Regularized federated learning (parameter, embedding, and KL-loss based) consistently outperforms FedAvg for dysarthric and elderly speech recognition by up to 0.55% absolute WER, and per-batch communication approaches centralized training accuracy.
Speech-Transformer: A No-Recurrence Sequence-to-Sequence Model for Speech Recognition,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
eess.AS 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Regularized Federated Learning for Privacy-Preserving Dysarthric and Elderly Speech Recognition
Regularized federated learning (parameter, embedding, and KL-loss based) consistently outperforms FedAvg for dysarthric and elderly speech recognition by up to 0.55% absolute WER, and per-batch communication approaches centralized training accuracy.