Matches in SemOpenAlex for { <https://semopenalex.org/work/W2970259623> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W2970259623 endingPage "7422" @default.
- W2970259623 startingPage "7411" @default.
- W2970259623 abstract "Efforts to understand the generalization mystery in deep learning have led to the belief that gradient-based optimization induces a form of implicit regularization, a bias towards models of low complexity. We study the implicit regularization of gradient descent over deep linear neural networks for matrix completion and sensing, a model referred to as deep matrix factorization. Our first finding, supported by theory and experiments, is that adding depth to a matrix factorization enhances an implicit tendency towards low-rank solutions, oftentimes leading to more accurate recovery. Secondly, we present theoretical and empirical arguments questioning a nascent view by which implicit regularization in matrix factorization can be captured using simple mathematical norms. Our results point to the possibility that the language of standard regularizers may not be rich enough to fully encompass the implicit regularization brought forth by gradient-based optimization." @default.
- W2970259623 created "2019-09-05" @default.
- W2970259623 creator A5019405863 @default.
- W2970259623 creator A5060414926 @default.
- W2970259623 creator A5062378128 @default.
- W2970259623 creator A5079951047 @default.
- W2970259623 date "2019-01-01" @default.
- W2970259623 modified "2023-09-24" @default.
- W2970259623 title "Implicit Regularization in Deep Matrix Factorization" @default.
- W2970259623 hasPublicationYear "2019" @default.
- W2970259623 type Work @default.
- W2970259623 sameAs 2970259623 @default.
- W2970259623 citedByCount "121" @default.
- W2970259623 countsByYear W29702596232019 @default.
- W2970259623 countsByYear W29702596232020 @default.
- W2970259623 countsByYear W29702596232021 @default.
- W2970259623 countsByYear W29702596232022 @default.
- W2970259623 crossrefType "proceedings-article" @default.
- W2970259623 hasAuthorship W2970259623A5019405863 @default.
- W2970259623 hasAuthorship W2970259623A5060414926 @default.
- W2970259623 hasAuthorship W2970259623A5062378128 @default.
- W2970259623 hasAuthorship W2970259623A5079951047 @default.
- W2970259623 hasConcept C106487976 @default.
- W2970259623 hasConcept C108583219 @default.
- W2970259623 hasConcept C11413529 @default.
- W2970259623 hasConcept C121332964 @default.
- W2970259623 hasConcept C126255220 @default.
- W2970259623 hasConcept C134306372 @default.
- W2970259623 hasConcept C136119220 @default.
- W2970259623 hasConcept C153258448 @default.
- W2970259623 hasConcept C154945302 @default.
- W2970259623 hasConcept C158693339 @default.
- W2970259623 hasConcept C159985019 @default.
- W2970259623 hasConcept C177148314 @default.
- W2970259623 hasConcept C187834632 @default.
- W2970259623 hasConcept C192562407 @default.
- W2970259623 hasConcept C202444582 @default.
- W2970259623 hasConcept C2776135515 @default.
- W2970259623 hasConcept C28826006 @default.
- W2970259623 hasConcept C33923547 @default.
- W2970259623 hasConcept C41008148 @default.
- W2970259623 hasConcept C42355184 @default.
- W2970259623 hasConcept C50644808 @default.
- W2970259623 hasConcept C62520636 @default.
- W2970259623 hasConceptScore W2970259623C106487976 @default.
- W2970259623 hasConceptScore W2970259623C108583219 @default.
- W2970259623 hasConceptScore W2970259623C11413529 @default.
- W2970259623 hasConceptScore W2970259623C121332964 @default.
- W2970259623 hasConceptScore W2970259623C126255220 @default.
- W2970259623 hasConceptScore W2970259623C134306372 @default.
- W2970259623 hasConceptScore W2970259623C136119220 @default.
- W2970259623 hasConceptScore W2970259623C153258448 @default.
- W2970259623 hasConceptScore W2970259623C154945302 @default.
- W2970259623 hasConceptScore W2970259623C158693339 @default.
- W2970259623 hasConceptScore W2970259623C159985019 @default.
- W2970259623 hasConceptScore W2970259623C177148314 @default.
- W2970259623 hasConceptScore W2970259623C187834632 @default.
- W2970259623 hasConceptScore W2970259623C192562407 @default.
- W2970259623 hasConceptScore W2970259623C202444582 @default.
- W2970259623 hasConceptScore W2970259623C2776135515 @default.
- W2970259623 hasConceptScore W2970259623C28826006 @default.
- W2970259623 hasConceptScore W2970259623C33923547 @default.
- W2970259623 hasConceptScore W2970259623C41008148 @default.
- W2970259623 hasConceptScore W2970259623C42355184 @default.
- W2970259623 hasConceptScore W2970259623C50644808 @default.
- W2970259623 hasConceptScore W2970259623C62520636 @default.
- W2970259623 hasLocation W29702596231 @default.
- W2970259623 hasOpenAccess W2970259623 @default.
- W2970259623 hasPrimaryLocation W29702596231 @default.
- W2970259623 hasRelatedWork W2194775991 @default.
- W2970259623 hasRelatedWork W2809090039 @default.
- W2970259623 hasRelatedWork W2886067286 @default.
- W2970259623 hasRelatedWork W2899476926 @default.
- W2970259623 hasRelatedWork W2899748887 @default.
- W2970259623 hasRelatedWork W2911742574 @default.
- W2970259623 hasRelatedWork W2947451636 @default.
- W2970259623 hasRelatedWork W2952204734 @default.
- W2970259623 hasRelatedWork W2962698540 @default.
- W2970259623 hasRelatedWork W2963376662 @default.
- W2970259623 hasRelatedWork W2963518130 @default.
- W2970259623 hasRelatedWork W2963695615 @default.
- W2970259623 hasRelatedWork W2963798163 @default.
- W2970259623 hasRelatedWork W2963826371 @default.
- W2970259623 hasRelatedWork W2963837241 @default.
- W2970259623 hasRelatedWork W2964031251 @default.
- W2970259623 hasRelatedWork W2965772785 @default.
- W2970259623 hasRelatedWork W2970166047 @default.
- W2970259623 hasRelatedWork W2995625976 @default.
- W2970259623 hasRelatedWork W3137695714 @default.
- W2970259623 hasVolume "32" @default.
- W2970259623 isParatext "false" @default.
- W2970259623 isRetracted "false" @default.
- W2970259623 magId "2970259623" @default.
- W2970259623 workType "article" @default.