Exploring Multi-Agent Debate for Zero-Shot Stance Detection: A Novel Approach

Zero-shot stance detection aims to identify the stance expressed in social media text aimed at specific targets without relying on annotated data. However, due to insufficient contextual information and the inherent ambiguity of language, this task faces numerous challenges in low-resource scenarios...

Full description

Saved in:
Bibliographic Details
Main Authors: Junxia Ma, Changjiang Wang, Lu Rong, Bo Wang, Yaoli Xu
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/9/4612
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849312803347759104
author Junxia Ma
Changjiang Wang
Lu Rong
Bo Wang
Yaoli Xu
author_facet Junxia Ma
Changjiang Wang
Lu Rong
Bo Wang
Yaoli Xu
author_sort Junxia Ma
collection DOAJ
description Zero-shot stance detection aims to identify the stance expressed in social media text aimed at specific targets without relying on annotated data. However, due to insufficient contextual information and the inherent ambiguity of language, this task faces numerous challenges in low-resource scenarios. This work proposes a novel zero-shot stance detection method based on multi-agent debate (ZSMD) to address the aforementioned challenges. Specifically, we construct two debater agents representing the supporting and opposing stances. A knowledge enhancement module supplements the original tweet and target with relevant background knowledge, providing richer contextual support for argument generation. Subsequently, the two agents engage in debate over a predetermined number of rounds, employing rebuttal strategies such as factual verification, logical analysis, and sentiment analysis. If no consensus is reached within the specified rounds, a referee agent synthesizes the debate process and original input information to deliver the final stance determination. We evaluate ZSMD on two benchmark datasets, SemEval-2016 Task 6 and P-Stance, and compare it against strong zero-shot baselines such as MB-Cal and COLA. The experimental results show that ZSMD not only achieves higher accuracy than these baselines, but also provides deeper insights into subtle differences in opinion expression, highlighting the potential of structured argumentation in low-resource settings.
format Article
id doaj-art-2acaf1d5edba45769db8c3b1b37e3e65
institution Kabale University
issn 2076-3417
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-2acaf1d5edba45769db8c3b1b37e3e652025-08-20T03:52:57ZengMDPI AGApplied Sciences2076-34172025-04-01159461210.3390/app15094612Exploring Multi-Agent Debate for Zero-Shot Stance Detection: A Novel ApproachJunxia Ma0Changjiang Wang1Lu Rong2Bo Wang3Yaoli Xu4College of Software Engineering, Zhengzhou University of Light Industry, Zhengzhou 450000, ChinaCollege of Software Engineering, Zhengzhou University of Light Industry, Zhengzhou 450000, ChinaSchool of Education, Tianjin University, Tianjin 300072, ChinaCollege of Software Engineering, Zhengzhou University of Light Industry, Zhengzhou 450000, ChinaCollege of Software Engineering, Zhengzhou University of Light Industry, Zhengzhou 450000, ChinaZero-shot stance detection aims to identify the stance expressed in social media text aimed at specific targets without relying on annotated data. However, due to insufficient contextual information and the inherent ambiguity of language, this task faces numerous challenges in low-resource scenarios. This work proposes a novel zero-shot stance detection method based on multi-agent debate (ZSMD) to address the aforementioned challenges. Specifically, we construct two debater agents representing the supporting and opposing stances. A knowledge enhancement module supplements the original tweet and target with relevant background knowledge, providing richer contextual support for argument generation. Subsequently, the two agents engage in debate over a predetermined number of rounds, employing rebuttal strategies such as factual verification, logical analysis, and sentiment analysis. If no consensus is reached within the specified rounds, a referee agent synthesizes the debate process and original input information to deliver the final stance determination. We evaluate ZSMD on two benchmark datasets, SemEval-2016 Task 6 and P-Stance, and compare it against strong zero-shot baselines such as MB-Cal and COLA. The experimental results show that ZSMD not only achieves higher accuracy than these baselines, but also provides deeper insights into subtle differences in opinion expression, highlighting the potential of structured argumentation in low-resource settings.https://www.mdpi.com/2076-3417/15/9/4612zero-shot stance detectionmulti-agent debateknowledge augmentZSMD
spellingShingle Junxia Ma
Changjiang Wang
Lu Rong
Bo Wang
Yaoli Xu
Exploring Multi-Agent Debate for Zero-Shot Stance Detection: A Novel Approach
Applied Sciences
zero-shot stance detection
multi-agent debate
knowledge augment
ZSMD
title Exploring Multi-Agent Debate for Zero-Shot Stance Detection: A Novel Approach
title_full Exploring Multi-Agent Debate for Zero-Shot Stance Detection: A Novel Approach
title_fullStr Exploring Multi-Agent Debate for Zero-Shot Stance Detection: A Novel Approach
title_full_unstemmed Exploring Multi-Agent Debate for Zero-Shot Stance Detection: A Novel Approach
title_short Exploring Multi-Agent Debate for Zero-Shot Stance Detection: A Novel Approach
title_sort exploring multi agent debate for zero shot stance detection a novel approach
topic zero-shot stance detection
multi-agent debate
knowledge augment
ZSMD
url https://www.mdpi.com/2076-3417/15/9/4612
work_keys_str_mv AT junxiama exploringmultiagentdebateforzeroshotstancedetectionanovelapproach
AT changjiangwang exploringmultiagentdebateforzeroshotstancedetectionanovelapproach
AT lurong exploringmultiagentdebateforzeroshotstancedetectionanovelapproach
AT bowang exploringmultiagentdebateforzeroshotstancedetectionanovelapproach
AT yaolixu exploringmultiagentdebateforzeroshotstancedetectionanovelapproach