Mispronunciation Detection and Diagnosis Without Model Training: A Retrieval-Based Approach
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912727916609536 |
|---|---|
| author | Tu, Huu Tuong Khanh, Ha Viet Dat, Tran Tien Huan, Vu Van Luong, Thien Cuong, Nguyen Tien Trang, Nguyen Thi Thu |
| author_facet | Tu, Huu Tuong Khanh, Ha Viet Dat, Tran Tien Huan, Vu Van Luong, Thien Cuong, Nguyen Tien Trang, Nguyen Thi Thu |
| contents | Mispronunciation Detection and Diagnosis (MDD) is crucial for language learning and speech therapy. Unlike conventional methods that require scoring models or training phoneme-level models, we propose a novel training-free framework that leverages retrieval techniques with a pretrained Automatic Speech Recognition model. Our method avoids phoneme-specific modeling or additional task-specific training, while still achieving accurate detection and diagnosis of pronunciation errors. Experiments on the L2-ARCTIC dataset show that our method achieves a superior F1 score of 69.60% while avoiding the complexity of model training. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_20107 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Mispronunciation Detection and Diagnosis Without Model Training: A Retrieval-Based Approach Tu, Huu Tuong Khanh, Ha Viet Dat, Tran Tien Huan, Vu Van Luong, Thien Cuong, Nguyen Tien Trang, Nguyen Thi Thu Computation and Language Sound Audio and Speech Processing Mispronunciation Detection and Diagnosis (MDD) is crucial for language learning and speech therapy. Unlike conventional methods that require scoring models or training phoneme-level models, we propose a novel training-free framework that leverages retrieval techniques with a pretrained Automatic Speech Recognition model. Our method avoids phoneme-specific modeling or additional task-specific training, while still achieving accurate detection and diagnosis of pronunciation errors. Experiments on the L2-ARCTIC dataset show that our method achieves a superior F1 score of 69.60% while avoiding the complexity of model training. |
| title | Mispronunciation Detection and Diagnosis Without Model Training: A Retrieval-Based Approach |
| topic | Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2511.20107 |