Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Something from Nothing: Data Augmentation for Robust Severity Level Estimation of Dysarthric Speech

About

Dysarthric speech quality assessment (DSQA) is critical for clinical diagnostics and inclusive speech technologies. However, subjective evaluation is costly and difficult to scale, and the scarcity of labeled data limits robust objective modeling. To address this, we propose a three-stage framework that leverages unlabeled dysarthric speech and large-scale typical speech datasets to scale training. A teacher model first generates pseudo-labels for unlabeled samples, followed by weakly supervised pretraining using a label-aware contrastive learning strategy that exposes the model to diverse speakers and acoustic conditions. The pretrained model is then fine-tuned for the downstream DSQA task. Experiments on five unseen datasets spanning multiple etiologies and languages demonstrate the robustness of our approach. Our Whisper-based baseline significantly outperforms SOTA DSQA predictors such as SpICE, and the full framework achieves an average SRCC of 0.761 across unseen test datasets.

Jaesung Bae, Xiuwen Zheng, Minje Kim, Chang D. Yoo, Mark Hasegawa-Johnson• 2026

Related benchmarks

TaskDatasetResultRank
Dysarthric speech severity assessmentUASpeech Cross-domain (test)
SRCC0.975
10
Dysarthric speech severity assessmentDysArinVox Cross-domain (test)
SRCC0.631
10
Dysarthric speech severity assessmentSAP In-domain (test)
SRCC0.724
10
Dysarthric speech severity assessmentEasyCall Cross-domain (test)
SRCC0.874
10
Dysarthric speech severity assessmentEWA-DB Cross-domain (test)
SRCC0.711
10
Dysarthric speech severity assessmentNeuroVoz Cross-domain (test)
SRCC0.617
10
Showing 6 of 6 rows

Other info

Follow for update