Matches in SemOpenAlex for { <https://semopenalex.org/work/W3207041304> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W3207041304 abstract "Multilingual end-to-end(E2E) models have shown a great potential in the expansion of the language coverage in the realm of automatic speech recognition(ASR). In this paper, we aim to enhance the multilingual ASR performance in two ways, 1)studying the impact of feeding a one-hot vector identifying the language, 2)formulating the task with a meta-learning objective combined with self-supervised learning (SSL). We associate every language with a distinct task manifold and attempt to improve the performance by transferring knowledge across learning processes itself as compared to transferring through final model parameters. We employ this strategy on a dataset comprising of 6 languages for an in-domain ASR task, by minimizing an objective related to expected gradient path length. Experimental results reveal the best pre-training strategy resulting in 3.55% relative reduction in overall WER. A combination of LEAP and SSL yields 3.51% relative reduction in overall WER when using language ID." @default.
- W3207041304 created "2021-10-25" @default.
- W3207041304 creator A5015824704 @default.
- W3207041304 creator A5037672539 @default.
- W3207041304 creator A5053915317 @default.
- W3207041304 creator A5090360853 @default.
- W3207041304 date "2021-10-15" @default.
- W3207041304 modified "2023-09-27" @default.
- W3207041304 title "Multilingual Speech Recognition using Knowledge Transfer across Learning Processes" @default.
- W3207041304 cites W1978660892 @default.
- W3207041304 cites W2025401819 @default.
- W3207041304 cites W2147768505 @default.
- W3207041304 cites W2276408190 @default.
- W3207041304 cites W2291975472 @default.
- W3207041304 cites W2601450892 @default.
- W3207041304 cites W2716988359 @default.
- W3207041304 cites W2786835190 @default.
- W3207041304 cites W2902253437 @default.
- W3207041304 cites W2962893195 @default.
- W3207041304 cites W2963649564 @default.
- W3207041304 cites W2964002616 @default.
- W3207041304 cites W2964112702 @default.
- W3207041304 cites W2964309797 @default.
- W3207041304 cites W2971840980 @default.
- W3207041304 cites W3008434450 @default.
- W3207041304 cites W3016010032 @default.
- W3207041304 cites W3037057938 @default.
- W3207041304 cites W3093579165 @default.
- W3207041304 cites W3095311338 @default.
- W3207041304 cites W3096215352 @default.
- W3207041304 cites W3156323585 @default.
- W3207041304 cites W3167207712 @default.
- W3207041304 cites W3178647810 @default.
- W3207041304 cites W3204696009 @default.
- W3207041304 doi "https://doi.org/10.48550/arxiv.2110.07909" @default.
- W3207041304 hasPublicationYear "2021" @default.
- W3207041304 type Work @default.
- W3207041304 sameAs 3207041304 @default.
- W3207041304 citedByCount "0" @default.
- W3207041304 crossrefType "posted-content" @default.
- W3207041304 hasAuthorship W3207041304A5015824704 @default.
- W3207041304 hasAuthorship W3207041304A5037672539 @default.
- W3207041304 hasAuthorship W3207041304A5053915317 @default.
- W3207041304 hasAuthorship W3207041304A5090360853 @default.
- W3207041304 hasBestOaLocation W32070413041 @default.
- W3207041304 hasConcept C111335779 @default.
- W3207041304 hasConcept C134306372 @default.
- W3207041304 hasConcept C137293760 @default.
- W3207041304 hasConcept C150899416 @default.
- W3207041304 hasConcept C154945302 @default.
- W3207041304 hasConcept C162324750 @default.
- W3207041304 hasConcept C17744445 @default.
- W3207041304 hasConcept C187736073 @default.
- W3207041304 hasConcept C199539241 @default.
- W3207041304 hasConcept C204321447 @default.
- W3207041304 hasConcept C2524010 @default.
- W3207041304 hasConcept C2778757428 @default.
- W3207041304 hasConcept C2780451532 @default.
- W3207041304 hasConcept C28490314 @default.
- W3207041304 hasConcept C33923547 @default.
- W3207041304 hasConcept C36503486 @default.
- W3207041304 hasConcept C41008148 @default.
- W3207041304 hasConceptScore W3207041304C111335779 @default.
- W3207041304 hasConceptScore W3207041304C134306372 @default.
- W3207041304 hasConceptScore W3207041304C137293760 @default.
- W3207041304 hasConceptScore W3207041304C150899416 @default.
- W3207041304 hasConceptScore W3207041304C154945302 @default.
- W3207041304 hasConceptScore W3207041304C162324750 @default.
- W3207041304 hasConceptScore W3207041304C17744445 @default.
- W3207041304 hasConceptScore W3207041304C187736073 @default.
- W3207041304 hasConceptScore W3207041304C199539241 @default.
- W3207041304 hasConceptScore W3207041304C204321447 @default.
- W3207041304 hasConceptScore W3207041304C2524010 @default.
- W3207041304 hasConceptScore W3207041304C2778757428 @default.
- W3207041304 hasConceptScore W3207041304C2780451532 @default.
- W3207041304 hasConceptScore W3207041304C28490314 @default.
- W3207041304 hasConceptScore W3207041304C33923547 @default.
- W3207041304 hasConceptScore W3207041304C36503486 @default.
- W3207041304 hasConceptScore W3207041304C41008148 @default.
- W3207041304 hasLocation W32070413041 @default.
- W3207041304 hasOpenAccess W3207041304 @default.
- W3207041304 hasPrimaryLocation W32070413041 @default.
- W3207041304 hasRelatedWork W2081647779 @default.
- W3207041304 hasRelatedWork W3093097038 @default.
- W3207041304 hasRelatedWork W3107474891 @default.
- W3207041304 hasRelatedWork W3164871893 @default.
- W3207041304 hasRelatedWork W3172706523 @default.
- W3207041304 hasRelatedWork W3185852197 @default.
- W3207041304 hasRelatedWork W3186237585 @default.
- W3207041304 hasRelatedWork W4205820553 @default.
- W3207041304 hasRelatedWork W4283219089 @default.
- W3207041304 hasRelatedWork W4226264439 @default.
- W3207041304 isParatext "false" @default.
- W3207041304 isRetracted "false" @default.
- W3207041304 magId "3207041304" @default.
- W3207041304 workType "article" @default.