Matches in SemOpenAlex for { <https://semopenalex.org/work/W3008587939> ?p ?o ?g. }
- W3008587939 abstract "In this paper, we present an end-to-end training framework for building state-of-the-art end-to-end speech recognition systems. Our training system utilizes a cluster of Central Processing Units (CPUs) and Graphics Processing Units (GPUs). The entire data reading, large scale data augmentation, neural network parameter updates are all performed “on-the-fly”. We use vocal tract length perturbation [1] and an acoustic simulator [2] for data augmentation. The processed features and labels are sent to the GPU cluster. The Horovod allreduce approach is employed to train neural network parameters. We evaluated the effectiveness of our system on the standard Librispeech corpus [3] and the 10,000-hr anonymized Bixby English dataset. Our end-to-end speech recognition system built using this training infrastructure showed a 2.44 % WER on test-clean of the LibriSpeech test set after applying shallow fusion with a Transformer language model (LM). For the proprietary English Bixby open domain test set, we obtained a WER of 7.92 % using a Bidirectional Full Attention (BFA) end-to-end model after applying shallow fusion with an RNN-LM. When the monotonic chunckwise attention (MoCha) based approach is employed for streaming speech recognition, we obtained a WER of 9.95 % on the same Bixby open domain test set." @default.
- W3008587939 created "2020-03-06" @default.
- W3008587939 creator A5003679010 @default.
- W3008587939 creator A5014613700 @default.
- W3008587939 creator A5021189702 @default.
- W3008587939 creator A5022480907 @default.
- W3008587939 creator A5040849068 @default.
- W3008587939 creator A5041486012 @default.
- W3008587939 creator A5043920881 @default.
- W3008587939 creator A5049274095 @default.
- W3008587939 creator A5052377938 @default.
- W3008587939 creator A5055878624 @default.
- W3008587939 creator A5065995328 @default.
- W3008587939 creator A5072645451 @default.
- W3008587939 creator A5079810318 @default.
- W3008587939 date "2019-12-01" @default.
- W3008587939 modified "2023-10-14" @default.
- W3008587939 title "End-to-End Training of a Large Vocabulary End-to-End Speech Recognition System" @default.
- W3008587939 cites W1494198834 @default.
- W3008587939 cites W1566425957 @default.
- W3008587939 cites W1607274768 @default.
- W3008587939 cites W1616590059 @default.
- W3008587939 cites W1979651826 @default.
- W3008587939 cites W2062164080 @default.
- W3008587939 cites W2063656732 @default.
- W3008587939 cites W2064675550 @default.
- W3008587939 cites W2088489891 @default.
- W3008587939 cites W2103681028 @default.
- W3008587939 cites W2107508685 @default.
- W3008587939 cites W2136439176 @default.
- W3008587939 cites W2160815625 @default.
- W3008587939 cites W2326699523 @default.
- W3008587939 cites W2327501763 @default.
- W3008587939 cites W2396889049 @default.
- W3008587939 cites W2398042854 @default.
- W3008587939 cites W2400570867 @default.
- W3008587939 cites W2515737026 @default.
- W3008587939 cites W2517616541 @default.
- W3008587939 cites W2545177271 @default.
- W3008587939 cites W2589857635 @default.
- W3008587939 cites W2617258110 @default.
- W3008587939 cites W2622203030 @default.
- W3008587939 cites W2702006285 @default.
- W3008587939 cites W2748584437 @default.
- W3008587939 cites W2750499125 @default.
- W3008587939 cites W2785572076 @default.
- W3008587939 cites W2799800213 @default.
- W3008587939 cites W2803023894 @default.
- W3008587939 cites W2936774411 @default.
- W3008587939 cites W2962728618 @default.
- W3008587939 cites W2962824709 @default.
- W3008587939 cites W2963122170 @default.
- W3008587939 cites W2963362078 @default.
- W3008587939 cites W2963747784 @default.
- W3008587939 cites W2963980003 @default.
- W3008587939 cites W2972426539 @default.
- W3008587939 cites W2972816482 @default.
- W3008587939 cites W2997454826 @default.
- W3008587939 cites W3007185811 @default.
- W3008587939 cites W3007528493 @default.
- W3008587939 doi "https://doi.org/10.1109/asru46091.2019.9003976" @default.
- W3008587939 hasPublicationYear "2019" @default.
- W3008587939 type Work @default.
- W3008587939 sameAs 3008587939 @default.
- W3008587939 citedByCount "25" @default.
- W3008587939 countsByYear W30085879392019 @default.
- W3008587939 countsByYear W30085879392020 @default.
- W3008587939 countsByYear W30085879392021 @default.
- W3008587939 countsByYear W30085879392023 @default.
- W3008587939 crossrefType "proceedings-article" @default.
- W3008587939 hasAuthorship W3008587939A5003679010 @default.
- W3008587939 hasAuthorship W3008587939A5014613700 @default.
- W3008587939 hasAuthorship W3008587939A5021189702 @default.
- W3008587939 hasAuthorship W3008587939A5022480907 @default.
- W3008587939 hasAuthorship W3008587939A5040849068 @default.
- W3008587939 hasAuthorship W3008587939A5041486012 @default.
- W3008587939 hasAuthorship W3008587939A5043920881 @default.
- W3008587939 hasAuthorship W3008587939A5049274095 @default.
- W3008587939 hasAuthorship W3008587939A5052377938 @default.
- W3008587939 hasAuthorship W3008587939A5055878624 @default.
- W3008587939 hasAuthorship W3008587939A5065995328 @default.
- W3008587939 hasAuthorship W3008587939A5072645451 @default.
- W3008587939 hasAuthorship W3008587939A5079810318 @default.
- W3008587939 hasBestOaLocation W30085879392 @default.
- W3008587939 hasConcept C127413603 @default.
- W3008587939 hasConcept C136764020 @default.
- W3008587939 hasConcept C138885662 @default.
- W3008587939 hasConcept C154945302 @default.
- W3008587939 hasConcept C15744967 @default.
- W3008587939 hasConcept C17744445 @default.
- W3008587939 hasConcept C199539241 @default.
- W3008587939 hasConcept C2777601683 @default.
- W3008587939 hasConcept C2778935963 @default.
- W3008587939 hasConcept C2780876879 @default.
- W3008587939 hasConcept C28490314 @default.
- W3008587939 hasConcept C2984155413 @default.
- W3008587939 hasConcept C2986709869 @default.
- W3008587939 hasConcept C41008148 @default.
- W3008587939 hasConcept C41895202 @default.
- W3008587939 hasConcept C523214423 @default.