Transition-Matrix Regularization for Next Dialogue Act Prediction in Counselling Conversations
cs.CL, cs.AI
Submitted: 2026-04-20
Updated: 2026-11-02
Comments: Findings of ACL 2026
Journal ref: Findings of the Association for Computational Linguistics: ACL 2026, pages 25442-25457
DOI: 10.18653/v1/2026.findings-acl.1271
Code: https://github.com/rudolpheric/tm-reg
License: http://creativecommons.org/licenses/by/4.0/
The gist: This paper studies how empirical dialogue-flow statistics can be incorporated into Next Dialogue Act Prediction (NDAP).
Terminology
Abstract
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.
Sources
- OnCoCo 1.0: A Public Dataset for Fine-Grained Message Classification in Online Counseling Conversations
- Controllable Multi-Objective Re-ranking with Policy Hypernetworks
- Explicit State Tracking with Semi-Supervision for Neural Dialogue Generation
- HDLTex: Hierarchical Deep Learning for Text Classification
- Evaluating Role-Consistency in LLMs for Counselor Training
- Meta-control of Dialogue Systems Using Large Language Models
- AI Systems in Text-Based Online Counselling: Ethical Considerations Across Three Implementation Approaches
- Rethinking the Inception Architecture for Computer Vision
- DiactTOD: Learning Generalizable Latent Dialogue Acts for Controllable Task-Oriented Dialogue Systems
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data