Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287100880> ?p ?o ?g. }
Showing items 1 to 72 of
72
with 100 items per page.
- W4287100880 abstract "We propose a cross-modal transformer-based neural correction models that refines the output of an automatic speech recognition (ASR) system so as to exclude ASR errors. Generally, neural correction models are composed of encoder-decoder networks, which can directly model sequence-to-sequence mapping problems. The most successful method is to use both input speech and its ASR output text as the input contexts for the encoder-decoder networks. However, the conventional method cannot take into account the relationships between these two different modal inputs because the input contexts are separately encoded for each modal. To effectively leverage the correlated information between the two different modal inputs, our proposed models encode two different contexts jointly on the basis of cross-modal self-attention using a transformer. We expect that cross-modal self-attention can effectively capture the relationships between two different modals for refining ASR hypotheses. We also introduce a shallow fusion technique to efficiently integrate the first-pass ASR model and our proposed neural correction model. Experiments on Japanese natural language ASR tasks demonstrated that our proposed models achieve better ASR performance than conventional neural correction models." @default.
- W4287100880 created "2022-07-25" @default.
- W4287100880 creator A5009756622 @default.
- W4287100880 creator A5013009514 @default.
- W4287100880 creator A5027118468 @default.
- W4287100880 creator A5033975068 @default.
- W4287100880 creator A5060644399 @default.
- W4287100880 creator A5080550833 @default.
- W4287100880 creator A5082466886 @default.
- W4287100880 creator A5087290011 @default.
- W4287100880 date "2021-07-04" @default.
- W4287100880 modified "2023-09-28" @default.
- W4287100880 title "Cross-Modal Transformer-Based Neural Correction Models for Automatic Speech Recognition" @default.
- W4287100880 hasPublicationYear "2021" @default.
- W4287100880 type Work @default.
- W4287100880 citedByCount "0" @default.
- W4287100880 crossrefType "posted-content" @default.
- W4287100880 hasAuthorship W4287100880A5009756622 @default.
- W4287100880 hasAuthorship W4287100880A5013009514 @default.
- W4287100880 hasAuthorship W4287100880A5027118468 @default.
- W4287100880 hasAuthorship W4287100880A5033975068 @default.
- W4287100880 hasAuthorship W4287100880A5060644399 @default.
- W4287100880 hasAuthorship W4287100880A5080550833 @default.
- W4287100880 hasAuthorship W4287100880A5082466886 @default.
- W4287100880 hasAuthorship W4287100880A5087290011 @default.
- W4287100880 hasBestOaLocation W42871008801 @default.
- W4287100880 hasConcept C111919701 @default.
- W4287100880 hasConcept C118505674 @default.
- W4287100880 hasConcept C119599485 @default.
- W4287100880 hasConcept C127413603 @default.
- W4287100880 hasConcept C153083717 @default.
- W4287100880 hasConcept C153180895 @default.
- W4287100880 hasConcept C154945302 @default.
- W4287100880 hasConcept C165801399 @default.
- W4287100880 hasConcept C185592680 @default.
- W4287100880 hasConcept C188027245 @default.
- W4287100880 hasConcept C28490314 @default.
- W4287100880 hasConcept C41008148 @default.
- W4287100880 hasConcept C50644808 @default.
- W4287100880 hasConcept C66322947 @default.
- W4287100880 hasConcept C71139939 @default.
- W4287100880 hasConceptScore W4287100880C111919701 @default.
- W4287100880 hasConceptScore W4287100880C118505674 @default.
- W4287100880 hasConceptScore W4287100880C119599485 @default.
- W4287100880 hasConceptScore W4287100880C127413603 @default.
- W4287100880 hasConceptScore W4287100880C153083717 @default.
- W4287100880 hasConceptScore W4287100880C153180895 @default.
- W4287100880 hasConceptScore W4287100880C154945302 @default.
- W4287100880 hasConceptScore W4287100880C165801399 @default.
- W4287100880 hasConceptScore W4287100880C185592680 @default.
- W4287100880 hasConceptScore W4287100880C188027245 @default.
- W4287100880 hasConceptScore W4287100880C28490314 @default.
- W4287100880 hasConceptScore W4287100880C41008148 @default.
- W4287100880 hasConceptScore W4287100880C50644808 @default.
- W4287100880 hasConceptScore W4287100880C66322947 @default.
- W4287100880 hasConceptScore W4287100880C71139939 @default.
- W4287100880 hasLocation W42871008801 @default.
- W4287100880 hasOpenAccess W4287100880 @default.
- W4287100880 hasPrimaryLocation W42871008801 @default.
- W4287100880 hasRelatedWork W11553578 @default.
- W4287100880 hasRelatedWork W13920906 @default.
- W4287100880 hasRelatedWork W2883085 @default.
- W4287100880 hasRelatedWork W2899263 @default.
- W4287100880 hasRelatedWork W2966661 @default.
- W4287100880 hasRelatedWork W3011399 @default.
- W4287100880 hasRelatedWork W5939791 @default.
- W4287100880 hasRelatedWork W654939 @default.
- W4287100880 hasRelatedWork W7958345 @default.
- W4287100880 hasRelatedWork W9531238 @default.
- W4287100880 isParatext "false" @default.
- W4287100880 isRetracted "false" @default.
- W4287100880 workType "article" @default.