Unifying Diarization, Separation, and ASR with Multi-Speaker Encoder

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shakeel, Muhammad, Sudo, Yui, Peng, Yifan, Lin, Chyi-Jiunn, Watanabe, Shinji
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911678066589696
author Shakeel, Muhammad
Sudo, Yui
Peng, Yifan
Lin, Chyi-Jiunn
Watanabe, Shinji
author_facet Shakeel, Muhammad
Sudo, Yui
Peng, Yifan
Lin, Chyi-Jiunn
Watanabe, Shinji
contents This paper presents a unified multi-speaker encoder (UME), a novel architecture that jointly learns representations for speaker diarization (SD), speech separation (SS), and multi-speaker automatic speech recognition (ASR) tasks using a shared speech foundational encoder. We leverage the hidden representations from multiple layers of UME as a residual weighted-sum encoding (RWSE) to effectively use information from different semantic levels, contributing to bottom-up alignment between tasks. This joint training approach captures the inherent interdependencies among the tasks, enhancing overall performance on overlapping speech data. Our evaluations demonstrate that UME substantially improves over the single-task baselines dedicated to SD, SS, and multi-speaker ASR on LibriMix evaluation sets. Notably, for SD, UME outperforms the previous studies, achieving diarization error rates of 1.37% and 2.29% on Libri2Mix and Libri3Mix evaluation sets, respectively.
format Preprint
id arxiv_https___arxiv_org_abs_2508_20474
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Unifying Diarization, Separation, and ASR with Multi-Speaker Encoder
Shakeel, Muhammad
Sudo, Yui
Peng, Yifan
Lin, Chyi-Jiunn
Watanabe, Shinji
Audio and Speech Processing
Computation and Language
Sound
This paper presents a unified multi-speaker encoder (UME), a novel architecture that jointly learns representations for speaker diarization (SD), speech separation (SS), and multi-speaker automatic speech recognition (ASR) tasks using a shared speech foundational encoder. We leverage the hidden representations from multiple layers of UME as a residual weighted-sum encoding (RWSE) to effectively use information from different semantic levels, contributing to bottom-up alignment between tasks. This joint training approach captures the inherent interdependencies among the tasks, enhancing overall performance on overlapping speech data. Our evaluations demonstrate that UME substantially improves over the single-task baselines dedicated to SD, SS, and multi-speaker ASR on LibriMix evaluation sets. Notably, for SD, UME outperforms the previous studies, achieving diarization error rates of 1.37% and 2.29% on Libri2Mix and Libri3Mix evaluation sets, respectively.
title Unifying Diarization, Separation, and ASR with Multi-Speaker Encoder
topic Audio and Speech Processing
Computation and Language
Sound
url https://arxiv.org/abs/2508.20474