Matches in SemOpenAlex for { <https://semopenalex.org/work/W2963746765> ?p ?o ?g. }
- W2963746765 abstract "Abstract: Hessian-free (HF) optimization has been successfully used for training deep autoencoders and recurrent networks. HF uses the conjugate gradient algorithm to construct update directions through curvature-vector products that can be computed on the same order of time as gradients. In this paper we exploit this property and study stochastic HF with gradient and curvature mini-batches independent of the dataset size. We modify Martens' HF for these settings and integrate dropout, a method for preventing co-adaptation of feature detectors, to guard against overfitting. Stochastic Hessian-free optimization gives an intermediary between SGD and HF that achieves competitive performance on both classification and deep autoencoder experiments." @default.
- W2963746765 created "2019-07-30" @default.
- W2963746765 creator A5055166817 @default.
- W2963746765 date "2013-01-16" @default.
- W2963746765 modified "2023-09-24" @default.
- W2963746765 title "Training Neural Networks with Stochastic Hessian-Free Optimization" @default.
- W2963746765 cites W1408639475 @default.
- W2963746765 cites W1533861849 @default.
- W2963746765 cites W1544532352 @default.
- W2963746765 cites W1889268436 @default.
- W2963746765 cites W1904365287 @default.
- W2963746765 cites W196214544 @default.
- W2963746765 cites W196761320 @default.
- W2963746765 cites W2006903949 @default.
- W2963746765 cites W2025768430 @default.
- W2963746765 cites W2061570747 @default.
- W2963746765 cites W2102017903 @default.
- W2963746765 cites W2110798204 @default.
- W2963746765 cites W2112104211 @default.
- W2963746765 cites W2113651538 @default.
- W2963746765 cites W2130984546 @default.
- W2963746765 cites W2135624716 @default.
- W2963746765 cites W2137515395 @default.
- W2963746765 cites W2156387975 @default.
- W2963746765 cites W2163605009 @default.
- W2963746765 cites W2164273299 @default.
- W2963746765 cites W2168231600 @default.
- W2963746765 cites W2252143850 @default.
- W2963746765 cites W2405601855 @default.
- W2963746765 cites W2914484425 @default.
- W2963746765 cites W2949427968 @default.
- W2963746765 cites W2950351588 @default.
- W2963746765 cites W2951282416 @default.
- W2963746765 cites W2951650375 @default.
- W2963746765 cites W2962936867 @default.
- W2963746765 cites W71081281 @default.
- W2963746765 hasPublicationYear "2013" @default.
- W2963746765 type Work @default.
- W2963746765 sameAs 2963746765 @default.
- W2963746765 citedByCount "5" @default.
- W2963746765 countsByYear W29637467652014 @default.
- W2963746765 countsByYear W29637467652015 @default.
- W2963746765 countsByYear W29637467652020 @default.
- W2963746765 crossrefType "proceedings-article" @default.
- W2963746765 hasAuthorship W2963746765A5055166817 @default.
- W2963746765 hasConcept C101738243 @default.
- W2963746765 hasConcept C11413529 @default.
- W2963746765 hasConcept C119857082 @default.
- W2963746765 hasConcept C126255220 @default.
- W2963746765 hasConcept C153180895 @default.
- W2963746765 hasConcept C154945302 @default.
- W2963746765 hasConcept C194387892 @default.
- W2963746765 hasConcept C203616005 @default.
- W2963746765 hasConcept C206688291 @default.
- W2963746765 hasConcept C22019652 @default.
- W2963746765 hasConcept C2776145597 @default.
- W2963746765 hasConcept C28826006 @default.
- W2963746765 hasConcept C33923547 @default.
- W2963746765 hasConcept C41008148 @default.
- W2963746765 hasConcept C50644808 @default.
- W2963746765 hasConcept C81184566 @default.
- W2963746765 hasConceptScore W2963746765C101738243 @default.
- W2963746765 hasConceptScore W2963746765C11413529 @default.
- W2963746765 hasConceptScore W2963746765C119857082 @default.
- W2963746765 hasConceptScore W2963746765C126255220 @default.
- W2963746765 hasConceptScore W2963746765C153180895 @default.
- W2963746765 hasConceptScore W2963746765C154945302 @default.
- W2963746765 hasConceptScore W2963746765C194387892 @default.
- W2963746765 hasConceptScore W2963746765C203616005 @default.
- W2963746765 hasConceptScore W2963746765C206688291 @default.
- W2963746765 hasConceptScore W2963746765C22019652 @default.
- W2963746765 hasConceptScore W2963746765C2776145597 @default.
- W2963746765 hasConceptScore W2963746765C28826006 @default.
- W2963746765 hasConceptScore W2963746765C33923547 @default.
- W2963746765 hasConceptScore W2963746765C41008148 @default.
- W2963746765 hasConceptScore W2963746765C50644808 @default.
- W2963746765 hasConceptScore W2963746765C81184566 @default.
- W2963746765 hasLocation W29637467651 @default.
- W2963746765 hasOpenAccess W2963746765 @default.
- W2963746765 hasPrimaryLocation W29637467651 @default.
- W2963746765 hasRelatedWork W1576933279 @default.
- W2963746765 hasRelatedWork W1932057668 @default.
- W2963746765 hasRelatedWork W196761320 @default.
- W2963746765 hasRelatedWork W1970789124 @default.
- W2963746765 hasRelatedWork W2102865215 @default.
- W2963746765 hasRelatedWork W2180776362 @default.
- W2963746765 hasRelatedWork W2403040875 @default.
- W2963746765 hasRelatedWork W2555215491 @default.
- W2963746765 hasRelatedWork W2560283133 @default.
- W2963746765 hasRelatedWork W2737818580 @default.
- W2963746765 hasRelatedWork W2777827707 @default.
- W2963746765 hasRelatedWork W2785921341 @default.
- W2963746765 hasRelatedWork W2786872741 @default.
- W2963746765 hasRelatedWork W2905519666 @default.
- W2963746765 hasRelatedWork W2955553835 @default.
- W2963746765 hasRelatedWork W2962780420 @default.
- W2963746765 hasRelatedWork W3001737550 @default.
- W2963746765 hasRelatedWork W3097919408 @default.
- W2963746765 hasRelatedWork W3134477793 @default.
- W2963746765 hasRelatedWork W3136329410 @default.