Pretrained ASR Pseudo-labeling for Noisy Police Audio
This work systematically assess the opportunities and limits of pseudo-labeling to adapt foundation ASR models (Whisper and Qwen3-ASR) to noisy BPC domain corpora from Baltimore and Chicago and introduces a new cross-model pseudo-labeling paradigm where one model is finetuned with pseudo-labels from the other.
Kaavya Chaparala, Su-Yuan Huang, Stephen L. Morgan et al.
· 0 citations