Matches in SemOpenAlex for { <https://semopenalex.org/work/W2963390466> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W2963390466 abstract "The performance of automatic speech recognition systems under noisy environments still leaves room for improvement. Speech enhancement or feature enhancement techniques for increasing noise robustness of these systems usually add components to the recognition system that need careful optimization. In this work, we propose the use of a relatively simple curriculum training strategy called accordion annealing (ACCAN). It uses a multi-stage training schedule where samples at signal-to-noise ratio (SNR) values as low as 0dB are first added and samples at increasing higher SNR values are gradually added up to an SNR value of 50dB. We also use a method called per-epoch noise mixing (PEM) that generates noisy training samples online during training and thus enables dynamically changing the SNR of our training data. Both the ACCAN and the PEM methods are evaluated on a end-to-end speech recognition pipeline on the Wall Street Journal corpus. ACCAN decreases the average word error rate (WER) on the 20dB to −10dB SNR range by up to 31.4% when compared to a conventional multi-condition training method." @default.
- W2963390466 created "2019-07-30" @default.
- W2963390466 creator A5030469075 @default.
- W2963390466 creator A5053813559 @default.
- W2963390466 creator A5074581209 @default.
- W2963390466 date "2017-08-01" @default.
- W2963390466 modified "2023-10-03" @default.
- W2963390466 title "A curriculum learning method for improved noise robustness in automatic speech recognition" @default.
- W2963390466 cites W1548957680 @default.
- W2963390466 cites W1973681148 @default.
- W2963390466 cites W1974387177 @default.
- W2963390466 cites W1992475611 @default.
- W2963390466 cites W2062164080 @default.
- W2963390466 cites W2064675550 @default.
- W2963390466 cites W2096653978 @default.
- W2963390466 cites W2127141656 @default.
- W2963390466 cites W2128653836 @default.
- W2963390466 cites W2160815625 @default.
- W2963390466 cites W2288669569 @default.
- W2963390466 cites W2289731793 @default.
- W2963390466 cites W2290318471 @default.
- W2963390466 cites W2296073425 @default.
- W2963390466 cites W2399733683 @default.
- W2963390466 cites W2963211739 @default.
- W2963390466 doi "https://doi.org/10.23919/eusipco.2017.8081267" @default.
- W2963390466 hasPublicationYear "2017" @default.
- W2963390466 type Work @default.
- W2963390466 sameAs 2963390466 @default.
- W2963390466 citedByCount "35" @default.
- W2963390466 countsByYear W29633904662016 @default.
- W2963390466 countsByYear W29633904662018 @default.
- W2963390466 countsByYear W29633904662019 @default.
- W2963390466 countsByYear W29633904662020 @default.
- W2963390466 countsByYear W29633904662021 @default.
- W2963390466 countsByYear W29633904662022 @default.
- W2963390466 countsByYear W29633904662023 @default.
- W2963390466 crossrefType "proceedings-article" @default.
- W2963390466 hasAuthorship W2963390466A5030469075 @default.
- W2963390466 hasAuthorship W2963390466A5053813559 @default.
- W2963390466 hasAuthorship W2963390466A5074581209 @default.
- W2963390466 hasBestOaLocation W29633904662 @default.
- W2963390466 hasConcept C104317684 @default.
- W2963390466 hasConcept C153180895 @default.
- W2963390466 hasConcept C154945302 @default.
- W2963390466 hasConcept C163294075 @default.
- W2963390466 hasConcept C185592680 @default.
- W2963390466 hasConcept C2776182073 @default.
- W2963390466 hasConcept C28490314 @default.
- W2963390466 hasConcept C29265498 @default.
- W2963390466 hasConcept C40969351 @default.
- W2963390466 hasConcept C41008148 @default.
- W2963390466 hasConcept C52622490 @default.
- W2963390466 hasConcept C55493867 @default.
- W2963390466 hasConcept C63479239 @default.
- W2963390466 hasConceptScore W2963390466C104317684 @default.
- W2963390466 hasConceptScore W2963390466C153180895 @default.
- W2963390466 hasConceptScore W2963390466C154945302 @default.
- W2963390466 hasConceptScore W2963390466C163294075 @default.
- W2963390466 hasConceptScore W2963390466C185592680 @default.
- W2963390466 hasConceptScore W2963390466C2776182073 @default.
- W2963390466 hasConceptScore W2963390466C28490314 @default.
- W2963390466 hasConceptScore W2963390466C29265498 @default.
- W2963390466 hasConceptScore W2963390466C40969351 @default.
- W2963390466 hasConceptScore W2963390466C41008148 @default.
- W2963390466 hasConceptScore W2963390466C52622490 @default.
- W2963390466 hasConceptScore W2963390466C55493867 @default.
- W2963390466 hasConceptScore W2963390466C63479239 @default.
- W2963390466 hasLocation W29633904661 @default.
- W2963390466 hasLocation W29633904662 @default.
- W2963390466 hasLocation W29633904663 @default.
- W2963390466 hasOpenAccess W2963390466 @default.
- W2963390466 hasPrimaryLocation W29633904661 @default.
- W2963390466 hasRelatedWork W101328856 @default.
- W2963390466 hasRelatedWork W1506140669 @default.
- W2963390466 hasRelatedWork W2024338893 @default.
- W2963390466 hasRelatedWork W2092305100 @default.
- W2963390466 hasRelatedWork W2151802380 @default.
- W2963390466 hasRelatedWork W2153435877 @default.
- W2963390466 hasRelatedWork W2775709066 @default.
- W2963390466 hasRelatedWork W2776559070 @default.
- W2963390466 hasRelatedWork W2908234883 @default.
- W2963390466 hasRelatedWork W3209376089 @default.
- W2963390466 isParatext "false" @default.
- W2963390466 isRetracted "false" @default.
- W2963390466 magId "2963390466" @default.
- W2963390466 workType "article" @default.