Joint Beamforming and Speaker-Attributed ASR for Real Distant-Microphone Meeting Transcription
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_ | 1866909677934084096 |
|---|---|
| author | Cui, Can Sheikh, Imran Ahamad Sadeghi, Mostafa Vincent, Emmanuel |
| author_facet | Cui, Can Sheikh, Imran Ahamad Sadeghi, Mostafa Vincent, Emmanuel |
| contents | Distant-microphone meeting transcription is a challenging task. State-of-the-art end-to-end speaker-attributed automatic speech recognition (SA-ASR) architectures lack a multichannel noise and reverberation reduction front-end, which limits their performance. In this paper, we introduce a joint beamforming and SA-ASR approach for real meeting transcription. We first describe a data alignment and augmentation method to pretrain a neural beamformer on real meeting data. We then compare fixed, hybrid, and fully neural beamformers as front-ends to the SA-ASR model. Finally, we jointly optimize the fully neural beamformer and the SA-ASR model. Experiments on the real AMI corpus show that, while state-of-the-art multi-frame cross-channel attention based channel fusion fails to improve ASR performance, fine-tuning SA-ASR on the fixed beamformer's output and jointly fine-tuning SA-ASR with the neural beamformer reduce the word error rate by 8% and 9% relative, respectively. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_21849 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Joint Beamforming and Speaker-Attributed ASR for Real Distant-Microphone Meeting Transcription Cui, Can Sheikh, Imran Ahamad Sadeghi, Mostafa Vincent, Emmanuel Computation and Language Distant-microphone meeting transcription is a challenging task. State-of-the-art end-to-end speaker-attributed automatic speech recognition (SA-ASR) architectures lack a multichannel noise and reverberation reduction front-end, which limits their performance. In this paper, we introduce a joint beamforming and SA-ASR approach for real meeting transcription. We first describe a data alignment and augmentation method to pretrain a neural beamformer on real meeting data. We then compare fixed, hybrid, and fully neural beamformers as front-ends to the SA-ASR model. Finally, we jointly optimize the fully neural beamformer and the SA-ASR model. Experiments on the real AMI corpus show that, while state-of-the-art multi-frame cross-channel attention based channel fusion fails to improve ASR performance, fine-tuning SA-ASR on the fixed beamformer's output and jointly fine-tuning SA-ASR with the neural beamformer reduce the word error rate by 8% and 9% relative, respectively. |
| title | Joint Beamforming and Speaker-Attributed ASR for Real Distant-Microphone Meeting Transcription |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2410.21849 |