Matches in SemOpenAlex for { <https://semopenalex.org/work/W2797969439> ?p ?o ?g. }
- W2797969439 abstract "In this study, we develop the keyword spotting (KWS) and acoustic model (AM) components in a far-field speaker system. Specifically, we use teacher-student (T/S) learning to adapt a close-talk well-trained production AM to far-field by using parallel close-talk and simulated far-field data. We also use T/S learning to compress a large-size KWS model into a small-size one to fit the device computational cost. Without the need of transcription, T/S learning well utilizes untranscribed data to boost the model performance in both the AM adaptation and KWS model compression. We further optimize the models with sequence discriminative training and live data to reach the best performance of systems. The adapted AM improved from the baseline by 72.60% and 57.16% relative word error rate reduction on play-back and live test data, respectively. The final KWS model size was reduced by 27 times from a large-size KWS model without losing accuracy." @default.
- W2797969439 created "2018-04-24" @default.
- W2797969439 creator A5008578397 @default.
- W2797969439 creator A5017825677 @default.
- W2797969439 creator A5022890229 @default.
- W2797969439 creator A5029670581 @default.
- W2797969439 creator A5057945477 @default.
- W2797969439 creator A5077401426 @default.
- W2797969439 creator A5086031494 @default.
- W2797969439 date "2018-04-14" @default.
- W2797969439 modified "2023-10-16" @default.
- W2797969439 title "Developing Far-Field Speaker System Via Teacher-Student Learning" @default.
- W2797969439 cites W1533416326 @default.
- W2797969439 cites W1553469512 @default.
- W2797969439 cites W1567690964 @default.
- W2797969439 cites W1571447047 @default.
- W2797969439 cites W1770758908 @default.
- W2797969439 cites W1821462560 @default.
- W2797969439 cites W1899458394 @default.
- W2797969439 cites W1987238397 @default.
- W2797969439 cites W1992475611 @default.
- W2797969439 cites W2010362084 @default.
- W2797969439 cites W2056738732 @default.
- W2797969439 cites W2064675550 @default.
- W2797969439 cites W2117239706 @default.
- W2797969439 cites W2117678320 @default.
- W2797969439 cites W2127141656 @default.
- W2797969439 cites W2136933783 @default.
- W2797969439 cites W2160815625 @default.
- W2797969439 cites W2289394825 @default.
- W2797969439 cites W2293009711 @default.
- W2797969439 cites W2293634267 @default.
- W2797969439 cites W2294543795 @default.
- W2797969439 cites W2296516251 @default.
- W2797969439 cites W2396384435 @default.
- W2797969439 cites W2400108189 @default.
- W2797969439 cites W2402040300 @default.
- W2797969439 cites W2507699225 @default.
- W2797969439 cites W2508162385 @default.
- W2797969439 cites W2512865187 @default.
- W2797969439 cites W2559260703 @default.
- W2797969439 cites W2735006420 @default.
- W2797969439 cites W2749659773 @default.
- W2797969439 cites W2749784707 @default.
- W2797969439 cites W2787290364 @default.
- W2797969439 cites W2791164268 @default.
- W2797969439 cites W2793056360 @default.
- W2797969439 doi "https://doi.org/10.48550/arxiv.1804.05166" @default.
- W2797969439 hasPublicationYear "2018" @default.
- W2797969439 type Work @default.
- W2797969439 sameAs 2797969439 @default.
- W2797969439 citedByCount "2" @default.
- W2797969439 countsByYear W27979694392018 @default.
- W2797969439 countsByYear W27979694392020 @default.
- W2797969439 crossrefType "posted-content" @default.
- W2797969439 hasAuthorship W2797969439A5008578397 @default.
- W2797969439 hasAuthorship W2797969439A5017825677 @default.
- W2797969439 hasAuthorship W2797969439A5022890229 @default.
- W2797969439 hasAuthorship W2797969439A5029670581 @default.
- W2797969439 hasAuthorship W2797969439A5057945477 @default.
- W2797969439 hasAuthorship W2797969439A5077401426 @default.
- W2797969439 hasAuthorship W2797969439A5086031494 @default.
- W2797969439 hasBestOaLocation W27979694391 @default.
- W2797969439 hasConcept C108583219 @default.
- W2797969439 hasConcept C111335779 @default.
- W2797969439 hasConcept C111368507 @default.
- W2797969439 hasConcept C119857082 @default.
- W2797969439 hasConcept C120665830 @default.
- W2797969439 hasConcept C121332964 @default.
- W2797969439 hasConcept C12725497 @default.
- W2797969439 hasConcept C127313418 @default.
- W2797969439 hasConcept C138885662 @default.
- W2797969439 hasConcept C139807058 @default.
- W2797969439 hasConcept C154945302 @default.
- W2797969439 hasConcept C179926584 @default.
- W2797969439 hasConcept C202444582 @default.
- W2797969439 hasConcept C2524010 @default.
- W2797969439 hasConcept C2779506182 @default.
- W2797969439 hasConcept C2781213101 @default.
- W2797969439 hasConcept C28490314 @default.
- W2797969439 hasConcept C33923547 @default.
- W2797969439 hasConcept C40969351 @default.
- W2797969439 hasConcept C41008148 @default.
- W2797969439 hasConcept C41895202 @default.
- W2797969439 hasConcept C90805587 @default.
- W2797969439 hasConcept C9652623 @default.
- W2797969439 hasConcept C97931131 @default.
- W2797969439 hasConceptScore W2797969439C108583219 @default.
- W2797969439 hasConceptScore W2797969439C111335779 @default.
- W2797969439 hasConceptScore W2797969439C111368507 @default.
- W2797969439 hasConceptScore W2797969439C119857082 @default.
- W2797969439 hasConceptScore W2797969439C120665830 @default.
- W2797969439 hasConceptScore W2797969439C121332964 @default.
- W2797969439 hasConceptScore W2797969439C12725497 @default.
- W2797969439 hasConceptScore W2797969439C127313418 @default.
- W2797969439 hasConceptScore W2797969439C138885662 @default.
- W2797969439 hasConceptScore W2797969439C139807058 @default.
- W2797969439 hasConceptScore W2797969439C154945302 @default.
- W2797969439 hasConceptScore W2797969439C179926584 @default.
- W2797969439 hasConceptScore W2797969439C202444582 @default.