@inproceedings{jahan-etal-2026-towards,
title = "Towards Fair Speech Recognition: Mitigating Demographic Bias in End-to-End {ASR} Systems",
author = "Jahan, Maliha and
Thebaud, Thomas and
Fagyal, Zsuzsanna and
Villalba, Jesus and
Hasegawa-Johnson, Mark and
Moro Velazquez, Laureano and
Dehak, Najim",
editor = "Piperidis, Stelios and
Bel, N{\'u}ria and
van den Heuvel, Henk and
Ide, Nancy and
Krek, Simon and
Toral, Antonio",
booktitle = "Proceedings of the Fifteenth Language Resources and Evaluation Conference",
month = may,
year = "2026",
address = "Palma de Mallorca, Spain",
publisher = "ELRA Language Resource Association",
url = "https://aclanthology.org/2026.lrec-1.325/",
doi = "10.63317/5e5f93jkhzt5",
pages = "4115--4125",
abstract = "Demographic bias in the performance of speech and language technology has been an active area of recent research. A lot of studies have shown the existence of demographic biases in Automatic Speech Recognition (ASR) systems. In this work, we propose a novel model-agnostic and demographic label-agnostic approach, called DARe, to mitigate any existing bias in an ASR system towards certain speaker groups. We built a debiasing module that goes between the feature extractor of an ASR and the rest of that ASR. The module includes content-group disentanglers to separate content and group, a demographic classifier, and adversarial reweighting. To eliminate the need for demographic labels, we generated pseudo-group labels by extracting speaker embeddings and clustering them. We worked with three ASR systems{--}Wav2Vec2 base, SEW tiny, and Whisper small. We used the FAI dataset, which contains naturalistic conversations with speakers who self-identify their demographic attributes. We used Word Error Rate (WER) as a metric of ASR performance and a Poisson regression-based approach to evaluate the racial fairness of the models. We compared the racial bias of the models before and after applying our proposed approach and observed a significant improvement in fairness."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jahan-etal-2026-towards">
<titleInfo>
<title>Towards Fair Speech Recognition: Mitigating Demographic Bias in End-to-End ASR Systems</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maliha</namePart>
<namePart type="family">Jahan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Thomas</namePart>
<namePart type="family">Thebaud</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zsuzsanna</namePart>
<namePart type="family">Fagyal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jesus</namePart>
<namePart type="family">Villalba</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mark</namePart>
<namePart type="family">Hasegawa-Johnson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Laureano</namePart>
<namePart type="family">Moro Velazquez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Najim</namePart>
<namePart type="family">Dehak</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Fifteenth Language Resources and Evaluation Conference</title>
</titleInfo>
<name type="personal">
<namePart type="given">Stelios</namePart>
<namePart type="family">Piperidis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Núria</namePart>
<namePart type="family">Bel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Henk</namePart>
<namePart type="family">van den Heuvel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nancy</namePart>
<namePart type="family">Ide</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Krek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Antonio</namePart>
<namePart type="family">Toral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resource Association</publisher>
<place>
<placeTerm type="text">Palma de Mallorca, Spain</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Demographic bias in the performance of speech and language technology has been an active area of recent research. A lot of studies have shown the existence of demographic biases in Automatic Speech Recognition (ASR) systems. In this work, we propose a novel model-agnostic and demographic label-agnostic approach, called DARe, to mitigate any existing bias in an ASR system towards certain speaker groups. We built a debiasing module that goes between the feature extractor of an ASR and the rest of that ASR. The module includes content-group disentanglers to separate content and group, a demographic classifier, and adversarial reweighting. To eliminate the need for demographic labels, we generated pseudo-group labels by extracting speaker embeddings and clustering them. We worked with three ASR systems–Wav2Vec2 base, SEW tiny, and Whisper small. We used the FAI dataset, which contains naturalistic conversations with speakers who self-identify their demographic attributes. We used Word Error Rate (WER) as a metric of ASR performance and a Poisson regression-based approach to evaluate the racial fairness of the models. We compared the racial bias of the models before and after applying our proposed approach and observed a significant improvement in fairness.</abstract>
<identifier type="citekey">jahan-etal-2026-towards</identifier>
<identifier type="doi">10.63317/5e5f93jkhzt5</identifier>
<location>
<url>https://aclanthology.org/2026.lrec-1.325/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>4115</start>
<end>4125</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Towards Fair Speech Recognition: Mitigating Demographic Bias in End-to-End ASR Systems
%A Jahan, Maliha
%A Thebaud, Thomas
%A Fagyal, Zsuzsanna
%A Villalba, Jesus
%A Hasegawa-Johnson, Mark
%A Moro Velazquez, Laureano
%A Dehak, Najim
%Y Piperidis, Stelios
%Y Bel, Núria
%Y van den Heuvel, Henk
%Y Ide, Nancy
%Y Krek, Simon
%Y Toral, Antonio
%S Proceedings of the Fifteenth Language Resources and Evaluation Conference
%D 2026
%8 May
%I ELRA Language Resource Association
%C Palma de Mallorca, Spain
%F jahan-etal-2026-towards
%X Demographic bias in the performance of speech and language technology has been an active area of recent research. A lot of studies have shown the existence of demographic biases in Automatic Speech Recognition (ASR) systems. In this work, we propose a novel model-agnostic and demographic label-agnostic approach, called DARe, to mitigate any existing bias in an ASR system towards certain speaker groups. We built a debiasing module that goes between the feature extractor of an ASR and the rest of that ASR. The module includes content-group disentanglers to separate content and group, a demographic classifier, and adversarial reweighting. To eliminate the need for demographic labels, we generated pseudo-group labels by extracting speaker embeddings and clustering them. We worked with three ASR systems–Wav2Vec2 base, SEW tiny, and Whisper small. We used the FAI dataset, which contains naturalistic conversations with speakers who self-identify their demographic attributes. We used Word Error Rate (WER) as a metric of ASR performance and a Poisson regression-based approach to evaluate the racial fairness of the models. We compared the racial bias of the models before and after applying our proposed approach and observed a significant improvement in fairness.
%R 10.63317/5e5f93jkhzt5
%U https://aclanthology.org/2026.lrec-1.325/
%U https://doi.org/10.63317/5e5f93jkhzt5
%P 4115-4125
Markdown (Informal)
[Towards Fair Speech Recognition: Mitigating Demographic Bias in End-to-End ASR Systems](https://aclanthology.org/2026.lrec-1.325/) (Jahan et al., LREC 2026)
ACL