MSA-ASR: Efficient Multilingual Speaker Attribution with frozen ASR Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Nguyen, Thai-Binh, Waibel, Alexander
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915103875530752
author Nguyen, Thai-Binh
Waibel, Alexander
author_facet Nguyen, Thai-Binh
Waibel, Alexander
contents Speaker-attributed automatic speech recognition (SA-ASR) aims to transcribe speech while assigning transcripts to the corresponding speakers accurately. Existing methods often rely on complex modular systems or require extensive fine-tuning of joint modules, limiting their adaptability and general efficiency. This paper introduces a novel approach, leveraging a frozen multilingual ASR model to incorporate speaker attribution into the transcriptions, using only standard monolingual ASR datasets. Our method involves training a speaker module to predict speaker embeddings based on weak labels without requiring additional ASR model modifications. Despite being trained exclusively with non-overlapping monolingual data, our approach effectively extracts speaker attributes across diverse multilingual datasets, including those with overlapping speech. Experimental results demonstrate competitive performance compared to strong baselines, highlighting the model's robustness and potential for practical applications.
format Preprint
id arxiv_https___arxiv_org_abs_2411_18152
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MSA-ASR: Efficient Multilingual Speaker Attribution with frozen ASR Models
Nguyen, Thai-Binh
Waibel, Alexander
Computation and Language
Sound
Audio and Speech Processing
Speaker-attributed automatic speech recognition (SA-ASR) aims to transcribe speech while assigning transcripts to the corresponding speakers accurately. Existing methods often rely on complex modular systems or require extensive fine-tuning of joint modules, limiting their adaptability and general efficiency. This paper introduces a novel approach, leveraging a frozen multilingual ASR model to incorporate speaker attribution into the transcriptions, using only standard monolingual ASR datasets. Our method involves training a speaker module to predict speaker embeddings based on weak labels without requiring additional ASR model modifications. Despite being trained exclusively with non-overlapping monolingual data, our approach effectively extracts speaker attributes across diverse multilingual datasets, including those with overlapping speech. Experimental results demonstrate competitive performance compared to strong baselines, highlighting the model's robustness and potential for practical applications.
title MSA-ASR: Efficient Multilingual Speaker Attribution with frozen ASR Models
topic Computation and Language
Sound
Audio and Speech Processing
url https://arxiv.org/abs/2411.18152