华侨大学学报（自然科学版）2026，Vol.47Issue(2)：183-192,10.DOI:10.11830/ISSN.1000-5013.202510024
多视角融合的无监督对话主题分割模型
Unsupervised Dialogue Topic Segmentation via Multi-View Fusion
摘要
Abstract
An unsupervised dialogue topic segmentation model based on multi-view fusion is proposed in this paper,which jointly models semantic similarity,logical coherence and summary consistency.The proposed framework adaptively integrates information from multiple perspectives through a hybrid mechanism combining static weighting with dynamic gating.Furthermore,a unified optimization objective is established by combi-ning neighboring utterance matching loss,summary consistency loss,and semantic-correlation modeling loss.Experimental results on three representative datasets show that the proposed model consistently achieves supe-rior performance,effectively improving robustness and global semantic coherence.关键词
对话主题分割/对话摘要/主题建模/无监督/相邻话语匹配Key words
dialogue topic segmentation/dialogue summarization/topic modeling/unsupervised/neighboring utterance matching分类
信息技术与安全科学引用本文复制引用
王吉豪,喻小光,陈霞..多视角融合的无监督对话主题分割模型[J].华侨大学学报（自然科学版）,2026,47(2):183-192,10.基金项目
国家自然科学基金面上资助项目(62476103) （62476103）
