Improving noisy student training for low-resource languages in End-to-End ASR using CycleGAN and inter-domain losses
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910548006797312 |
|---|---|
| author | Li, Chia-Yu Vu, Ngoc Thang |
| author_facet | Li, Chia-Yu Vu, Ngoc Thang |
| contents | Training a semi-supervised end-to-end speech recognition system using noisy student training has significantly improved performance. However, this approach requires a substantial amount of paired speech-text and unlabeled speech, which is costly for low-resource languages. Therefore, this paper considers a more extreme case of semi-supervised end-to-end automatic speech recognition where there are limited paired speech-text, unlabeled speech (less than five hours), and abundant external text. Firstly, we observe improved performance by training the model using our previous work on semi-supervised learning "CycleGAN and inter-domain losses" solely with external text. Secondly, we enhance "CycleGAN and inter-domain losses" by incorporating automatic hyperparameter tuning, calling it "enhanced CycleGAN inter-domain losses." Thirdly, we integrate it into the noisy student training approach pipeline for low-resource scenarios. Our experimental results, conducted on six non-English languages from Voxforge and Common Voice, show a 20% word error rate reduction compared to the baseline teacher model and a 10% word error rate reduction compared to the baseline best student model, highlighting the significant improvements achieved through our proposed method. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_21061 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Improving noisy student training for low-resource languages in End-to-End ASR using CycleGAN and inter-domain losses Li, Chia-Yu Vu, Ngoc Thang Computation and Language Sound Audio and Speech Processing Training a semi-supervised end-to-end speech recognition system using noisy student training has significantly improved performance. However, this approach requires a substantial amount of paired speech-text and unlabeled speech, which is costly for low-resource languages. Therefore, this paper considers a more extreme case of semi-supervised end-to-end automatic speech recognition where there are limited paired speech-text, unlabeled speech (less than five hours), and abundant external text. Firstly, we observe improved performance by training the model using our previous work on semi-supervised learning "CycleGAN and inter-domain losses" solely with external text. Secondly, we enhance "CycleGAN and inter-domain losses" by incorporating automatic hyperparameter tuning, calling it "enhanced CycleGAN inter-domain losses." Thirdly, we integrate it into the noisy student training approach pipeline for low-resource scenarios. Our experimental results, conducted on six non-English languages from Voxforge and Common Voice, show a 20% word error rate reduction compared to the baseline teacher model and a 10% word error rate reduction compared to the baseline best student model, highlighting the significant improvements achieved through our proposed method. |
| title | Improving noisy student training for low-resource languages in End-to-End ASR using CycleGAN and inter-domain losses |
| topic | Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2407.21061 |