Self-supervised speech encoders show speaker-group bias from the first latent layers, with SID bias lowest where overall error is lowest while ASR bias is highest where overall error is lowest; the ASR pattern persists after fine-tuning.
Where Do Self-Supervised Speech Models Become Unfair?
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
Speech encoder models are known to model members of some speaker groups (SGs) better than others. However, there has been little work in establishing why this occurs on a technological level. To our knowledge, we present the first layerwise fairness analysis of pretrained self-supervised speech encoder models (S3Ms), probing each embedding layer for speaker identification (SID) automatic speech recognition (ASR). We find S3Ms produce embeddings biased against certain SGs for both tasks, starting at the very first latent layers. Furthermore, we find opposite patterns of layerwise bias for SID vs ASR for all models in our study: SID bias is minimized in layers that minimize overall SID error; on the other hand, ASR bias is maximized in layers that minimize overall ASR error. The inverse bias/error relationship for ASR is unaffected when probing S3Ms that are finetuned for ASR, suggesting SG-level bias is established during pretraining and is difficult to remove.
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Where Do Self-Supervised Speech Models Become Unfair?
Self-supervised speech encoders show speaker-group bias from the first latent layers, with SID bias lowest where overall error is lowest while ASR bias is highest where overall error is lowest; the ASR pattern persists after fine-tuning.