Unsupervised Dialogue Act Induction using Gaussian Mixtures
Not on the map. Why?
Abstract (arXiv)
This paper introduces a new unsupervised approach for dialogue act induction. Given the sequence of dialogue utterances, the task is to assign them the labels representing their function in the dialogue. Utterances are represented as real-valued vectors encoding their meaning. We model the dialogue as Hidden Markov model with emission probabilities estimated by Gaussian mixtures. We use Gibbs sampling for posterior inference. We present the results on the standard Switchboard-DAMSL corpus. Our algorithm achieves promising results compared with strong supervised baselines and outperforms other unsupervised algorithms.
Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). First posted on arXiv 20 Dec 2016.
Builds on
No reference list yet.
Built on by
Citing papers per year: 2022: 5; 2023: 1.
6 papers.
- 2022Learning Discriminative Representations and Decision Boundaries for Open Intent Detectioncited by 15
- 2022Workflow Discovery from Dialogues in the Low Data Regimecited by 12
- 2022Unsupervised Learning of Hierarchical Conversation Structurecited by 5
- 2022Learning Interpretable Latent Dialogue Actions With Less Supervisioncited by 3
- 2022Learning Better Intent Representations for Financial Open Intent Classificationcited by 3
- 2023Effective Open Intent Classification with K-center Contrastive Learning and Adjustable Decision Boundarycited by 3
- 2023Effective Open Intent Classification with K-center Contrastive Learning and Adjustable Decision Boundary20 Apr 2023
- 2022Learning Better Intent Representations for Financial Open Intent Classification25 Oct 2022
- 2022Learning Interpretable Latent Dialogue Actions With Less Supervision22 Sep 2022
- 2022Unsupervised Learning of Hierarchical Conversation Structure24 May 2022
- 2022Workflow Discovery from Dialogues in the Low Data Regime24 May 2022
- 2022Learning Discriminative Representations and Decision Boundaries for Open Intent Detection11 Mar 2022