Automatic meeting participant role detection by dialogue patterns

Jing Su; Bridget Kane; Saturnino Luz

doi:10.1007/978-3-642-12397-9-27

Automatic meeting participant role detection by dialogue patterns

Jing Su^*, Bridget Kane, Saturnino Luz

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We introduce a new concept of 'Vocalization Horizon' for automatic speaker role detection in general meeting recordings. We demonstrate that classification accuracy reaches 38.5% when Vocalization Horizon and other features (i.e. vocalization duration and start time) are available. With another type of Horizon, the Pause - Overlap Horizon, the classification accuracy reaches 39.5%. Pauses and overlaps are also useful vocalization features for meeting structure analysis. In our experiments, the Bayesian Network classifier outperforms other classifiers, and is proposed for similar applications.

Original language	English
Title of host publication	Development of Multimodal Interfaces
Subtitle of host publication	Active Listening and Synchrony - Second COST 2102 International Training School, Revised Selected Papers
Pages	314-326
Number of pages	13
DOIs	https://doi.org/10.1007/978-3-642-12397-9-27
Publication status	Published - 2010
Event	2nd COST 2102 International Training School on Development of Multimodal Interfaces: Active Listening and Synchrony - Dublin, Ireland Duration: 23 Mar 2009 → 27 Mar 2009

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	5967 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	2nd COST 2102 International Training School on Development of Multimodal Interfaces: Active Listening and Synchrony
Country/Territory	Ireland
City	Dublin
Period	23/03/09 → 27/03/09

Keywords / Materials (for Non-textual outputs)

Content-free meeting analysis
Meeting segmentation
Multimedia information retrieval
Speaker role detection

Access to Document

10.1007/978-3-642-12397-9-27

https://link.springer.com/chapter/10.1007/978-3-642-12397-9_27

Cite this

Su, J., Kane, B., & Luz, S. (2010). Automatic meeting participant role detection by dialogue patterns. In Development of Multimodal Interfaces: Active Listening and Synchrony - Second COST 2102 International Training School, Revised Selected Papers (pp. 314-326). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5967 LNCS). https://doi.org/10.1007/978-3-642-12397-9-27

Su, Jing ; Kane, Bridget ; Luz, Saturnino. / Automatic meeting participant role detection by dialogue patterns. Development of Multimodal Interfaces: Active Listening and Synchrony - Second COST 2102 International Training School, Revised Selected Papers. 2010. pp. 314-326 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{8e5a91d85b49480fa7cc2a4b68056326,

title = "Automatic meeting participant role detection by dialogue patterns",

abstract = "We introduce a new concept of 'Vocalization Horizon' for automatic speaker role detection in general meeting recordings. We demonstrate that classification accuracy reaches 38.5% when Vocalization Horizon and other features (i.e. vocalization duration and start time) are available. With another type of Horizon, the Pause - Overlap Horizon, the classification accuracy reaches 39.5%. Pauses and overlaps are also useful vocalization features for meeting structure analysis. In our experiments, the Bayesian Network classifier outperforms other classifiers, and is proposed for similar applications.",

keywords = "Content-free meeting analysis, Meeting segmentation, Multimedia information retrieval, Speaker role detection",

author = "Jing Su and Bridget Kane and Saturnino Luz",

year = "2010",

doi = "10.1007/978-3-642-12397-9-27",

language = "English",

isbn = "3642123961",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "314--326",

booktitle = "Development of Multimodal Interfaces",

note = "2nd COST 2102 International Training School on Development of Multimodal Interfaces: Active Listening and Synchrony ; Conference date: 23-03-2009 Through 27-03-2009",

}

Su, J, Kane, B & Luz, S 2010, Automatic meeting participant role detection by dialogue patterns. in Development of Multimodal Interfaces: Active Listening and Synchrony - Second COST 2102 International Training School, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5967 LNCS, pp. 314-326, 2nd COST 2102 International Training School on Development of Multimodal Interfaces: Active Listening and Synchrony, Dublin, Ireland, 23/03/09. https://doi.org/10.1007/978-3-642-12397-9-27

Automatic meeting participant role detection by dialogue patterns. / Su, Jing; Kane, Bridget; Luz, Saturnino.
Development of Multimodal Interfaces: Active Listening and Synchrony - Second COST 2102 International Training School, Revised Selected Papers. 2010. p. 314-326 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5967 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Automatic meeting participant role detection by dialogue patterns

AU - Su, Jing

AU - Kane, Bridget

AU - Luz, Saturnino

PY - 2010

Y1 - 2010

N2 - We introduce a new concept of 'Vocalization Horizon' for automatic speaker role detection in general meeting recordings. We demonstrate that classification accuracy reaches 38.5% when Vocalization Horizon and other features (i.e. vocalization duration and start time) are available. With another type of Horizon, the Pause - Overlap Horizon, the classification accuracy reaches 39.5%. Pauses and overlaps are also useful vocalization features for meeting structure analysis. In our experiments, the Bayesian Network classifier outperforms other classifiers, and is proposed for similar applications.

AB - We introduce a new concept of 'Vocalization Horizon' for automatic speaker role detection in general meeting recordings. We demonstrate that classification accuracy reaches 38.5% when Vocalization Horizon and other features (i.e. vocalization duration and start time) are available. With another type of Horizon, the Pause - Overlap Horizon, the classification accuracy reaches 39.5%. Pauses and overlaps are also useful vocalization features for meeting structure analysis. In our experiments, the Bayesian Network classifier outperforms other classifiers, and is proposed for similar applications.

KW - Content-free meeting analysis

KW - Meeting segmentation

KW - Multimedia information retrieval

KW - Speaker role detection

UR - http://www.scopus.com/inward/record.url?scp=77952073202&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-12397-9-27

DO - 10.1007/978-3-642-12397-9-27

M3 - Conference contribution

AN - SCOPUS:77952073202

SN - 3642123961

SN - 9783642123962

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 314

EP - 326

BT - Development of Multimodal Interfaces

T2 - 2nd COST 2102 International Training School on Development of Multimodal Interfaces: Active Listening and Synchrony

Y2 - 23 March 2009 through 27 March 2009

ER -

Su J, Kane B, Luz S. Automatic meeting participant role detection by dialogue patterns. In Development of Multimodal Interfaces: Active Listening and Synchrony - Second COST 2102 International Training School, Revised Selected Papers. 2010. p. 314-326. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-642-12397-9-27

Automatic meeting participant role detection by dialogue patterns

Abstract

Publication series

Conference

Keywords / Materials (for Non-textual outputs)

Access to Document

Other files and links

Fingerprint

Cite this