Guessing demographic groups from speech with pseudo-labels or k-means, then applying group-based debiasing, reduces reported gender fairness gaps in multi-label SER on CREMA-D, with accuracy losses the abstract understates.
First, our experiments are based on the CREMA-D dataset, which comprises acted emotional expressions; thus, the generalizability to naturalistic settings remains to be vali- dated
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
dataset 1
citation-polarity summary
fields
eess.AS 1years
2025 1verdicts
CONDITIONAL 1roles
dataset 1polarities
use dataset 1representative citing papers
citing papers explorer
-
Mitigating Subgroup Disparities in Multi-Label Speech Emotion Recognition: A Pseudo-Labeling and Unsupervised Learning Approach
Guessing demographic groups from speech with pseudo-labels or k-means, then applying group-based debiasing, reduces reported gender fairness gaps in multi-label SER on CREMA-D, with accuracy losses the abstract understates.