Matches in SemOpenAlex for { <https://semopenalex.org/work/W2758103492> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W2758103492 abstract "Neural language models do not scale well when the vocabulary is large. Noise-contrastive estimation (NCE) is a sampling-based method that allows for fast learning with large vocabularies. Although NCE has shown promising performance in neural machine translation, it was considered to be an unsuccessful approach for language modelling. A sufficient investigation of the hyperparameters in the NCE-based neural language models was also missing. In this paper, we showed that NCE can be a successful approach in neural language modelling when the hyperparameters of a neural network are tuned appropriately. We introduced the 'search-then-converge' learning rate schedule for NCE and designed a heuristic that specifies how to use this schedule. The impact of the other important hyperparameters, such as the dropout rate and the weight initialisation range, was also demonstrated. We showed that appropriate tuning of NCE-based neural language models outperforms the state-of-the-art single-model methods on a popular benchmark." @default.
- W2758103492 created "2017-10-06" @default.
- W2758103492 creator A5052315074 @default.
- W2758103492 creator A5052957309 @default.
- W2758103492 date "2017-09-22" @default.
- W2758103492 modified "2023-09-27" @default.
- W2758103492 title "Improving Language Modelling with Noise-contrastive estimation" @default.
- W2758103492 cites W114517082 @default.
- W2758103492 cites W1558797106 @default.
- W2758103492 cites W2217098601 @default.
- W2758103492 cites W2951650375 @default.
- W2758103492 cites W2963084471 @default.
- W2758103492 hasPublicationYear "2017" @default.
- W2758103492 type Work @default.
- W2758103492 sameAs 2758103492 @default.
- W2758103492 citedByCount "0" @default.
- W2758103492 crossrefType "posted-content" @default.
- W2758103492 hasAuthorship W2758103492A5052315074 @default.
- W2758103492 hasAuthorship W2758103492A5052957309 @default.
- W2758103492 hasConcept C111919701 @default.
- W2758103492 hasConcept C115961682 @default.
- W2758103492 hasConcept C119857082 @default.
- W2758103492 hasConcept C13280743 @default.
- W2758103492 hasConcept C137293760 @default.
- W2758103492 hasConcept C138885662 @default.
- W2758103492 hasConcept C154945302 @default.
- W2758103492 hasConcept C173801870 @default.
- W2758103492 hasConcept C185798385 @default.
- W2758103492 hasConcept C203005215 @default.
- W2758103492 hasConcept C205649164 @default.
- W2758103492 hasConcept C2776145597 @default.
- W2758103492 hasConcept C2777601683 @default.
- W2758103492 hasConcept C28490314 @default.
- W2758103492 hasConcept C41008148 @default.
- W2758103492 hasConcept C41895202 @default.
- W2758103492 hasConcept C50644808 @default.
- W2758103492 hasConcept C68387754 @default.
- W2758103492 hasConcept C8642999 @default.
- W2758103492 hasConcept C99498987 @default.
- W2758103492 hasConceptScore W2758103492C111919701 @default.
- W2758103492 hasConceptScore W2758103492C115961682 @default.
- W2758103492 hasConceptScore W2758103492C119857082 @default.
- W2758103492 hasConceptScore W2758103492C13280743 @default.
- W2758103492 hasConceptScore W2758103492C137293760 @default.
- W2758103492 hasConceptScore W2758103492C138885662 @default.
- W2758103492 hasConceptScore W2758103492C154945302 @default.
- W2758103492 hasConceptScore W2758103492C173801870 @default.
- W2758103492 hasConceptScore W2758103492C185798385 @default.
- W2758103492 hasConceptScore W2758103492C203005215 @default.
- W2758103492 hasConceptScore W2758103492C205649164 @default.
- W2758103492 hasConceptScore W2758103492C2776145597 @default.
- W2758103492 hasConceptScore W2758103492C2777601683 @default.
- W2758103492 hasConceptScore W2758103492C28490314 @default.
- W2758103492 hasConceptScore W2758103492C41008148 @default.
- W2758103492 hasConceptScore W2758103492C41895202 @default.
- W2758103492 hasConceptScore W2758103492C50644808 @default.
- W2758103492 hasConceptScore W2758103492C68387754 @default.
- W2758103492 hasConceptScore W2758103492C8642999 @default.
- W2758103492 hasConceptScore W2758103492C99498987 @default.
- W2758103492 hasLocation W27581034921 @default.
- W2758103492 hasOpenAccess W2758103492 @default.
- W2758103492 hasPrimaryLocation W27581034921 @default.
- W2758103492 hasRelatedWork W2026383756 @default.
- W2758103492 hasRelatedWork W2105296092 @default.
- W2758103492 hasRelatedWork W2131241448 @default.
- W2758103492 hasRelatedWork W2140833774 @default.
- W2758103492 hasRelatedWork W2148862211 @default.
- W2758103492 hasRelatedWork W2604763608 @default.
- W2758103492 hasRelatedWork W2750881831 @default.
- W2758103492 hasRelatedWork W2782784481 @default.
- W2758103492 hasRelatedWork W2841543429 @default.
- W2758103492 hasRelatedWork W2938080845 @default.
- W2758103492 hasRelatedWork W2950000469 @default.
- W2758103492 hasRelatedWork W2951071805 @default.
- W2758103492 hasRelatedWork W3041009186 @default.
- W2758103492 hasRelatedWork W3120414668 @default.
- W2758103492 hasRelatedWork W3156622799 @default.
- W2758103492 hasRelatedWork W3158731518 @default.
- W2758103492 hasRelatedWork W3165418502 @default.
- W2758103492 hasRelatedWork W3170785217 @default.
- W2758103492 hasRelatedWork W3174220719 @default.
- W2758103492 hasRelatedWork W3180220823 @default.
- W2758103492 isParatext "false" @default.
- W2758103492 isRetracted "false" @default.
- W2758103492 magId "2758103492" @default.
- W2758103492 workType "article" @default.