Matches in SemOpenAlex for { <https://semopenalex.org/work/W3213664974> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W3213664974 abstract "We study the robust recovery of a low-rank matrix from sparsely and grossly corrupted Gaussian measurements, with no prior knowledge on the intrinsic rank. We consider the robust matrix factorization approach. We employ a robust $ell_1$ loss function and deal with the challenge of the unknown rank by using an overspecified factored representation of the matrix variable. We then solve the associated nonconvex nonsmooth problem using a subgradient method with diminishing stepsizes. We show that under a regularity condition on the sensing matrices and corruption, which we call restricted direction preserving property (RDPP), even with rank overspecified, the subgradient method converges to the exact low-rank solution at a sublinear rate. Moreover, our result is more general in the sense that it automatically speeds up to a linear rate once the factor rank matches the unknown rank. On the other hand, we show that the RDPP condition holds under generic settings, such as Gaussian measurements under independent or adversarial sparse corruptions, where the result could be of independent interest. Both the exact recovery and the convergence rate of the proposed subgradient method are numerically verified in the overspecified regime. Moreover, our experiment further shows that our particular design of diminishing stepsize effectively prevents overfitting for robust recovery under overparameterized models, such as robust matrix sensing and learning robust deep image prior. This regularization effect is worth further investigation." @default.
- W3213664974 created "2021-11-22" @default.
- W3213664974 creator A5011989964 @default.
- W3213664974 creator A5019924950 @default.
- W3213664974 creator A5036790472 @default.
- W3213664974 creator A5049904411 @default.
- W3213664974 creator A5060774236 @default.
- W3213664974 date "2021-09-23" @default.
- W3213664974 modified "2023-09-27" @default.
- W3213664974 title "Rank Overspecified Robust Matrix Recovery: Subgradient Method and Exact Recovery" @default.
- W3213664974 doi "https://doi.org/10.48550/arxiv.2109.11154" @default.
- W3213664974 hasPublicationYear "2021" @default.
- W3213664974 type Work @default.
- W3213664974 sameAs 3213664974 @default.
- W3213664974 citedByCount "0" @default.
- W3213664974 crossrefType "posted-content" @default.
- W3213664974 hasAuthorship W3213664974A5011989964 @default.
- W3213664974 hasAuthorship W3213664974A5019924950 @default.
- W3213664974 hasAuthorship W3213664974A5036790472 @default.
- W3213664974 hasAuthorship W3213664974A5049904411 @default.
- W3213664974 hasAuthorship W3213664974A5060774236 @default.
- W3213664974 hasBestOaLocation W32136649741 @default.
- W3213664974 hasConcept C106487976 @default.
- W3213664974 hasConcept C11413529 @default.
- W3213664974 hasConcept C114614502 @default.
- W3213664974 hasConcept C121332964 @default.
- W3213664974 hasConcept C126255220 @default.
- W3213664974 hasConcept C127162648 @default.
- W3213664974 hasConcept C154945302 @default.
- W3213664974 hasConcept C158968445 @default.
- W3213664974 hasConcept C159985019 @default.
- W3213664974 hasConcept C163716315 @default.
- W3213664974 hasConcept C164226766 @default.
- W3213664974 hasConcept C192562407 @default.
- W3213664974 hasConcept C2776135515 @default.
- W3213664974 hasConcept C2778459887 @default.
- W3213664974 hasConcept C28826006 @default.
- W3213664974 hasConcept C31258907 @default.
- W3213664974 hasConcept C33923547 @default.
- W3213664974 hasConcept C41008148 @default.
- W3213664974 hasConcept C57869625 @default.
- W3213664974 hasConcept C62520636 @default.
- W3213664974 hasConceptScore W3213664974C106487976 @default.
- W3213664974 hasConceptScore W3213664974C11413529 @default.
- W3213664974 hasConceptScore W3213664974C114614502 @default.
- W3213664974 hasConceptScore W3213664974C121332964 @default.
- W3213664974 hasConceptScore W3213664974C126255220 @default.
- W3213664974 hasConceptScore W3213664974C127162648 @default.
- W3213664974 hasConceptScore W3213664974C154945302 @default.
- W3213664974 hasConceptScore W3213664974C158968445 @default.
- W3213664974 hasConceptScore W3213664974C159985019 @default.
- W3213664974 hasConceptScore W3213664974C163716315 @default.
- W3213664974 hasConceptScore W3213664974C164226766 @default.
- W3213664974 hasConceptScore W3213664974C192562407 @default.
- W3213664974 hasConceptScore W3213664974C2776135515 @default.
- W3213664974 hasConceptScore W3213664974C2778459887 @default.
- W3213664974 hasConceptScore W3213664974C28826006 @default.
- W3213664974 hasConceptScore W3213664974C31258907 @default.
- W3213664974 hasConceptScore W3213664974C33923547 @default.
- W3213664974 hasConceptScore W3213664974C41008148 @default.
- W3213664974 hasConceptScore W3213664974C57869625 @default.
- W3213664974 hasConceptScore W3213664974C62520636 @default.
- W3213664974 hasLocation W32136649741 @default.
- W3213664974 hasOpenAccess W3213664974 @default.
- W3213664974 hasPrimaryLocation W32136649741 @default.
- W3213664974 hasRelatedWork W1712752320 @default.
- W3213664974 hasRelatedWork W2054519467 @default.
- W3213664974 hasRelatedWork W2108753466 @default.
- W3213664974 hasRelatedWork W2263999648 @default.
- W3213664974 hasRelatedWork W2788210652 @default.
- W3213664974 hasRelatedWork W2953159110 @default.
- W3213664974 hasRelatedWork W2972830027 @default.
- W3213664974 hasRelatedWork W2976194898 @default.
- W3213664974 hasRelatedWork W3093503721 @default.
- W3213664974 hasRelatedWork W4301205717 @default.
- W3213664974 isParatext "false" @default.
- W3213664974 isRetracted "false" @default.
- W3213664974 magId "3213664974" @default.
- W3213664974 workType "article" @default.