Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313483363> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4313483363 abstract "Recent studies have shown that using an external Language Model (LM) benefits the end-to-end Automatic Speech Recognition (ASR). However, predicting tokens that appear less frequently in the training set is still quite challenging. The long-tail prediction problems have been widely studied in many applications, but only been addressed by a few studies for ASR and LMs. In this paper, we propose a new memory augmented lookup dictionary based Transformer architecture for LM. The newly introduced lookup dictionary incorporates rich contextual information in training set, which is vital to correctly predict long-tail tokens. With intensive experiments on Chinese and English data sets, our proposed method is proved to outperform the baseline Transformer LM by a great margin on both word/character error rate and tail tokens error rate. This is achieved without impact on the decoding efficiency. Overall, we demonstrate the effectiveness of our proposed method in boosting the ASR decoding performance, especially for long-tail tokens." @default.
- W4313483363 created "2023-01-06" @default.
- W4313483363 creator A5003012119 @default.
- W4313483363 creator A5011370924 @default.
- W4313483363 creator A5018343941 @default.
- W4313483363 creator A5035742417 @default.
- W4313483363 creator A5040480051 @default.
- W4313483363 date "2022-12-30" @default.
- W4313483363 modified "2023-09-30" @default.
- W4313483363 title "Memory Augmented Lookup Dictionary based Language Modeling for Automatic Speech Recognition" @default.
- W4313483363 doi "https://doi.org/10.48550/arxiv.2301.00066" @default.
- W4313483363 hasPublicationYear "2022" @default.
- W4313483363 type Work @default.
- W4313483363 citedByCount "0" @default.
- W4313483363 crossrefType "posted-content" @default.
- W4313483363 hasAuthorship W4313483363A5003012119 @default.
- W4313483363 hasAuthorship W4313483363A5011370924 @default.
- W4313483363 hasAuthorship W4313483363A5018343941 @default.
- W4313483363 hasAuthorship W4313483363A5035742417 @default.
- W4313483363 hasAuthorship W4313483363A5040480051 @default.
- W4313483363 hasBestOaLocation W43134833631 @default.
- W4313483363 hasConcept C11413529 @default.
- W4313483363 hasConcept C121332964 @default.
- W4313483363 hasConcept C137293760 @default.
- W4313483363 hasConcept C138885662 @default.
- W4313483363 hasConcept C154945302 @default.
- W4313483363 hasConcept C165801399 @default.
- W4313483363 hasConcept C177264268 @default.
- W4313483363 hasConcept C199360897 @default.
- W4313483363 hasConcept C204321447 @default.
- W4313483363 hasConcept C28490314 @default.
- W4313483363 hasConcept C40969351 @default.
- W4313483363 hasConcept C41008148 @default.
- W4313483363 hasConcept C41895202 @default.
- W4313483363 hasConcept C46686674 @default.
- W4313483363 hasConcept C51632099 @default.
- W4313483363 hasConcept C57273362 @default.
- W4313483363 hasConcept C62520636 @default.
- W4313483363 hasConcept C66322947 @default.
- W4313483363 hasConcept C90805587 @default.
- W4313483363 hasConceptScore W4313483363C11413529 @default.
- W4313483363 hasConceptScore W4313483363C121332964 @default.
- W4313483363 hasConceptScore W4313483363C137293760 @default.
- W4313483363 hasConceptScore W4313483363C138885662 @default.
- W4313483363 hasConceptScore W4313483363C154945302 @default.
- W4313483363 hasConceptScore W4313483363C165801399 @default.
- W4313483363 hasConceptScore W4313483363C177264268 @default.
- W4313483363 hasConceptScore W4313483363C199360897 @default.
- W4313483363 hasConceptScore W4313483363C204321447 @default.
- W4313483363 hasConceptScore W4313483363C28490314 @default.
- W4313483363 hasConceptScore W4313483363C40969351 @default.
- W4313483363 hasConceptScore W4313483363C41008148 @default.
- W4313483363 hasConceptScore W4313483363C41895202 @default.
- W4313483363 hasConceptScore W4313483363C46686674 @default.
- W4313483363 hasConceptScore W4313483363C51632099 @default.
- W4313483363 hasConceptScore W4313483363C57273362 @default.
- W4313483363 hasConceptScore W4313483363C62520636 @default.
- W4313483363 hasConceptScore W4313483363C66322947 @default.
- W4313483363 hasConceptScore W4313483363C90805587 @default.
- W4313483363 hasLocation W43134833631 @default.
- W4313483363 hasOpenAccess W4313483363 @default.
- W4313483363 hasPrimaryLocation W43134833631 @default.
- W4313483363 hasRelatedWork W1578024259 @default.
- W4313483363 hasRelatedWork W1826521293 @default.
- W4313483363 hasRelatedWork W2055254539 @default.
- W4313483363 hasRelatedWork W2100441082 @default.
- W4313483363 hasRelatedWork W2597757402 @default.
- W4313483363 hasRelatedWork W3080136773 @default.
- W4313483363 hasRelatedWork W3107474891 @default.
- W4313483363 hasRelatedWork W3156915121 @default.
- W4313483363 hasRelatedWork W3197303984 @default.
- W4313483363 hasRelatedWork W3197898596 @default.
- W4313483363 isParatext "false" @default.
- W4313483363 isRetracted "false" @default.
- W4313483363 workType "article" @default.