Saved in:
Bibliographic Details
Main Authors: van Dalen, Rogier C., Zhang, Shucong, Parcollet, Titouan, Bhattacharya, Sourav
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.10653
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908405820555264
author van Dalen, Rogier C.
Zhang, Shucong
Parcollet, Titouan
Bhattacharya, Sourav
author_facet van Dalen, Rogier C.
Zhang, Shucong
Parcollet, Titouan
Bhattacharya, Sourav
contents Speech recognisers usually perform optimally only in a specific environment and need to be adapted to work well in another. For adaptation to a new speaker, there is often too little data for fine-tuning to be robust, and that data is usually unlabelled. This paper proposes a combination of approaches to make adaptation to a single minute of data robust. First, instead of estimating the adaptation parameters with cross-entropy on a single error-prone hypothesis or "pseudo-label", this paper proposes a novel loss function, the conditional entropy over complete hypotheses. Using multiple hypotheses makes adaptation more robust to errors in the initial recognition. Second, a "speaker code" characterises a speaker in a vector short enough that it requires little data to estimate. On a far-field noise-augmented version of Common Voice, the proposed scheme yields a 20% relative improvement in word error rate on one minute of adaptation data, increasing on 10 minutes to 29%.
format Preprint
id arxiv_https___arxiv_org_abs_2506_10653
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Robust Unsupervised Adaptation of a Speech Recogniser Using Entropy Minimisation and Speaker Codes
van Dalen, Rogier C.
Zhang, Shucong
Parcollet, Titouan
Bhattacharya, Sourav
Audio and Speech Processing
Computation and Language
Machine Learning
Speech recognisers usually perform optimally only in a specific environment and need to be adapted to work well in another. For adaptation to a new speaker, there is often too little data for fine-tuning to be robust, and that data is usually unlabelled. This paper proposes a combination of approaches to make adaptation to a single minute of data robust. First, instead of estimating the adaptation parameters with cross-entropy on a single error-prone hypothesis or "pseudo-label", this paper proposes a novel loss function, the conditional entropy over complete hypotheses. Using multiple hypotheses makes adaptation more robust to errors in the initial recognition. Second, a "speaker code" characterises a speaker in a vector short enough that it requires little data to estimate. On a far-field noise-augmented version of Common Voice, the proposed scheme yields a 20% relative improvement in word error rate on one minute of adaptation data, increasing on 10 minutes to 29%.
title Robust Unsupervised Adaptation of a Speech Recogniser Using Entropy Minimisation and Speaker Codes
topic Audio and Speech Processing
Computation and Language
Machine Learning
url https://arxiv.org/abs/2506.10653