Matches in SemOpenAlex for { <https://semopenalex.org/work/W2893370043> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W2893370043 abstract "Spectral mapping uses a deep neural network (DNN) to map directly from noisy speech to clean speech. Our previous study [1] found that the performance of spectral mapping improves greatly when using helpful cues from an acoustic model trained on clean speech. The mapper network learns to mimic the input favored by the spectral classifier and cleans the features accordingly. In this study, we explore two new innovations: we replace a DNN-based spectral mapper with a residual network that is more attuned to the goal of predicting clean speech. We also examine how integrating long term context in the mimic criterion (via wide-residual biL-STM networks) affects the performance of spectral mapping compared to DNNs. Our goal is to derive a model that can be used as a preprocessor for any recognition system; the features derived from our model are passed through the standard Kaldi ASR pipeline and achieve a WER of 9.3%, which is the lowest recorded word error rate for CHiME-2 dataset using only feature adaptation." @default.
- W2893370043 created "2018-10-05" @default.
- W2893370043 creator A5055583643 @default.
- W2893370043 creator A5056667180 @default.
- W2893370043 creator A5089667871 @default.
- W2893370043 date "2018-12-01" @default.
- W2893370043 modified "2023-09-26" @default.
- W2893370043 title "An Exploration of Mimic Architectures for Residual Network Based Spectral Mapping" @default.
- W2893370043 cites W1524333225 @default.
- W2893370043 cites W1600744878 @default.
- W2893370043 cites W1821462560 @default.
- W2893370043 cites W1897240248 @default.
- W2893370043 cites W2005708641 @default.
- W2893370043 cites W2006129368 @default.
- W2893370043 cites W2035576074 @default.
- W2893370043 cites W2057826000 @default.
- W2893370043 cites W2101045344 @default.
- W2893370043 cites W2134797427 @default.
- W2893370043 cites W2143612262 @default.
- W2893370043 cites W2144786298 @default.
- W2893370043 cites W2176412452 @default.
- W2893370043 cites W2194775991 @default.
- W2893370043 cites W2285420822 @default.
- W2893370043 cites W2288645994 @default.
- W2893370043 cites W2291652038 @default.
- W2893370043 cites W2296167893 @default.
- W2893370043 cites W2399557756 @default.
- W2893370043 cites W2402040300 @default.
- W2893370043 cites W2516342150 @default.
- W2893370043 cites W2519224033 @default.
- W2893370043 cites W2530876040 @default.
- W2893370043 cites W2587717635 @default.
- W2893370043 cites W2682975603 @default.
- W2893370043 cites W2962946266 @default.
- W2893370043 cites W2962998773 @default.
- W2893370043 cites W2963410064 @default.
- W2893370043 cites W2963453742 @default.
- W2893370043 cites W2964121744 @default.
- W2893370043 cites W2964137095 @default.
- W2893370043 cites W2966212213 @default.
- W2893370043 doi "https://doi.org/10.1109/slt.2018.8639632" @default.
- W2893370043 hasPublicationYear "2018" @default.
- W2893370043 type Work @default.
- W2893370043 sameAs 2893370043 @default.
- W2893370043 citedByCount "6" @default.
- W2893370043 countsByYear W28933700432019 @default.
- W2893370043 countsByYear W28933700432020 @default.
- W2893370043 countsByYear W28933700432021 @default.
- W2893370043 crossrefType "proceedings-article" @default.
- W2893370043 hasAuthorship W2893370043A5055583643 @default.
- W2893370043 hasAuthorship W2893370043A5056667180 @default.
- W2893370043 hasAuthorship W2893370043A5089667871 @default.
- W2893370043 hasBestOaLocation W28933700432 @default.
- W2893370043 hasConcept C11413529 @default.
- W2893370043 hasConcept C151730666 @default.
- W2893370043 hasConcept C153180895 @default.
- W2893370043 hasConcept C154945302 @default.
- W2893370043 hasConcept C155512373 @default.
- W2893370043 hasConcept C199360897 @default.
- W2893370043 hasConcept C2779343474 @default.
- W2893370043 hasConcept C28490314 @default.
- W2893370043 hasConcept C34736171 @default.
- W2893370043 hasConcept C41008148 @default.
- W2893370043 hasConcept C43521106 @default.
- W2893370043 hasConcept C50644808 @default.
- W2893370043 hasConcept C86803240 @default.
- W2893370043 hasConcept C95623464 @default.
- W2893370043 hasConceptScore W2893370043C11413529 @default.
- W2893370043 hasConceptScore W2893370043C151730666 @default.
- W2893370043 hasConceptScore W2893370043C153180895 @default.
- W2893370043 hasConceptScore W2893370043C154945302 @default.
- W2893370043 hasConceptScore W2893370043C155512373 @default.
- W2893370043 hasConceptScore W2893370043C199360897 @default.
- W2893370043 hasConceptScore W2893370043C2779343474 @default.
- W2893370043 hasConceptScore W2893370043C28490314 @default.
- W2893370043 hasConceptScore W2893370043C34736171 @default.
- W2893370043 hasConceptScore W2893370043C41008148 @default.
- W2893370043 hasConceptScore W2893370043C43521106 @default.
- W2893370043 hasConceptScore W2893370043C50644808 @default.
- W2893370043 hasConceptScore W2893370043C86803240 @default.
- W2893370043 hasConceptScore W2893370043C95623464 @default.
- W2893370043 hasLocation W28933700431 @default.
- W2893370043 hasLocation W28933700432 @default.
- W2893370043 hasOpenAccess W2893370043 @default.
- W2893370043 hasPrimaryLocation W28933700431 @default.
- W2893370043 hasRelatedWork W2039247195 @default.
- W2893370043 hasRelatedWork W2066259560 @default.
- W2893370043 hasRelatedWork W2185397297 @default.
- W2893370043 hasRelatedWork W2262783296 @default.
- W2893370043 hasRelatedWork W2780697200 @default.
- W2893370043 hasRelatedWork W2786849853 @default.
- W2893370043 hasRelatedWork W3129710645 @default.
- W2893370043 hasRelatedWork W3197541072 @default.
- W2893370043 hasRelatedWork W4211209597 @default.
- W2893370043 hasRelatedWork W7626849 @default.
- W2893370043 isParatext "false" @default.
- W2893370043 isRetracted "false" @default.
- W2893370043 magId "2893370043" @default.
- W2893370043 workType "article" @default.