StreamHear: Domain-Adapted Pseudo-Labeling for Semi-Supervised Streaming Speech Recognition
The paper introduces StreamHear, a semi-supervised pipeline for adapting pretrained streaming ASR models to domain-shifted target audio using fine-tuning and pseudo-labeling.
StreamHear uses a combination of fine-tuning, pseudo-labeling, and a prior-regularized dynamic-programming realignment step to adapt streaming ASR models to domain-shifted audio.
Before reading this…
Applications
- →Financial calls
- →Prepared read speech
- →Phone-quality dialogue
To understand this paper, make sure you know these concepts first:
- Automatic speech recognitionfind papers →
- Streaming modelsfind papers →
- Fine-tuningfind papers →
- Pseudo-labelsfind papers →
Abstract
More Like ThisStreaming automatic speech recognition (ASR) underperforms on domain-shifted target audio, where labeled in-domain data is costly to prepare while unlabeled audio is abundant. We present StreamHear, a semi-supervised pipeline that adapts a pretrained streaming student by fine-tuning an offline transducer teacher on the labeled training set, generating pseudo-labels on the unlabeled portion, and fine-tuning the student on the mixture. We further introduce a prior-regularized dynamic-programming realignment step that fixes chunk-level word placement using an ASR-hypothesis anchor. Across four datasets spanning financial calls, prepared read speech, and phone-quality dialogue, StreamHear consistently outperforms supervised student fine-tuning and narrows the gap to the offline teacher.