FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100910367> ?p ?o ?g. }

• W3100910367 abstract "Following the rationale of end-to-end modeling, CTC, RNN-T or encoder-decoder-attention models for automatic speech recognition (ASR) use graphemes or grapheme-based subword units based on e.g. byte-pair encoding (BPE). The mapping from pronunciation to spelling is learned completely from data. In contrast to this, classical approaches to ASR employ secondary knowledge sources in the form of phoneme lists to define phonetic output labels and pronunciation lexica. In this work, we do a systematic comparison between grapheme- and phoneme-based output labels for an encoder-decoder-attention ASR model. We investigate the use of single phonemes as well as BPE-based phoneme groups as output labels of our model. To preserve a simplified and efficient decoder design, we also extend the phoneme set by auxiliary units to be able to distinguish homophones. Experiments performed on the Switchboard 300h and LibriSpeech benchmarks show that phoneme-based modeling is competitive to grapheme-based encoder-decoder-attention modeling." @default.
• W3100910367 created "2020-11-23" @default.
• W3100910367 creator A5025049641 @default.
• W3100910367 creator A5059521407 @default.
• W3100910367 creator A5060987315 @default.
• W3100910367 creator A5087367411 @default.
• W3100910367 creator A5088968292 @default.
• W3100910367 creator A5089693342 @default.
• W3100910367 date "2020-05-19" @default.
• W3100910367 modified "2023-09-27" @default.
• W3100910367 title "A systematic comparison of grapheme-based vs. phoneme-based label units for encoder-decoder-attention models" @default.
• W3100910367 cites W1494198834 @default.
• W3100910367 cites W1510837697 @default.
• W3100910367 cites W1828163288 @default.
• W3100910367 cites W1902237438 @default.
• W3100910367 cites W2044223162 @default.
• W3100910367 cites W2064675550 @default.
• W3100910367 cites W2110871230 @default.
• W3100910367 cites W2117541506 @default.
• W3100910367 cites W2121879602 @default.
• W3100910367 cites W2127141656 @default.
• W3100910367 cites W2166637769 @default.
• W3100910367 cites W2193413348 @default.
• W3100910367 cites W2327501763 @default.
• W3100910367 cites W2395416438 @default.
• W3100910367 cites W2471933213 @default.
• W3100910367 cites W2520160253 @default.
• W3100910367 cites W2525778437 @default.
• W3100910367 cites W2530486890 @default.
• W3100910367 cites W2545177271 @default.
• W3100910367 cites W2597757402 @default.
• W3100910367 cites W2766219058 @default.
• W3100910367 cites W2799800213 @default.
• W3100910367 cites W2884483159 @default.
• W3100910367 cites W2899879954 @default.
• W3100910367 cites W2905489173 @default.
• W3100910367 cites W2937402758 @default.
• W3100910367 cites W2952862739 @default.
• W3100910367 cites W2962784628 @default.
• W3100910367 cites W2962961016 @default.
• W3100910367 cites W2963070863 @default.
• W3100910367 cites W2963850025 @default.
• W3100910367 cites W2963920996 @default.
• W3100910367 cites W2963979492 @default.
• W3100910367 cites W2964107261 @default.
• W3100910367 cites W2964308564 @default.
• W3100910367 cites W2966163367 @default.
• W3100910367 cites W3002595344 @default.
• W3100910367 cites W3008525923 @default.
• W3100910367 cites W3015349902 @default.
• W3100910367 cites W3015889230 @default.
• W3100910367 cites W3016190221 @default.
• W3100910367 cites W3016234571 @default.
• W3100910367 cites W3028545098 @default.
• W3100910367 cites W3101140821 @default.
• W3100910367 cites W3150637114 @default.
• W3100910367 cites W3160551958 @default.
• W3100910367 cites W46679369 @default.
• W3100910367 hasPublicationYear "2020" @default.
• W3100910367 type Work @default.
• W3100910367 sameAs 3100910367 @default.
• W3100910367 citedByCount "2" @default.
• W3100910367 countsByYear W31009103672021 @default.
• W3100910367 crossrefType "posted-content" @default.
• W3100910367 hasAuthorship W3100910367A5025049641 @default.
• W3100910367 hasAuthorship W3100910367A5059521407 @default.
• W3100910367 hasAuthorship W3100910367A5060987315 @default.
• W3100910367 hasAuthorship W3100910367A5087367411 @default.
• W3100910367 hasAuthorship W3100910367A5088968292 @default.
• W3100910367 hasAuthorship W3100910367A5089693342 @default.
• W3100910367 hasConcept C111919701 @default.
• W3100910367 hasConcept C118505674 @default.
• W3100910367 hasConcept C121332964 @default.
• W3100910367 hasConcept C137293760 @default.
• W3100910367 hasConcept C138885662 @default.
• W3100910367 hasConcept C154945302 @default.
• W3100910367 hasConcept C160253069 @default.
• W3100910367 hasConcept C177264268 @default.
• W3100910367 hasConcept C199360897 @default.
• W3100910367 hasConcept C204321447 @default.
• W3100910367 hasConcept C2776779415 @default.
• W3100910367 hasConcept C2780844864 @default.
• W3100910367 hasConcept C28490314 @default.
• W3100910367 hasConcept C30080830 @default.
• W3100910367 hasConcept C41008148 @default.
• W3100910367 hasConcept C41895202 @default.
• W3100910367 hasConcept C50644808 @default.
• W3100910367 hasConcept C62520636 @default.
• W3100910367 hasConcept C8521452 @default.
• W3100910367 hasConceptScore W3100910367C111919701 @default.
• W3100910367 hasConceptScore W3100910367C118505674 @default.
• W3100910367 hasConceptScore W3100910367C121332964 @default.
• W3100910367 hasConceptScore W3100910367C137293760 @default.
• W3100910367 hasConceptScore W3100910367C138885662 @default.
• W3100910367 hasConceptScore W3100910367C154945302 @default.
• W3100910367 hasConceptScore W3100910367C160253069 @default.
• W3100910367 hasConceptScore W3100910367C177264268 @default.
• W3100910367 hasConceptScore W3100910367C199360897 @default.
• W3100910367 hasConceptScore W3100910367C204321447 @default.
• W3100910367 hasConceptScore W3100910367C2776779415 @default.

• next