Matches in SemOpenAlex for { <https://semopenalex.org/work/W4319862426> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W4319862426 abstract "Automatic speech recognition (ASR) systems typically rely on an external endpointer (EP) model to identify speech boundaries. In this work, we propose a method to jointly train the ASR and EP tasks in a single end-to-end (E2E) multitask model, improving EP quality by optionally leveraging information from the ASR audio encoder. We introduce a “switch” connection, which trains the EP to consume either the audio frames directly or low-level latent representations from the ASR model. This results in a single E2E model that can be used during inference to perform frame filtering at low cost, and also make high quality end-of-query (EOQ) predictions based on ongoing ASR computation. We present results on a voice search test set showing that, compared to separate single-task models, this approach reduces median endpoint latency by 120 ms (30.8% reduction), and 90th percentile latency by 170 ms (23.0% reduction), without regressing word error rate. For continuous recognition, WER improves by 10.6% (relative)." @default.
- W4319862426 created "2023-02-11" @default.
- W4319862426 creator A5001306222 @default.
- W4319862426 creator A5006160562 @default.
- W4319862426 creator A5034995105 @default.
- W4319862426 creator A5057214137 @default.
- W4319862426 creator A5070513394 @default.
- W4319862426 creator A5089966579 @default.
- W4319862426 date "2023-01-09" @default.
- W4319862426 modified "2023-09-27" @default.
- W4319862426 title "Unified End-to-End Speech Recognition and Endpointing for Fast and Efficient Speech Systems" @default.
- W4319862426 cites W10731371 @default.
- W4319862426 cites W1999454387 @default.
- W4319862426 cites W2064675550 @default.
- W4319862426 cites W2100844198 @default.
- W4319862426 cites W2127141656 @default.
- W4319862426 cites W2157331050 @default.
- W4319862426 cites W2408468399 @default.
- W4319862426 cites W2606722458 @default.
- W4319862426 cites W2742061524 @default.
- W4319862426 cites W2891367150 @default.
- W4319862426 cites W2912003438 @default.
- W4319862426 cites W2935756939 @default.
- W4319862426 cites W2962760690 @default.
- W4319862426 cites W3015383801 @default.
- W4319862426 cites W3015927303 @default.
- W4319862426 cites W3016010032 @default.
- W4319862426 cites W3086863954 @default.
- W4319862426 cites W3097777922 @default.
- W4319862426 cites W3161080229 @default.
- W4319862426 cites W3163907627 @default.
- W4319862426 cites W3193786050 @default.
- W4319862426 cites W3196784225 @default.
- W4319862426 cites W4221147513 @default.
- W4319862426 cites W4225307083 @default.
- W4319862426 doi "https://doi.org/10.1109/slt54892.2023.10022338" @default.
- W4319862426 hasPublicationYear "2023" @default.
- W4319862426 type Work @default.
- W4319862426 citedByCount "0" @default.
- W4319862426 crossrefType "proceedings-article" @default.
- W4319862426 hasAuthorship W4319862426A5001306222 @default.
- W4319862426 hasAuthorship W4319862426A5006160562 @default.
- W4319862426 hasAuthorship W4319862426A5034995105 @default.
- W4319862426 hasAuthorship W4319862426A5057214137 @default.
- W4319862426 hasAuthorship W4319862426A5070513394 @default.
- W4319862426 hasAuthorship W4319862426A5089966579 @default.
- W4319862426 hasBestOaLocation W43198624262 @default.
- W4319862426 hasConcept C111919701 @default.
- W4319862426 hasConcept C118505674 @default.
- W4319862426 hasConcept C137293760 @default.
- W4319862426 hasConcept C154945302 @default.
- W4319862426 hasConcept C204201278 @default.
- W4319862426 hasConcept C2776214188 @default.
- W4319862426 hasConcept C28490314 @default.
- W4319862426 hasConcept C40969351 @default.
- W4319862426 hasConcept C41008148 @default.
- W4319862426 hasConcept C61328038 @default.
- W4319862426 hasConcept C74296488 @default.
- W4319862426 hasConcept C76155785 @default.
- W4319862426 hasConcept C82876162 @default.
- W4319862426 hasConceptScore W4319862426C111919701 @default.
- W4319862426 hasConceptScore W4319862426C118505674 @default.
- W4319862426 hasConceptScore W4319862426C137293760 @default.
- W4319862426 hasConceptScore W4319862426C154945302 @default.
- W4319862426 hasConceptScore W4319862426C204201278 @default.
- W4319862426 hasConceptScore W4319862426C2776214188 @default.
- W4319862426 hasConceptScore W4319862426C28490314 @default.
- W4319862426 hasConceptScore W4319862426C40969351 @default.
- W4319862426 hasConceptScore W4319862426C41008148 @default.
- W4319862426 hasConceptScore W4319862426C61328038 @default.
- W4319862426 hasConceptScore W4319862426C74296488 @default.
- W4319862426 hasConceptScore W4319862426C76155785 @default.
- W4319862426 hasConceptScore W4319862426C82876162 @default.
- W4319862426 hasLocation W43198624261 @default.
- W4319862426 hasLocation W43198624262 @default.
- W4319862426 hasOpenAccess W4319862426 @default.
- W4319862426 hasPrimaryLocation W43198624261 @default.
- W4319862426 hasRelatedWork W2782005958 @default.
- W4319862426 hasRelatedWork W3012645992 @default.
- W4319862426 hasRelatedWork W3160766462 @default.
- W4319862426 hasRelatedWork W3184187848 @default.
- W4319862426 hasRelatedWork W3197304116 @default.
- W4319862426 hasRelatedWork W3209371554 @default.
- W4319862426 hasRelatedWork W4300537895 @default.
- W4319862426 hasRelatedWork W4308165856 @default.
- W4319862426 hasRelatedWork W4319431239 @default.
- W4319862426 hasRelatedWork W4319862426 @default.
- W4319862426 isParatext "false" @default.
- W4319862426 isRetracted "false" @default.
- W4319862426 workType "article" @default.