Matches in SemOpenAlex for { <https://semopenalex.org/work/W4372260554> ?p ?o ?g. }
Showing items 1 to 84 of
84
with 100 items per page.
- W4372260554 abstract "We previously proposed contextual spelling correction (CSC) to correct the output of end-to-end (E2E) automatic speech recognition (ASR) models with contextual information such as name, place, etc. Although CSC has achieved reasonable improvement in the biasing problem, there are still two drawbacks for further accuracy improvement. First, due to information limitation in text only hypothesis or weak performance of ASR model on rare domains, the CSC model may fail to correct phrases with similar pronunciation or anti-context cases where all biasing phrases are not present in the utterance. Second, there is a discrepancy between the training and inference of CSC. The bias list in training is randomly selected but in inference there may be more similarity between ground truth phrase and other phrases. To solve above limitations, in this paper we propose an improved non-autoregressive (NAR) spelling correction model for contextual biasing in E2E neural transducer-based ASR systems to improve the previous CSC model from two perspectives: Firstly, we incorporate acoustics information with an external attention as well as text hypotheses into CSC to better distinguish target phrase from dissimilar or irrelevant phrases. Secondly, we design a semantic aware data augmentation schema in training phrase to reduce the mismatch between training and inference to further boost the biasing accuracy. Experiments show that the improved method outperforms the baseline ASR+Biasing system by as much as 20.3% relative name recall gain and achieves stable improvement compared to the previous CSC method over different bias list name coverage ratio." @default.
- W4372260554 created "2023-05-07" @default.
- W4372260554 creator A5007745122 @default.
- W4372260554 creator A5013702415 @default.
- W4372260554 creator A5027423694 @default.
- W4372260554 creator A5029670581 @default.
- W4372260554 date "2023-06-04" @default.
- W4372260554 modified "2023-09-27" @default.
- W4372260554 title "Improving Contextual Spelling Correction by External Acoustics Attention and Semantic Aware Data Augmentation" @default.
- W4372260554 cites W2402040300 @default.
- W4372260554 cites W2886319145 @default.
- W4372260554 cites W2889012072 @default.
- W4372260554 cites W2916997151 @default.
- W4372260554 cites W2962760690 @default.
- W4372260554 cites W2972625221 @default.
- W4372260554 cites W3011339933 @default.
- W4372260554 cites W3094667432 @default.
- W4372260554 cites W3096815019 @default.
- W4372260554 cites W3097794466 @default.
- W4372260554 cites W3140235797 @default.
- W4372260554 cites W3141464856 @default.
- W4372260554 cites W3161873870 @default.
- W4372260554 cites W3193959417 @default.
- W4372260554 cites W3198004110 @default.
- W4372260554 cites W3205495812 @default.
- W4372260554 cites W3211278025 @default.
- W4372260554 cites W4221165942 @default.
- W4372260554 cites W4224917454 @default.
- W4372260554 cites W4226292626 @default.
- W4372260554 cites W4226302523 @default.
- W4372260554 cites W4226462878 @default.
- W4372260554 doi "https://doi.org/10.1109/icassp49357.2023.10095434" @default.
- W4372260554 hasPublicationYear "2023" @default.
- W4372260554 type Work @default.
- W4372260554 citedByCount "0" @default.
- W4372260554 crossrefType "proceedings-article" @default.
- W4372260554 hasAuthorship W4372260554A5007745122 @default.
- W4372260554 hasAuthorship W4372260554A5013702415 @default.
- W4372260554 hasAuthorship W4372260554A5027423694 @default.
- W4372260554 hasAuthorship W4372260554A5029670581 @default.
- W4372260554 hasBestOaLocation W43722605541 @default.
- W4372260554 hasConcept C138885662 @default.
- W4372260554 hasConcept C151730666 @default.
- W4372260554 hasConcept C154945302 @default.
- W4372260554 hasConcept C204321447 @default.
- W4372260554 hasConcept C2775852435 @default.
- W4372260554 hasConcept C2776214188 @default.
- W4372260554 hasConcept C2776224158 @default.
- W4372260554 hasConcept C2777801307 @default.
- W4372260554 hasConcept C2779343474 @default.
- W4372260554 hasConcept C28490314 @default.
- W4372260554 hasConcept C41008148 @default.
- W4372260554 hasConcept C41895202 @default.
- W4372260554 hasConcept C86803240 @default.
- W4372260554 hasConceptScore W4372260554C138885662 @default.
- W4372260554 hasConceptScore W4372260554C151730666 @default.
- W4372260554 hasConceptScore W4372260554C154945302 @default.
- W4372260554 hasConceptScore W4372260554C204321447 @default.
- W4372260554 hasConceptScore W4372260554C2775852435 @default.
- W4372260554 hasConceptScore W4372260554C2776214188 @default.
- W4372260554 hasConceptScore W4372260554C2776224158 @default.
- W4372260554 hasConceptScore W4372260554C2777801307 @default.
- W4372260554 hasConceptScore W4372260554C2779343474 @default.
- W4372260554 hasConceptScore W4372260554C28490314 @default.
- W4372260554 hasConceptScore W4372260554C41008148 @default.
- W4372260554 hasConceptScore W4372260554C41895202 @default.
- W4372260554 hasConceptScore W4372260554C86803240 @default.
- W4372260554 hasLocation W43722605541 @default.
- W4372260554 hasLocation W43722605542 @default.
- W4372260554 hasOpenAccess W4372260554 @default.
- W4372260554 hasPrimaryLocation W43722605541 @default.
- W4372260554 hasRelatedWork W1596769518 @default.
- W4372260554 hasRelatedWork W1886600421 @default.
- W4372260554 hasRelatedWork W2065704406 @default.
- W4372260554 hasRelatedWork W2075207372 @default.
- W4372260554 hasRelatedWork W2083039973 @default.
- W4372260554 hasRelatedWork W2161862087 @default.
- W4372260554 hasRelatedWork W2292933197 @default.
- W4372260554 hasRelatedWork W2369308426 @default.
- W4372260554 hasRelatedWork W2886693075 @default.
- W4372260554 hasRelatedWork W3198503472 @default.
- W4372260554 isParatext "false" @default.
- W4372260554 isRetracted "false" @default.
- W4372260554 workType "article" @default.