Matches in SemOpenAlex for { <https://semopenalex.org/work/W3025165719> ?p ?o ?g. }
- W3025165719 abstract "Recently Transformer and Convolution neural network (CNN) based models have shown promising results in Automatic Speech Recognition (ASR), outperforming Recurrent neural networks (RNNs). Transformer models are good at capturing content-based global interactions, while CNNs exploit local features effectively. In this work, we achieve the best of both worlds by studying how to combine convolution neural networks and transformers to model both local and global dependencies of an audio sequence in a parameter-efficient way. To this regard, we propose the convolution-augmented transformer for speech recognition, named Conformer. Conformer significantly outperforms the previous Transformer and CNN based models achieving state-of-the-art accuracies. On the widely used LibriSpeech benchmark, our model achieves WER of 2.1%/4.3% without using a language model and 1.9%/3.9% with an external language model on test/testother. We also observe competitive performance of 2.7%/6.3% with a small model of only 10M parameters." @default.
- W3025165719 created "2020-05-21" @default.
- W3025165719 creator A5001205723 @default.
- W3025165719 creator A5002729731 @default.
- W3025165719 creator A5005777963 @default.
- W3025165719 creator A5006617799 @default.
- W3025165719 creator A5010253402 @default.
- W3025165719 creator A5027763497 @default.
- W3025165719 creator A5048771433 @default.
- W3025165719 creator A5067837926 @default.
- W3025165719 creator A5070108948 @default.
- W3025165719 creator A5071773009 @default.
- W3025165719 creator A5074220692 @default.
- W3025165719 date "2020-05-16" @default.
- W3025165719 modified "2023-10-16" @default.
- W3025165719 title "Conformer: Convolution-augmented Transformer for Speech Recognition" @default.
- W3025165719 cites W1494198834 @default.
- W3025165719 cites W1522301498 @default.
- W3025165719 cites W1828163288 @default.
- W3025165719 cites W1964175594 @default.
- W3025165719 cites W1995562189 @default.
- W3025165719 cites W2095705004 @default.
- W3025165719 cites W2112739286 @default.
- W3025165719 cites W2626778328 @default.
- W3025165719 cites W2767286248 @default.
- W3025165719 cites W2798858969 @default.
- W3025165719 cites W2892009249 @default.
- W3025165719 cites W2908336025 @default.
- W3025165719 cites W2928941594 @default.
- W3025165719 cites W2932319281 @default.
- W3025165719 cites W2936774411 @default.
- W3025165719 cites W2937843571 @default.
- W3025165719 cites W2948981900 @default.
- W3025165719 cites W2952180055 @default.
- W3025165719 cites W2962760690 @default.
- W3025165719 cites W2963414781 @default.
- W3025165719 cites W2963420686 @default.
- W3025165719 cites W2963542740 @default.
- W3025165719 cites W2963970792 @default.
- W3025165719 cites W2964110616 @default.
- W3025165719 cites W2972818416 @default.
- W3025165719 cites W2979636403 @default.
- W3025165719 cites W2981413347 @default.
- W3025165719 cites W2981581604 @default.
- W3025165719 cites W2981857663 @default.
- W3025165719 cites W2994771587 @default.
- W3025165719 cites W3015194534 @default.
- W3025165719 cites W3016010032 @default.
- W3025165719 cites W3019527251 @default.
- W3025165719 cites W3021469861 @default.
- W3025165719 doi "https://doi.org/10.48550/arxiv.2005.08100" @default.
- W3025165719 hasPublicationYear "2020" @default.
- W3025165719 type Work @default.
- W3025165719 sameAs 3025165719 @default.
- W3025165719 citedByCount "188" @default.
- W3025165719 countsByYear W30251657192020 @default.
- W3025165719 countsByYear W30251657192021 @default.
- W3025165719 countsByYear W30251657192022 @default.
- W3025165719 countsByYear W30251657192023 @default.
- W3025165719 crossrefType "posted-content" @default.
- W3025165719 hasAuthorship W3025165719A5001205723 @default.
- W3025165719 hasAuthorship W3025165719A5002729731 @default.
- W3025165719 hasAuthorship W3025165719A5005777963 @default.
- W3025165719 hasAuthorship W3025165719A5006617799 @default.
- W3025165719 hasAuthorship W3025165719A5010253402 @default.
- W3025165719 hasAuthorship W3025165719A5027763497 @default.
- W3025165719 hasAuthorship W3025165719A5048771433 @default.
- W3025165719 hasAuthorship W3025165719A5067837926 @default.
- W3025165719 hasAuthorship W3025165719A5070108948 @default.
- W3025165719 hasAuthorship W3025165719A5071773009 @default.
- W3025165719 hasAuthorship W3025165719A5074220692 @default.
- W3025165719 hasBestOaLocation W30251657191 @default.
- W3025165719 hasConcept C119599485 @default.
- W3025165719 hasConcept C127413603 @default.
- W3025165719 hasConcept C137293760 @default.
- W3025165719 hasConcept C147168706 @default.
- W3025165719 hasConcept C153180895 @default.
- W3025165719 hasConcept C154945302 @default.
- W3025165719 hasConcept C165801399 @default.
- W3025165719 hasConcept C28490314 @default.
- W3025165719 hasConcept C41008148 @default.
- W3025165719 hasConcept C50644808 @default.
- W3025165719 hasConcept C66322947 @default.
- W3025165719 hasConcept C81363708 @default.
- W3025165719 hasConceptScore W3025165719C119599485 @default.
- W3025165719 hasConceptScore W3025165719C127413603 @default.
- W3025165719 hasConceptScore W3025165719C137293760 @default.
- W3025165719 hasConceptScore W3025165719C147168706 @default.
- W3025165719 hasConceptScore W3025165719C153180895 @default.
- W3025165719 hasConceptScore W3025165719C154945302 @default.
- W3025165719 hasConceptScore W3025165719C165801399 @default.
- W3025165719 hasConceptScore W3025165719C28490314 @default.
- W3025165719 hasConceptScore W3025165719C41008148 @default.
- W3025165719 hasConceptScore W3025165719C50644808 @default.
- W3025165719 hasConceptScore W3025165719C66322947 @default.
- W3025165719 hasConceptScore W3025165719C81363708 @default.
- W3025165719 hasLocation W30251657191 @default.
- W3025165719 hasOpenAccess W3025165719 @default.
- W3025165719 hasPrimaryLocation W30251657191 @default.
- W3025165719 hasRelatedWork W2175746458 @default.