Revisiting ASR Error Correction with Specialized Models
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917347869065216 |
|---|---|
| author | Gu, Zijin Likhomanenko, Tatiana Bai, He McDermott, Erik Collobert, Ronan Jaitly, Navdeep |
| author_facet | Gu, Zijin Likhomanenko, Tatiana Bai, He McDermott, Erik Collobert, Ronan Jaitly, Navdeep |
| contents | Language models play a central role in automatic speech recognition (ASR), yet most methods rely on text-only models unaware of ASR error patterns. Recently, large language models (LLMs) have been applied to ASR correction, but introduce latency and hallucination concerns. We revisit ASR error correction with compact seq2seq models, trained on ASR errors from real and synthetic audio. To scale training, we construct synthetic corpora via cascaded TTS and ASR, finding that matching the diversity of realistic error distributions is key. We propose correction-first decoding, where the correction model generates candidates rescored using ASR acoustic scores. With 15x fewer parameters than LLMs, our model achieves 1.5/3.3% WER on LibriSpeech test-clean/other, outperforms LLMs, generalizes across ASR architectures (CTC, Seq2seq, Transducer) and diverse domains, and provides precise corrections in the low-error regime where LLMs struggle. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_15216 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Revisiting ASR Error Correction with Specialized Models Gu, Zijin Likhomanenko, Tatiana Bai, He McDermott, Erik Collobert, Ronan Jaitly, Navdeep Machine Learning Computation and Language Sound Audio and Speech Processing Language models play a central role in automatic speech recognition (ASR), yet most methods rely on text-only models unaware of ASR error patterns. Recently, large language models (LLMs) have been applied to ASR correction, but introduce latency and hallucination concerns. We revisit ASR error correction with compact seq2seq models, trained on ASR errors from real and synthetic audio. To scale training, we construct synthetic corpora via cascaded TTS and ASR, finding that matching the diversity of realistic error distributions is key. We propose correction-first decoding, where the correction model generates candidates rescored using ASR acoustic scores. With 15x fewer parameters than LLMs, our model achieves 1.5/3.3% WER on LibriSpeech test-clean/other, outperforms LLMs, generalizes across ASR architectures (CTC, Seq2seq, Transducer) and diverse domains, and provides precise corrections in the low-error regime where LLMs struggle. |
| title | Revisiting ASR Error Correction with Specialized Models |
| topic | Machine Learning Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2405.15216 |