Zero-Shot Conversational Stance Detection: Dataset and Approaches

2025-06-21

AI Summary

To address the poor generalization of existing dialogue stance detection models to unseen targets, this paper introduces zero-shot cross-target stance detection—a novel task. To support this research, we construct ZS-CSD, the first large-scale, high-quality zero-shot dialogue stance detection dataset, covering diverse targets and authentic conversational scenarios. We further propose the Speaker-Interactive and Target-Perceptive Prototype Contrastive Learning model (SITPCL), which integrates prototype networks, contrastive learning, speaker modeling, and target-aware representation learning. Experiments demonstrate that SITPCL achieves a macro-F1 score of 43.81% under zero-shot settings—significantly outperforming all baselines—and establishes the first benchmark for this task. This work formally defines and advances zero-shot transfer learning in dialogue stance detection, marking the first systematic study of its kind.

Technology Category

• Natural Language Processing: Conversational AI/Dialog Systems
• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Stance detection, which aims to identify public opinion towards specific targets using social media data, is an important yet challenging task. With the increasing number of online debates among social media users, conversational stance detection has become a crucial research area. However, existing conversational stance detection datasets are restricted to a limited set of specific targets, which constrains the effectiveness of stance detection models when encountering a large number of unseen targets in real-world applications. To bridge this gap, we manually curate a large-scale, high-quality zero-shot conversational stance detection dataset, named ZS-CSD, comprising 280 targets across two distinct target types. Leveraging the ZS-CSD dataset, we propose SITPCL, a speaker interaction and target-aware prototypical contrastive learning model, and establish the benchmark performance in the zero-shot setting. Experimental results demonstrate that our proposed SITPCL model achieves state-of-the-art performance in zero-shot conversational stance detection. Notably, the SITPCL model attains only an F1-macro score of 43.81%, highlighting the persistent challenges in zero-shot conversational stance detection.

Problem

• Detecting stance in conversations with unseen targets
• Overcoming limited target diversity in existing datasets
• Improving zero-shot performance for conversational stance detection

Innovation

• Large-scale zero-shot conversational stance detection dataset
• Speaker interaction and target-aware prototypical learning
• State-of-the-art performance in zero-shot setting