Leveraging kullbackLeibler divergence measures and information-rich cues for speech summarization

Shih Hsiang Lin; Yao Ming Yeh; Berlin Chen

doi:10.1109/TASL.2010.2066268

Leveraging kullbackLeibler divergence measures and information-rich cues for speech summarization

Shih Hsiang Lin, Yao Ming Yeh, Berlin Chen

Department of Computer science & Information Engineering

Research output: Contribution to journal › Article

15 Citations (Scopus)

Abstract

Imperfect speech recognition often leads to degraded performance when exploiting conventional text-based methods for speech summarization. To alleviate this problem, this paper investigates various ways to robustly represent the recognition hypotheses of spoken documents beyond the top scoring ones. Moreover, a summarization framework, building on the KullbackLeibler (KL) divergence measure and exploring both the relevance and topical information cues of spoken documents and sentences, is presented to work with such robust representations. Experiments on broadcast news speech summarization tasks appear to demonstrate the utility of the presented approaches.

Original language	English
Article number	5549862
Pages (from-to)	871-882
Number of pages	12
Journal	IEEE Transactions on Audio, Speech and Language Processing
Volume	19
Issue number	4
DOIs	https://doi.org/10.1109/TASL.2010.2066268
Publication status	Published - 2011 Apr 6

Fingerprint

Keywords

KullbackLeibler (KL) -divergence
multiple recognition hypotheses
relevance information
speech summarization
topical information

ASJC Scopus subject areas

Acoustics and Ultrasonics
Electrical and Electronic Engineering

Cite this

Lin, S. H., Yeh, Y. M., & Chen, B. (2011). Leveraging kullbackLeibler divergence measures and information-rich cues for speech summarization. IEEE Transactions on Audio, Speech and Language Processing, 19(4), 871-882. [5549862]. https://doi.org/10.1109/TASL.2010.2066268

@article{da8445ee4b864407be7b74d12ed2f582,

title = "Leveraging kullbackLeibler divergence measures and information-rich cues for speech summarization",

abstract = "Imperfect speech recognition often leads to degraded performance when exploiting conventional text-based methods for speech summarization. To alleviate this problem, this paper investigates various ways to robustly represent the recognition hypotheses of spoken documents beyond the top scoring ones. Moreover, a summarization framework, building on the KullbackLeibler (KL) divergence measure and exploring both the relevance and topical information cues of spoken documents and sentences, is presented to work with such robust representations. Experiments on broadcast news speech summarization tasks appear to demonstrate the utility of the presented approaches.",

keywords = "KullbackLeibler (KL) -divergence, multiple recognition hypotheses, relevance information, speech summarization, topical information",

author = "Lin, {Shih Hsiang} and Yeh, {Yao Ming} and Berlin Chen",

year = "2011",

month = apr

day = "6",

doi = "10.1109/TASL.2010.2066268",

language = "English",

volume = "19",

pages = "871--882",

journal = "IEEE Transactions on Audio, Speech and Language Processing",

issn = "1558-7916",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "4",

}

TY - JOUR

T1 - Leveraging kullbackLeibler divergence measures and information-rich cues for speech summarization

AU - Lin, Shih Hsiang

AU - Yeh, Yao Ming

AU - Chen, Berlin

PY - 2011/4/6

Y1 - 2011/4/6

N2 - Imperfect speech recognition often leads to degraded performance when exploiting conventional text-based methods for speech summarization. To alleviate this problem, this paper investigates various ways to robustly represent the recognition hypotheses of spoken documents beyond the top scoring ones. Moreover, a summarization framework, building on the KullbackLeibler (KL) divergence measure and exploring both the relevance and topical information cues of spoken documents and sentences, is presented to work with such robust representations. Experiments on broadcast news speech summarization tasks appear to demonstrate the utility of the presented approaches.

AB - Imperfect speech recognition often leads to degraded performance when exploiting conventional text-based methods for speech summarization. To alleviate this problem, this paper investigates various ways to robustly represent the recognition hypotheses of spoken documents beyond the top scoring ones. Moreover, a summarization framework, building on the KullbackLeibler (KL) divergence measure and exploring both the relevance and topical information cues of spoken documents and sentences, is presented to work with such robust representations. Experiments on broadcast news speech summarization tasks appear to demonstrate the utility of the presented approaches.

KW - KullbackLeibler (KL) -divergence

KW - multiple recognition hypotheses

KW - relevance information

KW - speech summarization

KW - topical information

UR - http://www.scopus.com/inward/record.url?scp=79953287641&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79953287641&partnerID=8YFLogxK

U2 - 10.1109/TASL.2010.2066268

DO - 10.1109/TASL.2010.2066268

M3 - Article

AN - SCOPUS:79953287641

VL - 19

SP - 871

EP - 882

JO - IEEE Transactions on Audio, Speech and Language Processing

JF - IEEE Transactions on Audio, Speech and Language Processing

SN - 1558-7916

IS - 4

M1 - 5549862

ER -

Access to Document

10.1109/TASL.2010.2066268