FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2981536480> ?p ?o ?g. }

• W2981536480 abstract "Supervised ASR models have reached unprecedented levels of accuracy, thanks in part to ever-increasing amounts of labelled training data. However, in many applications and locales, only moderate amounts of data are available, which has led to a surge in semi- and weakly-supervised learning research. In this paper, we conduct a large-scale study evaluating the effectiveness of weakly-supervised learning for speech recognition by using loosely related contextual information as a surrogate for ground-truth labels. For weakly supervised training, we use 50k hours of public English social media videos along with their respective titles and post text to train an encoder-decoder transformer model. Our best encoder-decoder models achieve an average of 20.8% WER reduction over a 1000 hours supervised baseline, and an average of 13.4% WER reduction when using only the weakly supervised encoder for CTC fine-tuning. Our results show that our setup for weak supervision improved both the encoder acoustic representations as well as the decoder language generation abilities." @default.
• W2981536480 created "2019-11-01" @default.
• W2981536480 creator A5014592284 @default.
• W2981536480 creator A5016113002 @default.
• W2981536480 creator A5022671092 @default.
• W2981536480 creator A5027835055 @default.
• W2981536480 creator A5029332181 @default.
• W2981536480 creator A5042186348 @default.
• W2981536480 creator A5047400593 @default.
• W2981536480 creator A5049246408 @default.
• W2981536480 creator A5051165898 @default.
• W2981536480 creator A5069954850 @default.
• W2981536480 creator A5085650357 @default.
• W2981536480 date "2019-10-27" @default.
• W2981536480 modified "2023-10-01" @default.
• W2981536480 title "Training ASR models by Generation of Contextual Information" @default.
• W2981536480 cites W1525264136 @default.
• W2981536480 cites W1978660892 @default.
• W2981536480 cites W1994606281 @default.
• W2981536480 cites W2025198378 @default.
• W2981536480 cites W2025482506 @default.
• W2981536480 cites W2033256038 @default.
• W2981536480 cites W2055408826 @default.
• W2981536480 cites W2100031962 @default.
• W2981536480 cites W2102113734 @default.
• W2981536480 cites W2106440210 @default.
• W2981536480 cites W2114347655 @default.
• W2981536480 cites W2124558353 @default.
• W2981536480 cites W2127141656 @default.
• W2981536480 cites W2160815625 @default.
• W2981536480 cites W2193413348 @default.
• W2981536480 cites W2250874882 @default.
• W2981536480 cites W2291975472 @default.
• W2981536480 cites W2327501763 @default.
• W2981536480 cites W2399093011 @default.
• W2981536480 cites W2514608284 @default.
• W2981536480 cites W2520160253 @default.
• W2981536480 cites W2556930864 @default.
• W2981536480 cites W2667408400 @default.
• W2981536480 cites W2759071281 @default.
• W2981536480 cites W2842511635 @default.
• W2981536480 cites W2892009249 @default.
• W2981536480 cites W2940322076 @default.
• W2981536480 cites W2941814890 @default.
• W2981536480 cites W2962756039 @default.
• W2981536480 cites W2962907457 @default.
• W2981536480 cites W2963096510 @default.
• W2981536480 cites W2963206148 @default.
• W2981536480 cites W2963250244 @default.
• W2981536480 cites W2963266252 @default.
• W2981536480 cites W2963341956 @default.
• W2981536480 cites W2963403868 @default.
• W2981536480 cites W2963425185 @default.
• W2981536480 cites W2963583362 @default.
• W2981536480 cites W2963703197 @default.
• W2981536480 cites W2964001192 @default.
• W2981536480 cites W2964115348 @default.
• W2981536480 cites W2965373594 @default.
• W2981536480 cites W2973049979 @default.
• W2981536480 cites W30845872 @default.
• W2981536480 cites W385555557 @default.
• W2981536480 cites W6908809 @default.
• W2981536480 cites W854541894 @default.
• W2981536480 hasPublicationYear "2019" @default.
• W2981536480 type Work @default.
• W2981536480 sameAs 2981536480 @default.
• W2981536480 citedByCount "4" @default.
• W2981536480 countsByYear W29815364802020 @default.
• W2981536480 countsByYear W29815364802021 @default.
• W2981536480 crossrefType "posted-content" @default.
• W2981536480 hasAuthorship W2981536480A5014592284 @default.
• W2981536480 hasAuthorship W2981536480A5016113002 @default.
• W2981536480 hasAuthorship W2981536480A5022671092 @default.
• W2981536480 hasAuthorship W2981536480A5027835055 @default.
• W2981536480 hasAuthorship W2981536480A5029332181 @default.
• W2981536480 hasAuthorship W2981536480A5042186348 @default.
• W2981536480 hasAuthorship W2981536480A5047400593 @default.
• W2981536480 hasAuthorship W2981536480A5049246408 @default.
• W2981536480 hasAuthorship W2981536480A5051165898 @default.
• W2981536480 hasAuthorship W2981536480A5069954850 @default.
• W2981536480 hasAuthorship W2981536480A5085650357 @default.
• W2981536480 hasConcept C111335779 @default.
• W2981536480 hasConcept C111368507 @default.
• W2981536480 hasConcept C111919701 @default.
• W2981536480 hasConcept C118505674 @default.
• W2981536480 hasConcept C119599485 @default.
• W2981536480 hasConcept C119857082 @default.
• W2981536480 hasConcept C12725497 @default.
• W2981536480 hasConcept C127313418 @default.
• W2981536480 hasConcept C127413603 @default.
• W2981536480 hasConcept C136389625 @default.
• W2981536480 hasConcept C137293760 @default.
• W2981536480 hasConcept C146849305 @default.
• W2981536480 hasConcept C154945302 @default.
• W2981536480 hasConcept C165801399 @default.
• W2981536480 hasConcept C204321447 @default.
• W2981536480 hasConcept C2524010 @default.
• W2981536480 hasConcept C2776145971 @default.
• W2981536480 hasConcept C28490314 @default.
• W2981536480 hasConcept C33923547 @default.

• next