Time vs. Layer: Locating Predictive Cues for Dysarthric Speech Descriptors in wav2vec 2.0
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914502290702336 |
|---|---|
| author | Engert, Natalie Wagner, Dominik Riedhammer, Korbinian Bocklet, Tobias |
| author_facet | Engert, Natalie Wagner, Dominik Riedhammer, Korbinian Bocklet, Tobias |
| contents | Wav2vec 2.0 (W2V2) has shown strong performance in pathological speech analysis by effectively capturing the characteristics of atypical speech. Despite its success, it remains unclear which components of its learned representations are most informative for specific downstream tasks. In this study, we address this question by investigating the regression of dysarthric speech descriptors using annotations from the Speech Accessibility Project dataset. We focus on five descriptors, each addressing a different aspect of speech or voice production: intelligibility, imprecise consonants, inappropriate silences, harsh voice and monoloudness. Speech representations are derived from a W2V2-based feature extractor, and we systematically compare layer-wise and time-wise aggregation strategies using attentive statistics pooling. Our results show that intelligibility is best captured through layer-wise representations, whereas imprecise consonants, harsh voice and monoloudness benefit from time-wise modeling. For inappropriate silences, no clear advantage could be observed for either approach. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_21628 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Time vs. Layer: Locating Predictive Cues for Dysarthric Speech Descriptors in wav2vec 2.0 Engert, Natalie Wagner, Dominik Riedhammer, Korbinian Bocklet, Tobias Sound Wav2vec 2.0 (W2V2) has shown strong performance in pathological speech analysis by effectively capturing the characteristics of atypical speech. Despite its success, it remains unclear which components of its learned representations are most informative for specific downstream tasks. In this study, we address this question by investigating the regression of dysarthric speech descriptors using annotations from the Speech Accessibility Project dataset. We focus on five descriptors, each addressing a different aspect of speech or voice production: intelligibility, imprecise consonants, inappropriate silences, harsh voice and monoloudness. Speech representations are derived from a W2V2-based feature extractor, and we systematically compare layer-wise and time-wise aggregation strategies using attentive statistics pooling. Our results show that intelligibility is best captured through layer-wise representations, whereas imprecise consonants, harsh voice and monoloudness benefit from time-wise modeling. For inappropriate silences, no clear advantage could be observed for either approach. |
| title | Time vs. Layer: Locating Predictive Cues for Dysarthric Speech Descriptors in wav2vec 2.0 |
| topic | Sound |
| url | https://arxiv.org/abs/2604.21628 |