/Conference proceedingPaper in conference proceedingpeer-review

Abstract

This paper considers the problem of recovering either a low rank matrix or a sparse vector from observations of linear combinations of the vector or matrix elements. Recent methods replace the non-convex regularization with ℓ1 or nuclear norm relaxations. It is well known that this approach recovers near optimal solutions if a so called restricted isometry property (RIP) holds. On the other hand it also has a shrinking bias which can degrade the solution. In this paper we study an alternative non-convex regularization term that does not suffer from this bias. Our main theoretical results show that if a RIP holds then the stationary points are often well separated, in the sense that their differences must be of high cardinality/rank. Thus, with a suitable initial solution the approach is unlikely to fall into a bad local minimum. Our numerical tests show that the approach is likely to converge to a better solution than standard ℓ1/nuclear-norm relaxation even when starting from trivial initializations. In many cases our results can also be used to verify global optimality of our method.

Original languageEnglish
Title of host publicationProceedings - 2017 IEEE International Conference on Computer Vision, ICCV 2017
PublisherIEEE - Institute of Electrical and Electronics Engineers Inc.
Pages332-340
Number of pages9
Volume2017-October
ISBN (Electronic)9781538610329
DOIs
Publication statusPublished - 2017 Dec 22
Event16th IEEE International Conference on Computer Vision, ICCV 2017 - Venice, Italy
Duration: 2017 Oct 222017 Oct 29

Conference

Conference16th IEEE International Conference on Computer Vision, ICCV 2017
Country/TerritoryItaly
CityVenice
Period2017/10/222017/10/29

Subject classification (UKÄ)

  • Computer Vision and Robotics (Autonomous Systems)

Fingerprint

Dive into the research topics of 'Non-convex Rank/Sparsity Regularization and Local Minima'. Together they form a unique fingerprint.

Cite this