Distillation and Pruning for Scalable Self-Supervised Representation-Based Speech Quality Assessment
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909483886706688 |
|---|---|
| author | Stahl, Benjamin Gamper, Hannes |
| author_facet | Stahl, Benjamin Gamper, Hannes |
| contents | In this paper, we investigate distillation and pruning methods to reduce model size for non-intrusive speech quality assessment based on self-supervised representations. Our experiments build on XLS-R-SQA, a speech quality assessment model using wav2vec 2.0 XLS-R embeddings. We retrain this model on a large compilation of mean opinion score datasets, encompassing over 100,000 labeled clips. For distillation, using this model as a teacher, we generate pseudo-labels on unlabeled degraded speech signals and train student models of varying sizes. For pruning, we use a data-driven strategy. While data-driven pruning performs better at larger model sizes, distillation on unlabeled data is more effective for smaller model sizes. Distillation can halve the gap between the baseline's correlation with ground-truth MOS labels and that of the XLS-R-based teacher model, while reducing model size by two orders of magnitude compared to the teacher model. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_05356 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Distillation and Pruning for Scalable Self-Supervised Representation-Based Speech Quality Assessment Stahl, Benjamin Gamper, Hannes Audio and Speech Processing Sound In this paper, we investigate distillation and pruning methods to reduce model size for non-intrusive speech quality assessment based on self-supervised representations. Our experiments build on XLS-R-SQA, a speech quality assessment model using wav2vec 2.0 XLS-R embeddings. We retrain this model on a large compilation of mean opinion score datasets, encompassing over 100,000 labeled clips. For distillation, using this model as a teacher, we generate pseudo-labels on unlabeled degraded speech signals and train student models of varying sizes. For pruning, we use a data-driven strategy. While data-driven pruning performs better at larger model sizes, distillation on unlabeled data is more effective for smaller model sizes. Distillation can halve the gap between the baseline's correlation with ground-truth MOS labels and that of the XLS-R-based teacher model, while reducing model size by two orders of magnitude compared to the teacher model. |
| title | Distillation and Pruning for Scalable Self-Supervised Representation-Based Speech Quality Assessment |
| topic | Audio and Speech Processing Sound |
| url | https://arxiv.org/abs/2502.05356 |