Matches in SemOpenAlex for { <https://semopenalex.org/work/W2791397041> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W2791397041 abstract "In this paper, we revisit the recurrent back-propagation (RBP) algorithm, discuss the conditions under which it applies as well as how to satisfy them in deep neural networks. We show that RBP can be unstable and propose two variants based on conjugate gradient on the normal equations (CG-RBP) and Neumann series (Neumann-RBP). We further investigate the relationship between Neumann-RBP and back propagation through time (BPTT) and its truncated version (TBPTT). Our Neumann-RBP has the same time complexity as TBPTT but only requires constant memory, whereas TBPTT's memory cost scales linearly with the number of truncation steps. We examine all RBP variants along with BPTT and TBPTT in three different application domains: associative memory with continuous Hopfield networks, document classification in citation networks using graph neural networks and hyperparameter optimization for fully connected networks. All experiments demonstrate that RBPs, especially the Neumann-RBP variant, are efficient and effective for optimizing convergent recurrent neural networks. Code is released at: url{this https URL}." @default.
- W2791397041 created "2018-03-29" @default.
- W2791397041 creator A5000111344 @default.
- W2791397041 creator A5048686150 @default.
- W2791397041 creator A5053007211 @default.
- W2791397041 creator A5057055829 @default.
- W2791397041 creator A5058557954 @default.
- W2791397041 creator A5060590657 @default.
- W2791397041 creator A5063688088 @default.
- W2791397041 creator A5066457130 @default.
- W2791397041 date "2018-03-16" @default.
- W2791397041 modified "2023-10-17" @default.
- W2791397041 title "Reviving and Improving Recurrent Back-Propagation" @default.
- W2791397041 hasPublicationYear "2018" @default.
- W2791397041 type Work @default.
- W2791397041 sameAs 2791397041 @default.
- W2791397041 citedByCount "4" @default.
- W2791397041 countsByYear W27913970412019 @default.
- W2791397041 countsByYear W27913970412020 @default.
- W2791397041 countsByYear W27913970412021 @default.
- W2791397041 crossrefType "posted-content" @default.
- W2791397041 hasAuthorship W2791397041A5000111344 @default.
- W2791397041 hasAuthorship W2791397041A5048686150 @default.
- W2791397041 hasAuthorship W2791397041A5053007211 @default.
- W2791397041 hasAuthorship W2791397041A5057055829 @default.
- W2791397041 hasAuthorship W2791397041A5058557954 @default.
- W2791397041 hasAuthorship W2791397041A5060590657 @default.
- W2791397041 hasAuthorship W2791397041A5063688088 @default.
- W2791397041 hasAuthorship W2791397041A5066457130 @default.
- W2791397041 hasConcept C108583219 @default.
- W2791397041 hasConcept C111919701 @default.
- W2791397041 hasConcept C11413529 @default.
- W2791397041 hasConcept C147168706 @default.
- W2791397041 hasConcept C154945302 @default.
- W2791397041 hasConcept C177264268 @default.
- W2791397041 hasConcept C199360897 @default.
- W2791397041 hasConcept C2776760102 @default.
- W2791397041 hasConcept C28826006 @default.
- W2791397041 hasConcept C33923547 @default.
- W2791397041 hasConcept C41008148 @default.
- W2791397041 hasConcept C46421273 @default.
- W2791397041 hasConcept C50644808 @default.
- W2791397041 hasConcept C53442348 @default.
- W2791397041 hasConcept C80444323 @default.
- W2791397041 hasConcept C80469333 @default.
- W2791397041 hasConcept C81184566 @default.
- W2791397041 hasConceptScore W2791397041C108583219 @default.
- W2791397041 hasConceptScore W2791397041C111919701 @default.
- W2791397041 hasConceptScore W2791397041C11413529 @default.
- W2791397041 hasConceptScore W2791397041C147168706 @default.
- W2791397041 hasConceptScore W2791397041C154945302 @default.
- W2791397041 hasConceptScore W2791397041C177264268 @default.
- W2791397041 hasConceptScore W2791397041C199360897 @default.
- W2791397041 hasConceptScore W2791397041C2776760102 @default.
- W2791397041 hasConceptScore W2791397041C28826006 @default.
- W2791397041 hasConceptScore W2791397041C33923547 @default.
- W2791397041 hasConceptScore W2791397041C41008148 @default.
- W2791397041 hasConceptScore W2791397041C46421273 @default.
- W2791397041 hasConceptScore W2791397041C50644808 @default.
- W2791397041 hasConceptScore W2791397041C53442348 @default.
- W2791397041 hasConceptScore W2791397041C80444323 @default.
- W2791397041 hasConceptScore W2791397041C80469333 @default.
- W2791397041 hasConceptScore W2791397041C81184566 @default.
- W2791397041 hasLocation W27913970411 @default.
- W2791397041 hasOpenAccess W2791397041 @default.
- W2791397041 hasPrimaryLocation W27913970411 @default.
- W2791397041 hasRelatedWork W1839868949 @default.
- W2791397041 hasRelatedWork W1888860385 @default.
- W2791397041 hasRelatedWork W2092582833 @default.
- W2791397041 hasRelatedWork W2172308869 @default.
- W2791397041 hasRelatedWork W2581506979 @default.
- W2791397041 hasRelatedWork W2619034422 @default.
- W2791397041 hasRelatedWork W2883819993 @default.
- W2791397041 hasRelatedWork W2952714364 @default.
- W2791397041 hasRelatedWork W2962849464 @default.
- W2791397041 hasRelatedWork W2964346435 @default.
- W2791397041 hasRelatedWork W2997081737 @default.
- W2791397041 hasRelatedWork W3007364240 @default.
- W2791397041 hasRelatedWork W3012276865 @default.
- W2791397041 hasRelatedWork W3020928088 @default.
- W2791397041 hasRelatedWork W3035204268 @default.
- W2791397041 hasRelatedWork W3038074040 @default.
- W2791397041 hasRelatedWork W3094346332 @default.
- W2791397041 hasRelatedWork W3102087754 @default.
- W2791397041 hasRelatedWork W3134483940 @default.
- W2791397041 hasRelatedWork W3213601708 @default.
- W2791397041 isParatext "false" @default.
- W2791397041 isRetracted "false" @default.
- W2791397041 magId "2791397041" @default.
- W2791397041 workType "article" @default.