Differentiable K-means for Fully-optimized Discrete Token-based ASR

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Onda, Kentaro, Kashiwagi, Yosuke, Tsunoo, Emiru, Futami, Hayato, Watanabe, Shinji
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915297871527936
author Onda, Kentaro
Kashiwagi, Yosuke
Tsunoo, Emiru
Futami, Hayato
Watanabe, Shinji
author_facet Onda, Kentaro
Kashiwagi, Yosuke
Tsunoo, Emiru
Futami, Hayato
Watanabe, Shinji
contents Recent studies have highlighted the potential of discrete tokens derived from self-supervised learning (SSL) models for various speech-related tasks. These tokens serve not only as substitutes for text in language modeling but also as intermediate representations for tasks such as automatic speech recognition (ASR). However, discrete tokens are typically obtained via k-means clustering of SSL features independently of downstream tasks, making them suboptimal for specific applications. This paper proposes the use of differentiable k-means, enabling the joint optimization of tokenization and downstream tasks. This approach enables the fine-tuning of the SSL parameters and learning weights for outputs from multiple SSL layers. Experiments were conducted with ASR as a downstream task. ASR accuracy successfully improved owing to the optimized tokens. The acquired tokens also exhibited greater purity of phonetic information, which were found to be useful even in speech resynthesis.
format Preprint
id arxiv_https___arxiv_org_abs_2505_16207
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Differentiable K-means for Fully-optimized Discrete Token-based ASR
Onda, Kentaro
Kashiwagi, Yosuke
Tsunoo, Emiru
Futami, Hayato
Watanabe, Shinji
Sound
Audio and Speech Processing
Recent studies have highlighted the potential of discrete tokens derived from self-supervised learning (SSL) models for various speech-related tasks. These tokens serve not only as substitutes for text in language modeling but also as intermediate representations for tasks such as automatic speech recognition (ASR). However, discrete tokens are typically obtained via k-means clustering of SSL features independently of downstream tasks, making them suboptimal for specific applications. This paper proposes the use of differentiable k-means, enabling the joint optimization of tokenization and downstream tasks. This approach enables the fine-tuning of the SSL parameters and learning weights for outputs from multiple SSL layers. Experiments were conducted with ASR as a downstream task. ASR accuracy successfully improved owing to the optimized tokens. The acquired tokens also exhibited greater purity of phonetic information, which were found to be useful even in speech resynthesis.
title Differentiable K-means for Fully-optimized Discrete Token-based ASR
topic Sound
Audio and Speech Processing
url https://arxiv.org/abs/2505.16207