Matches in SemOpenAlex for { <https://semopenalex.org/work/W2792156659> ?p ?o ?g. }
Showing items 1 to 73 of
73
with 100 items per page.
- W2792156659 abstract "End-to-end (E2E) automatic speech recognition (ASR) systems directly map acoustics to words using a unified model. Previous works mostly focus on E2E training a single model which integrates acoustic and language model into a whole. Although E2E training benefits from sequence modeling and simplified decoding pipelines, large amount of transcribed acoustic data is usually required, and traditional acoustic and language modelling techniques cannot be utilized. In this paper, a novel modular training framework of E2E ASR is proposed to separately train neural acoustic and language models during training stage, while still performing end-to-end inference in decoding stage. Here, an acoustics-to-phoneme model (A2P) and a phoneme-to-word model (P2W) are trained using acoustic data and text data respectively. A phone synchronous decoding (PSD) module is inserted between A2P and P2W to reduce sequence lengths without precision loss. Finally, modules are integrated into an acousticsto-word model (A2W) and jointly optimized using acoustic data to retain the advantage of sequence modeling. Experiments on a 300- hour Switchboard task show significant improvement over the direct A2W model. The efficiency in both training and decoding also benefits from the proposed method." @default.
- W2792156659 created "2018-03-29" @default.
- W2792156659 creator A5002433660 @default.
- W2792156659 creator A5019560977 @default.
- W2792156659 creator A5043098653 @default.
- W2792156659 creator A5085028455 @default.
- W2792156659 date "2018-03-02" @default.
- W2792156659 modified "2023-09-26" @default.
- W2792156659 title "On Modular Training of Neural Acoustics-to-Word Model for LVCSR" @default.
- W2792156659 doi "https://doi.org/10.48550/arxiv.1803.01090" @default.
- W2792156659 hasPublicationYear "2018" @default.
- W2792156659 type Work @default.
- W2792156659 sameAs 2792156659 @default.
- W2792156659 citedByCount "1" @default.
- W2792156659 countsByYear W27921566592018 @default.
- W2792156659 crossrefType "posted-content" @default.
- W2792156659 hasAuthorship W2792156659A5002433660 @default.
- W2792156659 hasAuthorship W2792156659A5019560977 @default.
- W2792156659 hasAuthorship W2792156659A5043098653 @default.
- W2792156659 hasAuthorship W2792156659A5085028455 @default.
- W2792156659 hasBestOaLocation W27921566591 @default.
- W2792156659 hasConcept C11413529 @default.
- W2792156659 hasConcept C127413603 @default.
- W2792156659 hasConcept C137293760 @default.
- W2792156659 hasConcept C138885662 @default.
- W2792156659 hasConcept C154945302 @default.
- W2792156659 hasConcept C155635449 @default.
- W2792156659 hasConcept C201995342 @default.
- W2792156659 hasConcept C2778112365 @default.
- W2792156659 hasConcept C2780451532 @default.
- W2792156659 hasConcept C28490314 @default.
- W2792156659 hasConcept C41008148 @default.
- W2792156659 hasConcept C41895202 @default.
- W2792156659 hasConcept C54355233 @default.
- W2792156659 hasConcept C57273362 @default.
- W2792156659 hasConcept C61328038 @default.
- W2792156659 hasConcept C86803240 @default.
- W2792156659 hasConcept C90805587 @default.
- W2792156659 hasConceptScore W2792156659C11413529 @default.
- W2792156659 hasConceptScore W2792156659C127413603 @default.
- W2792156659 hasConceptScore W2792156659C137293760 @default.
- W2792156659 hasConceptScore W2792156659C138885662 @default.
- W2792156659 hasConceptScore W2792156659C154945302 @default.
- W2792156659 hasConceptScore W2792156659C155635449 @default.
- W2792156659 hasConceptScore W2792156659C201995342 @default.
- W2792156659 hasConceptScore W2792156659C2778112365 @default.
- W2792156659 hasConceptScore W2792156659C2780451532 @default.
- W2792156659 hasConceptScore W2792156659C28490314 @default.
- W2792156659 hasConceptScore W2792156659C41008148 @default.
- W2792156659 hasConceptScore W2792156659C41895202 @default.
- W2792156659 hasConceptScore W2792156659C54355233 @default.
- W2792156659 hasConceptScore W2792156659C57273362 @default.
- W2792156659 hasConceptScore W2792156659C61328038 @default.
- W2792156659 hasConceptScore W2792156659C86803240 @default.
- W2792156659 hasConceptScore W2792156659C90805587 @default.
- W2792156659 hasLocation W27921566591 @default.
- W2792156659 hasLocation W27921566592 @default.
- W2792156659 hasOpenAccess W2792156659 @default.
- W2792156659 hasPrimaryLocation W27921566591 @default.
- W2792156659 hasRelatedWork W1486956970 @default.
- W2792156659 hasRelatedWork W1822699154 @default.
- W2792156659 hasRelatedWork W1826521293 @default.
- W2792156659 hasRelatedWork W248303808 @default.
- W2792156659 hasRelatedWork W2953291251 @default.
- W2792156659 hasRelatedWork W3080136773 @default.
- W2792156659 hasRelatedWork W3146812467 @default.
- W2792156659 hasRelatedWork W4249316903 @default.
- W2792156659 hasRelatedWork W4290682478 @default.
- W2792156659 hasRelatedWork W62743518 @default.
- W2792156659 isParatext "false" @default.
- W2792156659 isRetracted "false" @default.
- W2792156659 magId "2792156659" @default.
- W2792156659 workType "article" @default.