Detection of agreement vs. disagreement in meetings: training with unlabeled data
D. Hillard, M. Ostendorf, and E. Shriberg. NAACL '03: Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology, page 34--36. Morristown, NJ, USA, Association for Computational Linguistics, (2003)
DOI: http://dx.doi.org/10.3115/1073483.1073495
Abstract
To support summarization of automatically transcribed meetings, we introduce a classifier to recognize agreement or disagreement utterances, utilizing both word-based and prosodic cues. We show that hand-labeling efforts can be minimized by using unsupervised training on a large unlabeled data set combined with supervised training on a small amount of data. For ASR transcripts with over 45% WER, the system recovers nearly 80% of agree/disagree utterances with a confusion rate of only 3%.
Description
Detection of agreement vs. disagreement in meetings
NAACL '03: Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology
%0 Conference Paper
%1 1073495
%A Hillard, Dustin
%A Ostendorf, Mari
%A Shriberg, Elizabeth
%B NAACL '03: Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology
%C Morristown, NJ, USA
%D 2003
%I Association for Computational Linguistics
%K agreement imported machinelearning
%P 34--36
%R http://dx.doi.org/10.3115/1073483.1073495
%T Detection of agreement vs. disagreement in meetings: training with unlabeled data
%U http://portal.acm.org/citation.cfm?id=1073483.1073495
%X To support summarization of automatically transcribed meetings, we introduce a classifier to recognize agreement or disagreement utterances, utilizing both word-based and prosodic cues. We show that hand-labeling efforts can be minimized by using unsupervised training on a large unlabeled data set combined with supervised training on a small amount of data. For ASR transcripts with over 45% WER, the system recovers nearly 80% of agree/disagree utterances with a confusion rate of only 3%.
@inproceedings{1073495,
abstract = {To support summarization of automatically transcribed meetings, we introduce a classifier to recognize agreement or disagreement utterances, utilizing both word-based and prosodic cues. We show that hand-labeling efforts can be minimized by using unsupervised training on a large unlabeled data set combined with supervised training on a small amount of data. For ASR transcripts with over 45% WER, the system recovers nearly 80% of agree/disagree utterances with a confusion rate of only 3%.},
added-at = {2007-10-21T14:13:29.000+0200},
address = {Morristown, NJ, USA},
author = {Hillard, Dustin and Ostendorf, Mari and Shriberg, Elizabeth},
biburl = {https://www.bibsonomy.org/bibtex/2052086fcc54bcd05d78760afea9d67d0/wnpxrz},
booktitle = {NAACL '03: Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology},
description = {Detection of agreement vs. disagreement in meetings},
doi = {http://dx.doi.org/10.3115/1073483.1073495},
interhash = {bba52b6c79e2cf95337cd34a967e5d8e},
intrahash = {052086fcc54bcd05d78760afea9d67d0},
keywords = {agreement imported machinelearning},
location = {Edmonton, Canada},
pages = {34--36},
publisher = {Association for Computational Linguistics},
timestamp = {2007-10-21T14:13:29.000+0200},
title = {Detection of agreement vs. disagreement in meetings: training with unlabeled data},
url = {http://portal.acm.org/citation.cfm?id=1073483.1073495},
year = 2003
}