Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2011.02160 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910680820482048 |
|---|---|
| author | Du, Chenpeng Li, Hao Lu, Yizhou Wang, Lan Qian, Yanmin |
| author_facet | Du, Chenpeng Li, Hao Lu, Yizhou Wang, Lan Qian, Yanmin |
| contents | Training a code-switching end-to-end automatic speech recognition (ASR) model normally requires a large amount of data, while code-switching data is often limited. In this paper, three novel approaches are proposed for code-switching data augmentation. Specifically, they are audio splicing with the existing code-switching data, and TTS with new code-switching texts generated by word translation or word insertion. Our experiments on 200 hours Mandarin-English code-switching dataset show that all the three proposed approaches yield significant improvements on code-switching ASR individually. Moreover, all the proposed approaches can be combined with recent popular SpecAugment, and an addition gain can be obtained. WER is significantly reduced by relative 24.0% compared to the system without any data augmentation, and still relative 13.0% gain compared to the system with only SpecAugment |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2011_02160 |
| institution | arXiv |
| publishDate | 2020 |
| record_format | arxiv |
| spellingShingle | Data Augmentation for End-to-end Code-switching Speech Recognition Du, Chenpeng Li, Hao Lu, Yizhou Wang, Lan Qian, Yanmin Computation and Language Audio and Speech Processing Training a code-switching end-to-end automatic speech recognition (ASR) model normally requires a large amount of data, while code-switching data is often limited. In this paper, three novel approaches are proposed for code-switching data augmentation. Specifically, they are audio splicing with the existing code-switching data, and TTS with new code-switching texts generated by word translation or word insertion. Our experiments on 200 hours Mandarin-English code-switching dataset show that all the three proposed approaches yield significant improvements on code-switching ASR individually. Moreover, all the proposed approaches can be combined with recent popular SpecAugment, and an addition gain can be obtained. WER is significantly reduced by relative 24.0% compared to the system without any data augmentation, and still relative 13.0% gain compared to the system with only SpecAugment |
| title | Data Augmentation for End-to-end Code-switching Speech Recognition |
| topic | Computation and Language Audio and Speech Processing |
| url | https://arxiv.org/abs/2011.02160 |