A probabilistic model of meetings that combines words and discourse features

Mike Dowman, Virginia Savova, Thomas L. Griffiths, Konrad P. Körding, Joshua B. Tenenbaum, Matthew Purver

Research output: Contribution to journalArticlepeer-review

17 Scopus citations

Abstract

In order to determine the points at which meeting discourse changes from one topic to another, probabilistic models were used to approximate the process through which meeting transcripts were produced. Gibbs sampling was used to estimate the values of random variables in the models, including the locations of topic boundaries. This paper shows how discourse features were integrated into the Bayesian model and reports empirical evaluations of the benefit obtained through the inclusion of each feature and of the suitability of alternative models of the placement of topic boundaries. It demonstrates howmultiple cues to segmentation can be combined in a principled way, and empirical tests show a clear improvement over previous work.

Original languageEnglish (US)
Pages (from-to)1238-1248
Number of pages11
JournalIEEE Transactions on Audio, Speech and Language Processing
Volume16
Issue number7
DOIs
StatePublished - Sep 2008
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • Acoustics and Ultrasonics
  • Electrical and Electronic Engineering

Keywords

  • Gibbs sampling
  • Hierarchical bayesian models
  • Latent dirichlet allocation
  • Markov chain monte carlo
  • Topical segmentation

Fingerprint

Dive into the research topics of 'A probabilistic model of meetings that combines words and discourse features'. Together they form a unique fingerprint.

Cite this