@inproceedings{cheng-etal-2026-mard,
title = "{MARD}: A Multi-Agent Reinforced Debiasing Framework for Large Language Models",
author = "Cheng, Xiaoqing and
Liu, Tao and
Lan, Tian and
Song, Jinwang and
Chen, Chen and
Zan, Hongying and
Zhang, Wen and
Chen, Huajun",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2630/",
pages = "59812--59823",
ISBN = "979-8-89176-459-0",
abstract = "Mitigating social bias in large language models (LLMs) has become an increasingly important research objective. However, existing methods often rely on prompt-level control or optimize bias-aware guidance and final prediction within a single model, which may lead to coupled optimization and task interference. To address these limitations, we propose MARD, a Multi-Agent Reinforced Debiasing framework that decomposes debiasing into two role-specialized agents. The Bias-Guard Agent identifies stereotype cues and generates bias-control guidance, while the Evidence-Decision Agent produces evidence-grounded predictions without relying on identity-based assumptions. MARD further introduces retrieved group-related references as negative warning signals to expose potential stereotype cues, and adopts an alternating GRPO optimization strategy to improve role specialization. Experiments on five datasets across four task types demonstrate that MARD achieves strong overall debiasing performance. Further analyses highlight the effectiveness of role decoupling, alternating optimization, format control, and retrieved warning references."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="cheng-etal-2026-mard">
<titleInfo>
<title>MARD: A Multi-Agent Reinforced Debiasing Framework for Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xiaoqing</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tao</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tian</namePart>
<namePart type="family">Lan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jinwang</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chen</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongying</namePart>
<namePart type="family">Zan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wen</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Huajun</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Mitigating social bias in large language models (LLMs) has become an increasingly important research objective. However, existing methods often rely on prompt-level control or optimize bias-aware guidance and final prediction within a single model, which may lead to coupled optimization and task interference. To address these limitations, we propose MARD, a Multi-Agent Reinforced Debiasing framework that decomposes debiasing into two role-specialized agents. The Bias-Guard Agent identifies stereotype cues and generates bias-control guidance, while the Evidence-Decision Agent produces evidence-grounded predictions without relying on identity-based assumptions. MARD further introduces retrieved group-related references as negative warning signals to expose potential stereotype cues, and adopts an alternating GRPO optimization strategy to improve role specialization. Experiments on five datasets across four task types demonstrate that MARD achieves strong overall debiasing performance. Further analyses highlight the effectiveness of role decoupling, alternating optimization, format control, and retrieved warning references.</abstract>
<identifier type="citekey">cheng-etal-2026-mard</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2630/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>59812</start>
<end>59823</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MARD: A Multi-Agent Reinforced Debiasing Framework for Large Language Models
%A Cheng, Xiaoqing
%A Liu, Tao
%A Lan, Tian
%A Song, Jinwang
%A Chen, Chen
%A Zan, Hongying
%A Zhang, Wen
%A Chen, Huajun
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F cheng-etal-2026-mard
%X Mitigating social bias in large language models (LLMs) has become an increasingly important research objective. However, existing methods often rely on prompt-level control or optimize bias-aware guidance and final prediction within a single model, which may lead to coupled optimization and task interference. To address these limitations, we propose MARD, a Multi-Agent Reinforced Debiasing framework that decomposes debiasing into two role-specialized agents. The Bias-Guard Agent identifies stereotype cues and generates bias-control guidance, while the Evidence-Decision Agent produces evidence-grounded predictions without relying on identity-based assumptions. MARD further introduces retrieved group-related references as negative warning signals to expose potential stereotype cues, and adopts an alternating GRPO optimization strategy to improve role specialization. Experiments on five datasets across four task types demonstrate that MARD achieves strong overall debiasing performance. Further analyses highlight the effectiveness of role decoupling, alternating optimization, format control, and retrieved warning references.
%U https://aclanthology.org/2026.emnlp-main.2630/
%P 59812-59823
Markdown (Informal)
[MARD: A Multi-Agent Reinforced Debiasing Framework for Large Language Models](https://aclanthology.org/2026.emnlp-main.2630/) (Cheng et al., EMNLP 2026)
ACL
- Xiaoqing Cheng, Tao Liu, Tian Lan, Jinwang Song, Chen Chen, Hongying Zan, Wen Zhang, and Huajun Chen. 2026. MARD: A Multi-Agent Reinforced Debiasing Framework for Large Language Models. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 59812–59823, Budapest, Hungary. Association for Computational Linguistics.