FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2772048672> ?p ?o ?g. }

• W2772048672 abstract "Having a sequence-to-sequence model which can operate in an online fashion is important for streaming applications such as Voice Search. Neural transducer is a streaming sequence-to-sequence model, but has shown a significant degradation in performance compared to non-streaming models such as Listen, Attend and Spell (LAS). In this paper, we present various improvements to NT. Specifically, we look at increasing the window over which NT computes attention, mainly by looking backwards in time so the model still remains online. In addition, we explore initializing a NT model from a LAS-trained model so that it is guided with a better alignment. Finally, we explore including stronger language models such as using wordpiece models, and applying an external LM during the beam search. On a Voice Search task, we find with these improvements we can get NT to match the performance of LAS." @default.
• W2772048672 created "2017-12-22" @default.
• W2772048672 creator A5010253402 @default.
• W2772048672 creator A5023098355 @default.
• W2772048672 creator A5027763497 @default.
• W2772048672 creator A5032640894 @default.
• W2772048672 creator A5056830118 @default.
• W2772048672 creator A5070513394 @default.
• W2772048672 creator A5083465496 @default.
• W2772048672 date "2017-12-05" @default.
• W2772048672 modified "2023-10-18" @default.
• W2772048672 title "Improving the Performance of Online Neural Transducer Models" @default.
• W2772048672 cites W1828163288 @default.
• W2772048672 cites W1971453323 @default.
• W2772048672 cites W2064675550 @default.
• W2772048672 cites W2121879602 @default.
• W2772048672 cites W2515439472 @default.
• W2772048672 cites W2525778437 @default.
• W2772048672 cites W2617258110 @default.
• W2772048672 cites W2625979394 @default.
• W2772048672 cites W2626778328 @default.
• W2772048672 cites W2739427748 @default.
• W2772048672 cites W2748679025 @default.
• W2772048672 cites W2750499125 @default.
• W2772048672 cites W2953022181 @default.
• W2772048672 cites W2963414781 @default.
• W2772048672 cites W2963729263 @default.
• W2772048672 cites W2963920996 @default.
• W2772048672 cites W2964121744 @default.
• W2772048672 cites W2964308564 @default.
• W2772048672 doi "https://doi.org/10.48550/arxiv.1712.01807" @default.
• W2772048672 hasPublicationYear "2017" @default.
• W2772048672 type Work @default.
• W2772048672 sameAs 2772048672 @default.
• W2772048672 citedByCount "3" @default.
• W2772048672 countsByYear W27720486722018 @default.
• W2772048672 countsByYear W27720486722019 @default.
• W2772048672 countsByYear W27720486722020 @default.
• W2772048672 crossrefType "posted-content" @default.
• W2772048672 hasAuthorship W2772048672A5010253402 @default.
• W2772048672 hasAuthorship W2772048672A5023098355 @default.
• W2772048672 hasAuthorship W2772048672A5027763497 @default.
• W2772048672 hasAuthorship W2772048672A5032640894 @default.
• W2772048672 hasAuthorship W2772048672A5056830118 @default.
• W2772048672 hasAuthorship W2772048672A5070513394 @default.
• W2772048672 hasAuthorship W2772048672A5083465496 @default.
• W2772048672 hasBestOaLocation W27720486721 @default.
• W2772048672 hasConcept C11413529 @default.
• W2772048672 hasConcept C114466953 @default.
• W2772048672 hasConcept C119857082 @default.
• W2772048672 hasConcept C125583679 @default.
• W2772048672 hasConcept C127413603 @default.
• W2772048672 hasConcept C136764020 @default.
• W2772048672 hasConcept C137293760 @default.
• W2772048672 hasConcept C144024400 @default.
• W2772048672 hasConcept C154945302 @default.
• W2772048672 hasConcept C19165224 @default.
• W2772048672 hasConcept C19889080 @default.
• W2772048672 hasConcept C199360897 @default.
• W2772048672 hasConcept C201995342 @default.
• W2772048672 hasConcept C2778112365 @default.
• W2772048672 hasConcept C2778751112 @default.
• W2772048672 hasConcept C2780451532 @default.
• W2772048672 hasConcept C2780957641 @default.
• W2772048672 hasConcept C28490314 @default.
• W2772048672 hasConcept C41008148 @default.
• W2772048672 hasConcept C54355233 @default.
• W2772048672 hasConcept C86803240 @default.
• W2772048672 hasConceptScore W2772048672C11413529 @default.
• W2772048672 hasConceptScore W2772048672C114466953 @default.
• W2772048672 hasConceptScore W2772048672C119857082 @default.
• W2772048672 hasConceptScore W2772048672C125583679 @default.
• W2772048672 hasConceptScore W2772048672C127413603 @default.
• W2772048672 hasConceptScore W2772048672C136764020 @default.
• W2772048672 hasConceptScore W2772048672C137293760 @default.
• W2772048672 hasConceptScore W2772048672C144024400 @default.
• W2772048672 hasConceptScore W2772048672C154945302 @default.
• W2772048672 hasConceptScore W2772048672C19165224 @default.
• W2772048672 hasConceptScore W2772048672C19889080 @default.
• W2772048672 hasConceptScore W2772048672C199360897 @default.
• W2772048672 hasConceptScore W2772048672C201995342 @default.
• W2772048672 hasConceptScore W2772048672C2778112365 @default.
• W2772048672 hasConceptScore W2772048672C2778751112 @default.
• W2772048672 hasConceptScore W2772048672C2780451532 @default.
• W2772048672 hasConceptScore W2772048672C2780957641 @default.
• W2772048672 hasConceptScore W2772048672C28490314 @default.
• W2772048672 hasConceptScore W2772048672C41008148 @default.
• W2772048672 hasConceptScore W2772048672C54355233 @default.
• W2772048672 hasConceptScore W2772048672C86803240 @default.
• W2772048672 hasLocation W27720486721 @default.
• W2772048672 hasOpenAccess W2772048672 @default.
• W2772048672 hasPrimaryLocation W27720486721 @default.
• W2772048672 hasRelatedWork W1855892484 @default.
• W2772048672 hasRelatedWork W2034490042 @default.
• W2772048672 hasRelatedWork W2374918184 @default.
• W2772048672 hasRelatedWork W2414484917 @default.
• W2772048672 hasRelatedWork W2772048672 @default.
• W2772048672 hasRelatedWork W2775766866 @default.
• W2772048672 hasRelatedWork W2936774411 @default.
• W2772048672 hasRelatedWork W2962742956 @default.

• next