FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2496955520> ?p ?o ?g. }

• W2496955520 endingPage "2157" @default.
• W2496955520 startingPage "2146" @default.
• W2496955520 abstract "Recurrent neural network language models RNNLMs are becoming increasingly popular for a range of applications including automatic speech recognition. An important issue that limits their possible application areas is the computational cost incurred in training and evaluation. This paper describes a series of new efficiency improving approaches that allows RNNLMs to be more efficiently trained on graphics processing units GPUs and evaluated on CPUs. First, a modified RNNLM architecture with a nonclass-based, full output layer structure F-RNNLM is proposed. This modified architecture facilitates a novel spliced sentence bunch mode parallelization of F-RNNLM training using large quantities of data on a GPU. Second, two efficient RNNLM training criteria based on variance regularization and noise contrastive estimation are explored to specifically reduce the computation associated with the RNNLM output layer softmax normalisation term. Finally, a pipelined training algorithm utilizing multiple GPUs is also used to further improve the training speed. Initially, RNNLMs were trained on a moderate dataset with 20M words from a large vocabulary conversational telephone speech recognition task. The training time of RNNLM is reduced by up to a factor of 53 on a single GPU over the standard CPU-based RNNLM toolkit. A 56 times speed up in test time evaluation on a CPU was obtained over the baseline F-RNNLMs. Consistent improvements in both recognition accuracy and perplexity were also obtained over C-RNNLMs. Experiments on Google's one billion corpus also reveals that the training of RNNLM scales well." @default.
• W2496955520 created "2016-08-23" @default.
• W2496955520 creator A5002191410 @default.
• W2496955520 creator A5007473150 @default.
• W2496955520 creator A5014592284 @default.
• W2496955520 creator A5037109470 @default.
• W2496955520 creator A5050766679 @default.
• W2496955520 date "2016-11-01" @default.
• W2496955520 modified "2023-09-22" @default.
• W2496955520 title "Efficient Training and Evaluation of Recurrent Neural Network Language Models for Automatic Speech Recognition" @default.
• W2496955520 cites W1488980900 @default.
• W2496955520 cites W1494108583 @default.
• W2496955520 cites W1520465330 @default.
• W2496955520 cites W1585876329 @default.
• W2496955520 cites W1773599773 @default.
• W2496955520 cites W1917432393 @default.
• W2496955520 cites W1964742985 @default.
• W2496955520 cites W1970689298 @default.
• W2496955520 cites W1985258458 @default.
• W2496955520 cites W1994536225 @default.
• W2496955520 cites W2008398646 @default.
• W2496955520 cites W2026149468 @default.
• W2496955520 cites W2026383756 @default.
• W2496955520 cites W2058695628 @default.
• W2496955520 cites W2066378046 @default.
• W2496955520 cites W2076094076 @default.
• W2496955520 cites W2091981305 @default.
• W2496955520 cites W2094233035 @default.
• W2496955520 cites W2094971681 @default.
• W2496955520 cites W2110415041 @default.
• W2496955520 cites W2134950286 @default.
• W2496955520 cites W2144790469 @default.
• W2496955520 cites W2154137718 @default.
• W2496955520 cites W2161199684 @default.
• W2496955520 cites W2171928131 @default.
• W2496955520 cites W2251682575 @default.
• W2496955520 cites W2394932179 @default.
• W2496955520 doi "https://doi.org/10.1109/taslp.2016.2598304" @default.
• W2496955520 hasPublicationYear "2016" @default.
• W2496955520 type Work @default.
• W2496955520 sameAs 2496955520 @default.
• W2496955520 citedByCount "50" @default.
• W2496955520 countsByYear W24969555202016 @default.
• W2496955520 countsByYear W24969555202017 @default.
• W2496955520 countsByYear W24969555202018 @default.
• W2496955520 countsByYear W24969555202019 @default.
• W2496955520 countsByYear W24969555202020 @default.
• W2496955520 countsByYear W24969555202021 @default.
• W2496955520 countsByYear W24969555202022 @default.
• W2496955520 countsByYear W24969555202023 @default.
• W2496955520 crossrefType "journal-article" @default.
• W2496955520 hasAuthorship W2496955520A5002191410 @default.
• W2496955520 hasAuthorship W2496955520A5007473150 @default.
• W2496955520 hasAuthorship W2496955520A5014592284 @default.
• W2496955520 hasAuthorship W2496955520A5037109470 @default.
• W2496955520 hasAuthorship W2496955520A5050766679 @default.
• W2496955520 hasBestOaLocation W24969555202 @default.
• W2496955520 hasConcept C100279451 @default.
• W2496955520 hasConcept C119857082 @default.
• W2496955520 hasConcept C137293760 @default.
• W2496955520 hasConcept C138885662 @default.
• W2496955520 hasConcept C154945302 @default.
• W2496955520 hasConcept C188441871 @default.
• W2496955520 hasConcept C2777601683 @default.
• W2496955520 hasConcept C28490314 @default.
• W2496955520 hasConcept C41008148 @default.
• W2496955520 hasConcept C41895202 @default.
• W2496955520 hasConcept C50644808 @default.
• W2496955520 hasConceptScore W2496955520C100279451 @default.
• W2496955520 hasConceptScore W2496955520C119857082 @default.
• W2496955520 hasConceptScore W2496955520C137293760 @default.
• W2496955520 hasConceptScore W2496955520C138885662 @default.
• W2496955520 hasConceptScore W2496955520C154945302 @default.
• W2496955520 hasConceptScore W2496955520C188441871 @default.
• W2496955520 hasConceptScore W2496955520C2777601683 @default.
• W2496955520 hasConceptScore W2496955520C28490314 @default.
• W2496955520 hasConceptScore W2496955520C41008148 @default.
• W2496955520 hasConceptScore W2496955520C41895202 @default.
• W2496955520 hasConceptScore W2496955520C50644808 @default.
• W2496955520 hasFunder F4320334627 @default.
• W2496955520 hasIssue "11" @default.
• W2496955520 hasLocation W24969555201 @default.
• W2496955520 hasLocation W24969555202 @default.
• W2496955520 hasOpenAccess W2496955520 @default.
• W2496955520 hasPrimaryLocation W24969555201 @default.
• W2496955520 hasRelatedWork W1595652908 @default.
• W2496955520 hasRelatedWork W1679636228 @default.
• W2496955520 hasRelatedWork W2086999410 @default.
• W2496955520 hasRelatedWork W2154859999 @default.
• W2496955520 hasRelatedWork W2329734087 @default.
• W2496955520 hasRelatedWork W2370984647 @default.
• W2496955520 hasRelatedWork W2373874059 @default.
• W2496955520 hasRelatedWork W2386387936 @default.
• W2496955520 hasRelatedWork W3107474891 @default.
• W2496955520 hasRelatedWork W3175075966 @default.
• W2496955520 hasVolume "24" @default.
• W2496955520 isParatext "false" @default.
• W2496955520 isRetracted "false" @default.

• next