Dialogue Act Recognition using Reweighted Speaker Adaptation

Abstract

In this work we study the effectiveness of speaker adaptation for dialogue act recognition in multiparty meetings. First, we analyze idiosyncracy in dialogue verbal acts by qualitatively studying the differences and conflicts among speakers and by quantitively comparing speaker-specific models. Based on these observations, we propose a new approach for dialogue act recognition based on reweighted domain adaptation which effectively balance the influence of speaker specific and other speakers? data. Our experiments on a real-world meeting dataset show that with even only 200 speaker-specific annotated dialogue acts, the performances on dialogue act recognition are significantly improved when compared to several baseline algorithms. To our knowledge, this work is the first 1 to tackle this promising research direction of speaker adaptation for dialogue act recognition.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 2012
Accession Number
ADA563217

Entities

People

  • Congkai Sun
  • Louis-Philippe Morency

Organizations

  • University of Southern California

Tags

Communities of Interest

  • Autonomy

DTIC Thesaurus Topics

  • Age Distribution
  • Algorithms
  • Artificial Intelligence Software
  • Automated Speech Recognition
  • Automatic
  • Computational Linguistics
  • Computational Science
  • Computer Languages
  • Language
  • Linguistics
  • Machine Learning
  • Natural Language Processing
  • Natural Languages
  • Observation
  • Recognition
  • Standards
  • Test Sets

Fields of Study

  • Computer science

Readers

  • Defense Acquisition Program Management
  • Neural Network Machine Learning.
  • Speech Processing/Speech Recognition.