MT²-CSD and LLM-CRAN: A new dataset and an LLM-based multi-semantic knowledge fusion model for conversational stance...

In contemporary social media, automatic stance detection is pivotal for opinion mining, as it synthesizes and examines user perspectives on contentious topics to uncover prevailing trends and sentiments. Traditional stance detection research often targets individual instances, thereby limiting its capacity to model multi-party discussions typical in real social media scenarios. This shortcoming largely stems from the scarcity of datasets that authentically capture the dynamics of social media interactions, hindering advancements in conversational stance detection. In this paper, we introduce MT²-CSD, a comprehensive dataset for multi-target, multi-turn conversational stance detection. To the best of our knowledge, MT²-CSD is the largest dataset available for this purpose, comprising 24,457 annotated instances and exhibiting substantial conversational depth, thereby presenting new challenges for stance detection. To address these challenges, we propose the Large Language Model enhanced Conversational Relational Attention Network (LLM-CRAN), which uses LLM-derived relational knowledge to improve conversational understanding. We conduct extensive experiments to evaluate the efficacy of LLM-CRAN on the MT²-CSD dataset. The experimental results indicate that LLM-CRAN significantly outperforms strong baseline models in the task of conversational stance detection. The dataset and code for MT²-CSD are available at https://github.com/nfq729/MT-CSD.