发表机构
Université Paris-Saclay; CEA; List（巴黎-萨克雷大学; 法国原子能和替代能源委员会; List（此处为CEA下属研究机构，保留英文原名））
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究提出结合话轮转换熵辅助任务的多任务学习方法，在多方对话意图识别任务中提升了预训练模型的性能，优于忽略多方交互动态的现有方法。
AI 中文摘要
我们提出一种用于多方对话意图识别的多任务学习方法，该方法利用建模话轮转换动态的辅助任务。具体而言，我们引入话轮转换熵，这是一种从说话人转换序列计算得到的自监督目标，用于量化交互模式的可预测性。在多个预训练模型上开展的实验表明，纳入该辅助任务可提升意图识别性能，优于那些忽略多方交互动态的现有方法。我们发现，所提出的连续目标可作为单任务目标学习，表明它是携带有用信息的实际信号。
英文摘要
We propose a multi-task learning approach for multi-party dialogue intent recognition that leverages an auxiliary task that models turn-taking dynamics. Specifically, we introduce turn-transition entropy, a self-supervised target computed from the sequence of speaker transitions, which quantifies the predictability of interaction patterns. Experiments on multiple pre-trained models demonstrate that incorporating this auxiliary task improves intent recognition performance, outperforming existing approaches which ignore multi-party interaction dynamics. We find that our proposed continuous target can be learned as a single-task objective, suggesting that it is an actual signal carrying useful information.
CommentsAccepted for publication at EMNLP Industry Track 2026
