Query-oriented summarization based on neighborhood graph model

Furu Wei, Yanxiang He, Wenjie Li, Lei Huang

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

2 Citations (Scopus)

Abstract

In this paper, we investigate how to combine the link-aware and link-free information in sentence ranking for query-oriented summarization. Although the link structure has been emphasized in the existing graph-based summarization models, there is lack of pertinent analysis on how to use the links. By contrasting the text graph with the web graph, we propose to evaluate significance of sentences based on neighborhood graph model. Taking the advantage of the link information provided on the graph, each sentence is evaluated according to its own value as well as the cumulative impacts from its neighbors. For a task like query-oriented summarization, it is critical to explore how to reflect the influence of the query. To better incorporate query information into the model, we further design a query-sensitive similarity measure to estimate the association between a pair of sentences. When evaluated on DUC 2005 dataset, the results of the pro-posed approach are promising.
Original languageEnglish
Title of host publicationComputer Processing of Oriental Languages
Subtitle of host publicationLanguage Technology for the Knowledge-based Economy - 22nd International Conference, ICCPOL 2009, Proceedings
Pages156-167
Number of pages12
DOIs
Publication statusPublished - 9 Nov 2009
Event22nd International Conference on Computer Processing of Oriental Languages, ICCPOL 2009 - , Hong Kong
Duration: 26 Mar 200927 Mar 2009

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5459 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd International Conference on Computer Processing of Oriental Languages, ICCPOL 2009
Country/TerritoryHong Kong
Period26/03/0927/03/09

Keywords

  • Neighborhood Graph
  • Query-Oriented Summarization
  • Query-Sensitive Similarity

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Query-oriented summarization based on neighborhood graph model'. Together they form a unique fingerprint.

Cite this