ML Anthology

A Boosting Approach to Topic Spotting on Subdialogues

Kary L. Myers, Michael J. Kearns, Satinder Singh, Marilyn A. Walker
ICML 2000 pp. 655-662
/icml/2000/myers2000icml-boosting/

Abstract

We report the results of a study on topic spotting in conversational speech. Using a machine learn-ing approach, we build classifiers that accept an audio file of conversational human speech as in-put, and output an estimate of the topic being dis-cussed. Our methodology makes use of a well-known corpus of transcribed and topic-labeled speech (the Switchboard corpus), and involves an interesting double use of the BOOSTEXTER learning algorithm. Our work is distinguished from previous efforts in topic spotting by our ex-plicit study of the effects of dialogue length on classifier performance, and by our use of off-the-shelf speech recognition technology. One of our main results is the identification of a single clas-sifier with good performance (relative to our clas-sifier space) across all subdialogue lengths. 1.

Semantic Scholar

Cite

Text

Myers et al. "A Boosting Approach to Topic Spotting on Subdialogues." International Conference on Machine Learning, 2000.

Markdown

[Myers et al. "A Boosting Approach to Topic Spotting on Subdialogues." International Conference on Machine Learning, 2000.](https://mlanthology.org/icml/2000/myers2000icml-boosting/)

BibTeX

@inproceedings{myers2000icml-boosting,
  title     = {{A Boosting Approach to Topic Spotting on Subdialogues}},
  author    = {Myers, Kary L. and Kearns, Michael J. and Singh, Satinder and Walker, Marilyn A.},
  booktitle = {International Conference on Machine Learning},
  year      = {2000},
  pages     = {655-662},
  url       = {https://mlanthology.org/icml/2000/myers2000icml-boosting/}
}