Unsupervised Dialogue Act Induction using Gaussian Mixtures

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.

• Publication year

  2016

• Venue

  Conference of the European Chapter of the Association for Computational Linguistics

• Publication date

  2016-12-20

• Fields of study

  Computer Science

• Identifiers
  DOI 10.18653/V1/E17-2078arXiv 1612.06572
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Artificial Intelligence

  2019cited by this paper
• UWB at SemEval-2016 Task 1: Semantic Textual Similarity using Lexical, Syntactic, and Semantic Information

  2016cited by this paper
• Gaussian LDA for Topic Models with Word Embeddings

  2015cited by this paper
• Distributed Representations for Unsupervised Semantic Role Labeling

  2015cited by this paper
• GloVe: Global Vectors for Word Representation

  2014cited by this paper
• Recurrent Convolutional Neural Networks for Discourse Compositionality

  2013influential reference
• Machine learning - a probabilistic perspective

  2012cited by this paper
• Mixed Membership Markov Models for Unsupervised Conversation Modeling

  2012influential reference
• Predicting Thread Discourse Structure over Technical Web Forums

  2011cited by this paper
• Unsupervised Modeling of Dialog Acts in Asynchronous Conversations

  2011cited by this paper
• Two Decades of Unsupervised POS Induction: How Far Have We Come?

  2010cited by this paper
• Using Syntactic and Semantic based Relations for Dialogue Act Recognition

  2010cited by this paper
• Unsupervised Modeling of Twitter Conversations

  2010influential reference
• Online community search using thread structure

  2009cited by this paper
• Unsupervised Classification of Dialogue Acts using a Dirichlet Process Mixture Model

  2009cited by this paper
• Speech and Language Processing, 2nd Edition

  2008cited by this paper
• V-Measure: A Conditional Entropy-Based External Cluster Evaluation Measure

  2007cited by this paper
• PLOW: A Collaborative Task Learning Agent

  2007cited by this paper
• Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL)

  2007cited by this paper
• Relationship Identification for Social Network Discovery

  2007cited by this paper
• Finding scientific topics

  2004cited by this paper
• Dialogue act modeling for automatic tagging and recognition of conversational speech

  2000cited by this paper
• Speech and language processing: an introduction to natural language processing, computational linguistics, and speech recognition, 2nd Edition

  2000cited by this paper
• Switchboard SWBD-DAMSL shallow-discourse-function annotation coders manual

  1997cited by this paper
• A Maximum Entropy Approach to Natural Language Processing

  1996cited by this paper
• The Principle of Semantic Compositionality

  1994cited by this paper
• of the Association for Computational Linguistics

  year unknowncited by this paper

• Multi-path feature fusion and constrained boundary optimization for open intent detection

  2025cites this paper
• Revisiting clustering for efficient unsupervised dialogue structure induction

  2024cites this paper
• Effective Open Intent Classification with K-center Contrastive Learning and Adjustable Decision Boundary

  2023cites this paper
• Class Lifelong Learning for Intent Detection via Structure Consolidation Networks

  2023cites this paper
• Towards Open Intent Detection

  2022cites this paper
• Unsupervised Learning of Hierarchical Conversation Structure

  2022influential citation
• Workflow Discovery from Dialogues in the Low Data Regime

  2022cites this paper
• Learning Discriminative Representations and Decision Boundaries for Open Intent Detection

  2022influential citation
• Learning Interpretable Latent Dialogue Actions With Less Supervision

  2022cites this paper
• Learning Better Intent Representations for Financial Open Intent Classification

  2022cites this paper
• Deep Open Intent Classification with Adaptive Decision Boundary

  2020cites this paper
• Dialogue Systems in a Low-resource Scenario PhD. Thesis Proposal

  2020cites this paper
• Emerging Intention Mining Inspired by Semantic Reconstruction

  2020cites this paper
• Discovering New Intents with Deep Aligned Clustering

  2020cites this paper
• Using Functional Schemas to Understand Social Media Narratives

  2019cites this paper
• A post-processing method for detecting unknown intent of dialogue system via pre-trained deep neural network classifier

  2019cites this paper
• Deep Unknown Intent Detection with Margin Loss

  2019cites this paper
• Dialogue Act Recognition for Open-Domain Based on Word-Level Sequence Annotation with CRF

  2018cites this paper
• Unsupervised Dialogue Act Classification with Optimum-Path Forest

  2018influential citation
• Cross-lingual Word Analogies using Linear Transformations between Semantic Spaces

  2018cites this paper
• Modeling Dialogue Acts with Content Word Filtering and Speaker Preferences

  2017influential citation
• Open Text Classi�cation Based on Dynamic Boundary Balance

  year unknowncites this paper