@inproceedings{jain-etal-2024-mafia,
title = "{MAFIA}: Multi-Adapter Fused Inclusive Language Models",
author = "Jain, Prachi and
Sathe, Ashutosh and
Gumma, Varun and
Ahuja, Kabir and
Sitaram, Sunayana",
editor = "Graham, Yvette and
Purver, Matthew",
booktitle = "Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = mar,
year = "2024",
address = "St. Julian{'}s, Malta",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.eacl-long.37",
pages = "627--645",
abstract = "Pretrained Language Models (PLMs) are widely used in NLP for various tasks. Recent studies have identified various biases that such models exhibit and have proposed methods to correct these biases. However, most of the works address a limited set of bias dimensions independently such as gender, race, or religion. Moreover, the methods typically involve finetuning the full model in order to maintain the performance on the downstream task. In this work, we aim to modularly debias a pre-trained language model across multiple dimensions. Previous works extensively explored debiasing PLMs by using limited US-centric counterfactual data augmentation (CDA). We use structured knowledge and a large generative model to build a diverse CDA across multiple bias dimensions in a semi-automated way. We highlight how existing debiasing methods do not consider interactions between multiple societal biases and propose a debiasing model that exploits the synergy amongst various societal biases and enables multi-bias debiasing simultaneously. An extensive evaluation on multiple tasks and languages demonstrates the efficacy of the approach.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jain-etal-2024-mafia">
<titleInfo>
<title>MAFIA: Multi-Adapter Fused Inclusive Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Prachi</namePart>
<namePart type="family">Jain</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ashutosh</namePart>
<namePart type="family">Sathe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Varun</namePart>
<namePart type="family">Gumma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kabir</namePart>
<namePart type="family">Ahuja</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunayana</namePart>
<namePart type="family">Sitaram</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-03</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yvette</namePart>
<namePart type="family">Graham</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Matthew</namePart>
<namePart type="family">Purver</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">St. Julian’s, Malta</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Pretrained Language Models (PLMs) are widely used in NLP for various tasks. Recent studies have identified various biases that such models exhibit and have proposed methods to correct these biases. However, most of the works address a limited set of bias dimensions independently such as gender, race, or religion. Moreover, the methods typically involve finetuning the full model in order to maintain the performance on the downstream task. In this work, we aim to modularly debias a pre-trained language model across multiple dimensions. Previous works extensively explored debiasing PLMs by using limited US-centric counterfactual data augmentation (CDA). We use structured knowledge and a large generative model to build a diverse CDA across multiple bias dimensions in a semi-automated way. We highlight how existing debiasing methods do not consider interactions between multiple societal biases and propose a debiasing model that exploits the synergy amongst various societal biases and enables multi-bias debiasing simultaneously. An extensive evaluation on multiple tasks and languages demonstrates the efficacy of the approach.</abstract>
<identifier type="citekey">jain-etal-2024-mafia</identifier>
<location>
<url>https://aclanthology.org/2024.eacl-long.37</url>
</location>
<part>
<date>2024-03</date>
<extent unit="page">
<start>627</start>
<end>645</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MAFIA: Multi-Adapter Fused Inclusive Language Models
%A Jain, Prachi
%A Sathe, Ashutosh
%A Gumma, Varun
%A Ahuja, Kabir
%A Sitaram, Sunayana
%Y Graham, Yvette
%Y Purver, Matthew
%S Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2024
%8 March
%I Association for Computational Linguistics
%C St. Julian’s, Malta
%F jain-etal-2024-mafia
%X Pretrained Language Models (PLMs) are widely used in NLP for various tasks. Recent studies have identified various biases that such models exhibit and have proposed methods to correct these biases. However, most of the works address a limited set of bias dimensions independently such as gender, race, or religion. Moreover, the methods typically involve finetuning the full model in order to maintain the performance on the downstream task. In this work, we aim to modularly debias a pre-trained language model across multiple dimensions. Previous works extensively explored debiasing PLMs by using limited US-centric counterfactual data augmentation (CDA). We use structured knowledge and a large generative model to build a diverse CDA across multiple bias dimensions in a semi-automated way. We highlight how existing debiasing methods do not consider interactions between multiple societal biases and propose a debiasing model that exploits the synergy amongst various societal biases and enables multi-bias debiasing simultaneously. An extensive evaluation on multiple tasks and languages demonstrates the efficacy of the approach.
%U https://aclanthology.org/2024.eacl-long.37
%P 627-645
Markdown (Informal)
[MAFIA: Multi-Adapter Fused Inclusive Language Models](https://aclanthology.org/2024.eacl-long.37) (Jain et al., EACL 2024)
ACL
- Prachi Jain, Ashutosh Sathe, Varun Gumma, Kabir Ahuja, and Sunayana Sitaram. 2024. MAFIA: Multi-Adapter Fused Inclusive Language Models. In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pages 627–645, St. Julian’s, Malta. Association for Computational Linguistics.