Matches in SemOpenAlex for { <https://semopenalex.org/work/W2994771587> ?p ?o ?g. }
- W2994771587 abstract "Recently, SpecAugment, an augmentation scheme for automatic speech recognition that acts directly on the spectrogram of input utterances, has shown to be highly effective in enhancing the performance of end-to-end networks on public datasets. In this paper, we demonstrate its effectiveness on tasks with large scale datasets by investigating its application to the Google Multidomain Dataset (Narayanan et al., 2018). We achieve improvement across all test domains by mixing raw training data augmented with SpecAugment and noise-perturbed training data when training the acoustic model. We also introduce a modification of SpecAugment that adapts the time mask size and/or multiplicity depending on the length of the utterance, which can potentially benefit large scale tasks. By using adaptive masking, we are able to further improve the performance of the Listen, Attend and Spell model on LibriSpeech to 2.2% WER on test-clean and 5.2% WER on test-other." @default.
- W2994771587 created "2019-12-26" @default.
- W2994771587 creator A5007069562 @default.
- W2994771587 creator A5010253402 @default.
- W2994771587 creator A5016918393 @default.
- W2994771587 creator A5018757219 @default.
- W2994771587 creator A5027763497 @default.
- W2994771587 creator A5034995105 @default.
- W2994771587 creator A5071773009 @default.
- W2994771587 creator A5088551093 @default.
- W2994771587 date "2019-12-11" @default.
- W2994771587 modified "2023-10-16" @default.
- W2994771587 title "SpecAugment on Large Scale Datasets" @default.
- W2994771587 cites W1485222997 @default.
- W2994771587 cites W1494198834 @default.
- W2994771587 cites W1915251500 @default.
- W2994771587 cites W1922655562 @default.
- W2994771587 cites W1989674786 @default.
- W2994771587 cites W2099621636 @default.
- W2994771587 cites W2121879602 @default.
- W2994771587 cites W2184343439 @default.
- W2994771587 cites W2327501763 @default.
- W2994771587 cites W2397147568 @default.
- W2994771587 cites W2407080277 @default.
- W2994771587 cites W2577366047 @default.
- W2994771587 cites W2617258110 @default.
- W2994771587 cites W2622203030 @default.
- W2994771587 cites W2759921250 @default.
- W2994771587 cites W2887528618 @default.
- W2994771587 cites W2888493875 @default.
- W2994771587 cites W2936774411 @default.
- W2994771587 cites W2962760690 @default.
- W2994771587 cites W2963122170 @default.
- W2994771587 cites W2963590452 @default.
- W2994771587 cites W2972816482 @default.
- W2994771587 cites W2972818416 @default.
- W2994771587 cites W2972991469 @default.
- W2994771587 cites W2977728428 @default.
- W2994771587 cites W3103005696 @default.
- W2994771587 doi "https://doi.org/10.48550/arxiv.1912.05533" @default.
- W2994771587 hasPublicationYear "2019" @default.
- W2994771587 type Work @default.
- W2994771587 sameAs 2994771587 @default.
- W2994771587 citedByCount "3" @default.
- W2994771587 countsByYear W29947715872019 @default.
- W2994771587 countsByYear W29947715872020 @default.
- W2994771587 countsByYear W29947715872021 @default.
- W2994771587 crossrefType "posted-content" @default.
- W2994771587 hasAuthorship W2994771587A5007069562 @default.
- W2994771587 hasAuthorship W2994771587A5010253402 @default.
- W2994771587 hasAuthorship W2994771587A5016918393 @default.
- W2994771587 hasAuthorship W2994771587A5018757219 @default.
- W2994771587 hasAuthorship W2994771587A5027763497 @default.
- W2994771587 hasAuthorship W2994771587A5034995105 @default.
- W2994771587 hasAuthorship W2994771587A5071773009 @default.
- W2994771587 hasAuthorship W2994771587A5088551093 @default.
- W2994771587 hasBestOaLocation W29947715871 @default.
- W2994771587 hasConcept C119857082 @default.
- W2994771587 hasConcept C121332964 @default.
- W2994771587 hasConcept C134306372 @default.
- W2994771587 hasConcept C137293760 @default.
- W2994771587 hasConcept C142362112 @default.
- W2994771587 hasConcept C144024400 @default.
- W2994771587 hasConcept C153349607 @default.
- W2994771587 hasConcept C154945302 @default.
- W2994771587 hasConcept C16910744 @default.
- W2994771587 hasConcept C19165224 @default.
- W2994771587 hasConcept C199360897 @default.
- W2994771587 hasConcept C2775852435 @default.
- W2994771587 hasConcept C2777402240 @default.
- W2994771587 hasConcept C2778755073 @default.
- W2994771587 hasConcept C2780957641 @default.
- W2994771587 hasConcept C28490314 @default.
- W2994771587 hasConcept C33923547 @default.
- W2994771587 hasConcept C41008148 @default.
- W2994771587 hasConcept C45273575 @default.
- W2994771587 hasConcept C62520636 @default.
- W2994771587 hasConcept C77618280 @default.
- W2994771587 hasConceptScore W2994771587C119857082 @default.
- W2994771587 hasConceptScore W2994771587C121332964 @default.
- W2994771587 hasConceptScore W2994771587C134306372 @default.
- W2994771587 hasConceptScore W2994771587C137293760 @default.
- W2994771587 hasConceptScore W2994771587C142362112 @default.
- W2994771587 hasConceptScore W2994771587C144024400 @default.
- W2994771587 hasConceptScore W2994771587C153349607 @default.
- W2994771587 hasConceptScore W2994771587C154945302 @default.
- W2994771587 hasConceptScore W2994771587C16910744 @default.
- W2994771587 hasConceptScore W2994771587C19165224 @default.
- W2994771587 hasConceptScore W2994771587C199360897 @default.
- W2994771587 hasConceptScore W2994771587C2775852435 @default.
- W2994771587 hasConceptScore W2994771587C2777402240 @default.
- W2994771587 hasConceptScore W2994771587C2778755073 @default.
- W2994771587 hasConceptScore W2994771587C2780957641 @default.
- W2994771587 hasConceptScore W2994771587C28490314 @default.
- W2994771587 hasConceptScore W2994771587C33923547 @default.
- W2994771587 hasConceptScore W2994771587C41008148 @default.
- W2994771587 hasConceptScore W2994771587C45273575 @default.
- W2994771587 hasConceptScore W2994771587C62520636 @default.
- W2994771587 hasConceptScore W2994771587C77618280 @default.
- W2994771587 hasLocation W29947715871 @default.