← 研踪 PaperTrailarXiv:2604.18539 · v2
Transition-Matrix Regularization for Next Dialogue Act Prediction in Counselling Conversations
首次提交 2026-04-21 01:33
版本更新 2026-11-02 21:52
雷达观测 2026-11-04 23:41
摘要
This paper studies how empirical dialogue-flow statistics can be incorporated into Next Dialogue Act Prediction (NDAP). A KL regularization term is proposed that aligns predicted act distributions with corpus-derived transition patterns. Evaluated on a 60-class German counselling taxonomy using 5-fold cross-validation, this improves macro-F1 by 9--42% relative depending on encoder and substantially improves dialogue-flow alignment. Cross-dataset validation on HOPE suggests that improvements transfer across languages and counselling domains. In systematic ablations across pretrained encoders and architectures, the findings indicate that transition regularization provides consistent gains and disproportionately benefits weaker baseline models. The results suggest that lightweight discourse-flow priors complement pretrained encoders, especially in fine-grained, data-sparse dialogue tasks.
作者备注：Findings of ACL 2026
期刊信息：Findings of the Association for Computational Linguistics: ACL 2026, pages 25442-25457
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-02 21:52 · 最新观测版本
Transition-Matrix Regularization for Next Dialogue Act Prediction in Counselling Conversations
This paper studies how empirical dialogue-flow statistics can be incorporated into Next Dialogue Act Prediction (NDAP). A KL regularization term is proposed that aligns predicted act distributions with corpus-derived transition patterns. Evaluated on a 60-class German counselling taxonomy using 5-fold cross-validation, this improves macro-F1 by 9--42% relative depending on encoder and substantially improves dialogue-flow alignment. Cross-dataset validation on HOPE suggests that improvements transfer across languages and counselling domains. In systematic ablations across pretrained encoders and architectures, the findings indicate that transition regularization provides consistent gains and disproportionately benefits weaker baseline models. The results suggest that lightweight discourse-flow priors complement pretrained encoders, especially in fine-grained, data-sparse dialogue tasks.
查看此版本解读 →查看此版本原文 ↗