Not All or None: Dynamic Construction of Target-aware Memory Graph for Conversational Stance Detection
Computation and LanguageArtificial Intelligence
Summary
The authors present a new method called Memory Graph TamGraph for detecting a person's opinion in conversations on social media. Their approach focuses on important past statements related to the topic, rather than all previous messages, to better understand stance without adding confusion. They use a smart backtracking system to pick out useful information and build a graph that shows how opinions connect. Tests on English and Chinese data sets show that this method improves performance of large language models in this task.
stance detectionconversational stance detectionmemory graphbacktrackingentropy-guidedtarget-awaresocial media analysislarge language modelsutterancegraph modeling
Authors
Yifan Xiang, Bin Liang, Yuqi Huang, Ruifeng Xu, Kam-Fai Wong
Abstract
Stance detection is crucial for understanding the underlying attitude of an expression towards a target. Conversational stance detection is a more challenging stance detection task in real-world social media scenarios, as it involves detecting the user's stance by leveraging the target-related historical statements across conversational sessions. In this paper, we propose target-aware Memory Graph TamGraph, a novel method that dynamically leverages target-related statements for conversational stance detection. Instead of considering all preceding historical conversations or using no prior conversation information for stance detection, our TamGraph employs a stepwise, entropy-guided backtracking mechanism to selectively activate memory from historical conversations and dynamically constructs a target-aware graph to model the stance relations among utterances. This allows the exploitation of target-related information from the conversation history for stance detection while preventing the introduction of noise. Experimental results on both English and Chinese benchmarks demonstrate that our TamGraph substantially improves LLM performance on conversational stance detection.