Exploring SSL Discrete Speech Features for Zipformer-based Contextual ASR
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908401079943168 |
|---|---|
| author | Cui, Mingyu Yang, Yifan Deng, Jiajun Kang, Jiawen Hu, Shujie Wang, Tianzi Li, Zhaoqing Zhang, Shiliang Chen, Xie Liu, Xunying |
| author_facet | Cui, Mingyu Yang, Yifan Deng, Jiajun Kang, Jiawen Hu, Shujie Wang, Tianzi Li, Zhaoqing Zhang, Shiliang Chen, Xie Liu, Xunying |
| contents | Self-supervised learning (SSL) based discrete speech representations are highly compact and domain adaptable. In this paper, SSL discrete speech features extracted from WavLM models are used as additional cross-utterance acoustic context features in Zipformer-Transducer ASR systems. The efficacy of replacing Fbank features with discrete token features for modelling either cross-utterance contexts (from preceding and future segments), or current utterance's internal contexts alone, or both at the same time, are demonstrated thoroughly on the Gigaspeech 1000-hr corpus. The best Zipformer-Transducer system using discrete tokens based cross-utterance context features outperforms the baseline using utterance internal context only with statistically significant word error rate (WER) reductions of 0.32% to 0.41% absolute (2.78% to 3.54% relative) on the dev and test data. The lowest published WER of 11.15% and 11.14% were obtained on the dev and test sets. Our work is open-source and publicly available at https://github.com/open-creator/icefall/tree/master/egs/gigaspeech/Context\_ASR. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_08797 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Exploring SSL Discrete Speech Features for Zipformer-based Contextual ASR Cui, Mingyu Yang, Yifan Deng, Jiajun Kang, Jiawen Hu, Shujie Wang, Tianzi Li, Zhaoqing Zhang, Shiliang Chen, Xie Liu, Xunying Computation and Language Sound Audio and Speech Processing Self-supervised learning (SSL) based discrete speech representations are highly compact and domain adaptable. In this paper, SSL discrete speech features extracted from WavLM models are used as additional cross-utterance acoustic context features in Zipformer-Transducer ASR systems. The efficacy of replacing Fbank features with discrete token features for modelling either cross-utterance contexts (from preceding and future segments), or current utterance's internal contexts alone, or both at the same time, are demonstrated thoroughly on the Gigaspeech 1000-hr corpus. The best Zipformer-Transducer system using discrete tokens based cross-utterance context features outperforms the baseline using utterance internal context only with statistically significant word error rate (WER) reductions of 0.32% to 0.41% absolute (2.78% to 3.54% relative) on the dev and test data. The lowest published WER of 11.15% and 11.14% were obtained on the dev and test sets. Our work is open-source and publicly available at https://github.com/open-creator/icefall/tree/master/egs/gigaspeech/Context\_ASR. |
| title | Exploring SSL Discrete Speech Features for Zipformer-based Contextual ASR |
| topic | Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2409.08797 |