Dual Debiasing: Remove Stereotypes and Keep Factual Gender for Fair Language Modeling and Translation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Limisiewicz, Tomasz, Mareček, David, Musil, Tomáš
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909470242635776
author Limisiewicz, Tomasz
Mareček, David
Musil, Tomáš
author_facet Limisiewicz, Tomasz
Mareček, David
Musil, Tomáš
contents Mitigation of biases, such as language models' reliance on gender stereotypes, is a crucial endeavor required for the creation of reliable and useful language technology. The crucial aspect of debiasing is to ensure that the models preserve their versatile capabilities, including their ability to solve language tasks and equitably represent various genders. To address this issue, we introduce a streamlined Dual Dabiasing Algorithm through Model Adaptation (2DAMA). Novel Dual Debiasing enables robust reduction of stereotypical bias while preserving desired factual gender information encoded by language models. We show that 2DAMA effectively reduces gender bias in English and is one of the first approaches facilitating the mitigation of stereotypical tendencies in translation. The proposed method's key advantage is the preservation of factual gender cues, which are useful in a wide range of natural language processing tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2501_10150
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dual Debiasing: Remove Stereotypes and Keep Factual Gender for Fair Language Modeling and Translation
Limisiewicz, Tomasz
Mareček, David
Musil, Tomáš
Computation and Language
Artificial Intelligence
Mitigation of biases, such as language models' reliance on gender stereotypes, is a crucial endeavor required for the creation of reliable and useful language technology. The crucial aspect of debiasing is to ensure that the models preserve their versatile capabilities, including their ability to solve language tasks and equitably represent various genders. To address this issue, we introduce a streamlined Dual Dabiasing Algorithm through Model Adaptation (2DAMA). Novel Dual Debiasing enables robust reduction of stereotypical bias while preserving desired factual gender information encoded by language models. We show that 2DAMA effectively reduces gender bias in English and is one of the first approaches facilitating the mitigation of stereotypical tendencies in translation. The proposed method's key advantage is the preservation of factual gender cues, which are useful in a wide range of natural language processing tasks.
title Dual Debiasing: Remove Stereotypes and Keep Factual Gender for Fair Language Modeling and Translation
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2501.10150