Matches in SemOpenAlex for { <https://semopenalex.org/work/W2970337468> ?p ?o ?g. }
Showing items 1 to 74 of
74
with 100 items per page.
- W2970337468 endingPage "11400" @default.
- W2970337468 startingPage "11389" @default.
- W2970337468 abstract "Feed-forward neural networks can be understood as a combination of an intermediate representation and a linear hypothesis. While most previous works aim to diversify the representations, we explore the complementary direction by performing an adaptive and data-dependent regularization motivated by the empirical Bayes method. Specifically, we propose to construct a matrix-variate normal prior (on weights) whose covariance matrix has a Kronecker product structure. This structure is designed to capture the correlations in neurons through backpropagation. Under the assumption of this Kronecker factorization, the prior encourages neurons to borrow statistical strength from one another. Hence, it leads to an adaptive and data-dependent regularization when training networks on small datasets. To optimize the model, we present an efficient block coordinate descent algorithm with analytical solutions. Empirically, we demonstrate that the proposed method helps networks converge to local optima with smaller stable ranks and spectral norms. These properties suggest better generalizations and we present empirical results to support this expectation. We also verify the effectiveness of the approach on multiclass classification and multitask regression problems with various network structures. Our code is publicly available at:~url{https://github.com/yaohungt/Adaptive-Regularization-Neural-Network}." @default.
- W2970337468 created "2019-09-05" @default.
- W2970337468 creator A5012188239 @default.
- W2970337468 creator A5012830032 @default.
- W2970337468 creator A5026062552 @default.
- W2970337468 creator A5071983998 @default.
- W2970337468 date "2019-07-14" @default.
- W2970337468 modified "2023-09-24" @default.
- W2970337468 title "Learning Neural Networks with Adaptive Regularization" @default.
- W2970337468 hasPublicationYear "2019" @default.
- W2970337468 type Work @default.
- W2970337468 sameAs 2970337468 @default.
- W2970337468 citedByCount "2" @default.
- W2970337468 countsByYear W29703374682019 @default.
- W2970337468 countsByYear W29703374682020 @default.
- W2970337468 crossrefType "proceedings-article" @default.
- W2970337468 hasAuthorship W2970337468A5012188239 @default.
- W2970337468 hasAuthorship W2970337468A5012830032 @default.
- W2970337468 hasAuthorship W2970337468A5026062552 @default.
- W2970337468 hasAuthorship W2970337468A5071983998 @default.
- W2970337468 hasConcept C11413529 @default.
- W2970337468 hasConcept C119857082 @default.
- W2970337468 hasConcept C121332964 @default.
- W2970337468 hasConcept C154945302 @default.
- W2970337468 hasConcept C157553263 @default.
- W2970337468 hasConcept C158693339 @default.
- W2970337468 hasConcept C185142706 @default.
- W2970337468 hasConcept C2776135515 @default.
- W2970337468 hasConcept C41008148 @default.
- W2970337468 hasConcept C42355184 @default.
- W2970337468 hasConcept C50644808 @default.
- W2970337468 hasConcept C62520636 @default.
- W2970337468 hasConceptScore W2970337468C11413529 @default.
- W2970337468 hasConceptScore W2970337468C119857082 @default.
- W2970337468 hasConceptScore W2970337468C121332964 @default.
- W2970337468 hasConceptScore W2970337468C154945302 @default.
- W2970337468 hasConceptScore W2970337468C157553263 @default.
- W2970337468 hasConceptScore W2970337468C158693339 @default.
- W2970337468 hasConceptScore W2970337468C185142706 @default.
- W2970337468 hasConceptScore W2970337468C2776135515 @default.
- W2970337468 hasConceptScore W2970337468C41008148 @default.
- W2970337468 hasConceptScore W2970337468C42355184 @default.
- W2970337468 hasConceptScore W2970337468C50644808 @default.
- W2970337468 hasConceptScore W2970337468C62520636 @default.
- W2970337468 hasLocation W29703374681 @default.
- W2970337468 hasOpenAccess W2970337468 @default.
- W2970337468 hasPrimaryLocation W29703374681 @default.
- W2970337468 hasRelatedWork W2053559248 @default.
- W2970337468 hasRelatedWork W2396730968 @default.
- W2970337468 hasRelatedWork W2415243320 @default.
- W2970337468 hasRelatedWork W2471962125 @default.
- W2970337468 hasRelatedWork W2885143803 @default.
- W2970337468 hasRelatedWork W2898077722 @default.
- W2970337468 hasRelatedWork W2950297419 @default.
- W2970337468 hasRelatedWork W2963691249 @default.
- W2970337468 hasRelatedWork W2972096249 @default.
- W2970337468 hasRelatedWork W2983189838 @default.
- W2970337468 hasRelatedWork W2983595028 @default.
- W2970337468 hasRelatedWork W3030137591 @default.
- W2970337468 hasRelatedWork W3082378091 @default.
- W2970337468 hasRelatedWork W3093231382 @default.
- W2970337468 hasRelatedWork W3105997200 @default.
- W2970337468 hasRelatedWork W3126265048 @default.
- W2970337468 hasRelatedWork W3200863653 @default.
- W2970337468 hasRelatedWork W192573243 @default.
- W2970337468 hasRelatedWork W2888376184 @default.
- W2970337468 hasRelatedWork W3096979466 @default.
- W2970337468 hasVolume "32" @default.
- W2970337468 isParatext "false" @default.
- W2970337468 isRetracted "false" @default.
- W2970337468 magId "2970337468" @default.
- W2970337468 workType "article" @default.