Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2507.17918 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914077142417408 |
|---|---|
| author | Phan, Nhan Porwal, Anusha Getman, Yaroslav Voskoboinik, Ekaterina Grósz, Tamás Kurimo, Mikko |
| author_facet | Phan, Nhan Porwal, Anusha Getman, Yaroslav Voskoboinik, Ekaterina Grósz, Tamás Kurimo, Mikko |
| contents | We present an efficient end-to-end approach for holistic Automatic Speaking Assessment (ASA) of multi-part second-language tests, developed for the 2025 Speak & Improve Challenge. Our system's main novelty is the ability to process all four spoken responses with a single Whisper-small encoder, combine all information via a lightweight aggregator, and predict the final score. This architecture removes the need for transcription and per-part models, cuts inference time, and makes ASA practical for large-scale Computer-Assisted Language Learning systems.
Our system achieved a Root Mean Squared Error (RMSE) of 0.384, outperforming the text-based baseline (0.44) while using at most 168M parameters (about 70% of Whisper-small). Furthermore, we propose a data sampling strategy, allowing the model to train on only 44.8% of the speakers in the corpus and still reach 0.383 RMSE, demonstrating improved performance on imbalanced classes and strong data efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_17918 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | One Whisper to Grade Them All Phan, Nhan Porwal, Anusha Getman, Yaroslav Voskoboinik, Ekaterina Grósz, Tamás Kurimo, Mikko Computation and Language Audio and Speech Processing We present an efficient end-to-end approach for holistic Automatic Speaking Assessment (ASA) of multi-part second-language tests, developed for the 2025 Speak & Improve Challenge. Our system's main novelty is the ability to process all four spoken responses with a single Whisper-small encoder, combine all information via a lightweight aggregator, and predict the final score. This architecture removes the need for transcription and per-part models, cuts inference time, and makes ASA practical for large-scale Computer-Assisted Language Learning systems. Our system achieved a Root Mean Squared Error (RMSE) of 0.384, outperforming the text-based baseline (0.44) while using at most 168M parameters (about 70% of Whisper-small). Furthermore, we propose a data sampling strategy, allowing the model to train on only 44.8% of the speakers in the corpus and still reach 0.383 RMSE, demonstrating improved performance on imbalanced classes and strong data efficiency. |
| title | One Whisper to Grade Them All |
| topic | Computation and Language Audio and Speech Processing |
| url | https://arxiv.org/abs/2507.17918 |