Topic segmentation and labeling in asynchronous conversations

Shafiq Rayhan Joty, Giuseppe Carenini, Raymond T. Ng

Research output: Contribution to journalArticle

27 Citations (Scopus)

Abstract

Topic segmentation and labeling is often considered a prerequisite for higher-level conversation analysis and has been shown to be useful in many Natural Language Processing (NLP) applications. We present two new corpora of email and blog conversations annotated with topics, and evaluate annotator reliability for the segmentation and labeling tasks in these asynchronous conversations. We propose a complete computational framework for topic segmentation and labeling in asynchronous conversations. Our approach extends state-of-the-art methods by considering a fine-grained structure of an asynchronous conversation, along with other conversational features by applying recent graph-based methods for NLP. For topic segmentation, we propose two novel unsupervised models that exploit the fine-grained conversational structure, and a novel graph-theoretic supervised model that combines lexical, conversational and topic features. For topic labeling, we propose two novel (unsupervised) random walk models that respectively capture conversation specific clues from two different sources: the leading sentences and the fine-grained conversational structure. Empirical evaluation shows that the segmentation and the labeling performed by our best models beat the state-of-the-art, and are highly correlated with human annotations.

Original languageEnglish
Pages (from-to)521-573
Number of pages53
JournalJournal of Artificial Intelligence Research
Volume47
Publication statusPublished - 31 Jul 2013

Fingerprint

Labeling
Blogs
Electronic mail
Processing

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Topic segmentation and labeling in asynchronous conversations. / Rayhan Joty, Shafiq; Carenini, Giuseppe; Ng, Raymond T.

In: Journal of Artificial Intelligence Research, Vol. 47, 31.07.2013, p. 521-573.

Research output: Contribution to journalArticle

@article{94defa9b5bca49c6a9675c419c92398b,
title = "Topic segmentation and labeling in asynchronous conversations",
abstract = "Topic segmentation and labeling is often considered a prerequisite for higher-level conversation analysis and has been shown to be useful in many Natural Language Processing (NLP) applications. We present two new corpora of email and blog conversations annotated with topics, and evaluate annotator reliability for the segmentation and labeling tasks in these asynchronous conversations. We propose a complete computational framework for topic segmentation and labeling in asynchronous conversations. Our approach extends state-of-the-art methods by considering a fine-grained structure of an asynchronous conversation, along with other conversational features by applying recent graph-based methods for NLP. For topic segmentation, we propose two novel unsupervised models that exploit the fine-grained conversational structure, and a novel graph-theoretic supervised model that combines lexical, conversational and topic features. For topic labeling, we propose two novel (unsupervised) random walk models that respectively capture conversation specific clues from two different sources: the leading sentences and the fine-grained conversational structure. Empirical evaluation shows that the segmentation and the labeling performed by our best models beat the state-of-the-art, and are highly correlated with human annotations.",
author = "{Rayhan Joty}, Shafiq and Giuseppe Carenini and Ng, {Raymond T.}",
year = "2013",
month = "7",
day = "31",
language = "English",
volume = "47",
pages = "521--573",
journal = "Journal of Artificial Intelligence Research",
issn = "1076-9757",
publisher = "Morgan Kaufmann Publishers, Inc.",

}

TY - JOUR

T1 - Topic segmentation and labeling in asynchronous conversations

AU - Rayhan Joty, Shafiq

AU - Carenini, Giuseppe

AU - Ng, Raymond T.

PY - 2013/7/31

Y1 - 2013/7/31

N2 - Topic segmentation and labeling is often considered a prerequisite for higher-level conversation analysis and has been shown to be useful in many Natural Language Processing (NLP) applications. We present two new corpora of email and blog conversations annotated with topics, and evaluate annotator reliability for the segmentation and labeling tasks in these asynchronous conversations. We propose a complete computational framework for topic segmentation and labeling in asynchronous conversations. Our approach extends state-of-the-art methods by considering a fine-grained structure of an asynchronous conversation, along with other conversational features by applying recent graph-based methods for NLP. For topic segmentation, we propose two novel unsupervised models that exploit the fine-grained conversational structure, and a novel graph-theoretic supervised model that combines lexical, conversational and topic features. For topic labeling, we propose two novel (unsupervised) random walk models that respectively capture conversation specific clues from two different sources: the leading sentences and the fine-grained conversational structure. Empirical evaluation shows that the segmentation and the labeling performed by our best models beat the state-of-the-art, and are highly correlated with human annotations.

AB - Topic segmentation and labeling is often considered a prerequisite for higher-level conversation analysis and has been shown to be useful in many Natural Language Processing (NLP) applications. We present two new corpora of email and blog conversations annotated with topics, and evaluate annotator reliability for the segmentation and labeling tasks in these asynchronous conversations. We propose a complete computational framework for topic segmentation and labeling in asynchronous conversations. Our approach extends state-of-the-art methods by considering a fine-grained structure of an asynchronous conversation, along with other conversational features by applying recent graph-based methods for NLP. For topic segmentation, we propose two novel unsupervised models that exploit the fine-grained conversational structure, and a novel graph-theoretic supervised model that combines lexical, conversational and topic features. For topic labeling, we propose two novel (unsupervised) random walk models that respectively capture conversation specific clues from two different sources: the leading sentences and the fine-grained conversational structure. Empirical evaluation shows that the segmentation and the labeling performed by our best models beat the state-of-the-art, and are highly correlated with human annotations.

UR - http://www.scopus.com/inward/record.url?scp=84880660500&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84880660500&partnerID=8YFLogxK

M3 - Article

VL - 47

SP - 521

EP - 573

JO - Journal of Artificial Intelligence Research

JF - Journal of Artificial Intelligence Research

SN - 1076-9757

ER -