Query snowball: A co-occurrence-based approach to multi-document summarization for question answering

Hajime Morita, Tetsuya Sakai, Manabu Okumura

Research output: Contribution to journalArticle

Abstract

We propose a new method for query-oriented extractive multi-document summarization. To enrich the information need representation of a given query, we build a co-occurrence graph to obtain words that augment the original query terms. We then formulate the summarization problem as a Maximum Coverage Problem with Knapsack Constraints based on word pairs rather than single words. Our experiments with the NTCIR ACLIA question answering test collections show that our method achieves a pyramid F3-score of up to 0.313, a 36% improvement over a baseline using Maximal Marginal Relevance.

Original languageEnglish
Pages (from-to)124-129
Number of pages6
JournalIPSJ Online Transactions
Volume5
Issue number2012
DOIs
Publication statusPublished - 2012
Externally publishedYes

Fingerprint

Experiments

Keywords

  • Information need representation
  • Multi-document summarization
  • Query-oriented
  • Question answering

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Query snowball : A co-occurrence-based approach to multi-document summarization for question answering. / Morita, Hajime; Sakai, Tetsuya; Okumura, Manabu.

In: IPSJ Online Transactions, Vol. 5, No. 2012, 2012, p. 124-129.

Research output: Contribution to journalArticle

@article{e058765e3a844d809130d748b4632a4a,
title = "Query snowball: A co-occurrence-based approach to multi-document summarization for question answering",
abstract = "We propose a new method for query-oriented extractive multi-document summarization. To enrich the information need representation of a given query, we build a co-occurrence graph to obtain words that augment the original query terms. We then formulate the summarization problem as a Maximum Coverage Problem with Knapsack Constraints based on word pairs rather than single words. Our experiments with the NTCIR ACLIA question answering test collections show that our method achieves a pyramid F3-score of up to 0.313, a 36{\%} improvement over a baseline using Maximal Marginal Relevance.",
keywords = "Information need representation, Multi-document summarization, Query-oriented, Question answering",
author = "Hajime Morita and Tetsuya Sakai and Manabu Okumura",
year = "2012",
doi = "10.2197/ipsjtrans.5.124",
language = "English",
volume = "5",
pages = "124--129",
journal = "IPSJ Online Transactions",
issn = "1882-6660",
publisher = "Information Processing Society of Japan",
number = "2012",

}

TY - JOUR

T1 - Query snowball

T2 - A co-occurrence-based approach to multi-document summarization for question answering

AU - Morita, Hajime

AU - Sakai, Tetsuya

AU - Okumura, Manabu

PY - 2012

Y1 - 2012

N2 - We propose a new method for query-oriented extractive multi-document summarization. To enrich the information need representation of a given query, we build a co-occurrence graph to obtain words that augment the original query terms. We then formulate the summarization problem as a Maximum Coverage Problem with Knapsack Constraints based on word pairs rather than single words. Our experiments with the NTCIR ACLIA question answering test collections show that our method achieves a pyramid F3-score of up to 0.313, a 36% improvement over a baseline using Maximal Marginal Relevance.

AB - We propose a new method for query-oriented extractive multi-document summarization. To enrich the information need representation of a given query, we build a co-occurrence graph to obtain words that augment the original query terms. We then formulate the summarization problem as a Maximum Coverage Problem with Knapsack Constraints based on word pairs rather than single words. Our experiments with the NTCIR ACLIA question answering test collections show that our method achieves a pyramid F3-score of up to 0.313, a 36% improvement over a baseline using Maximal Marginal Relevance.

KW - Information need representation

KW - Multi-document summarization

KW - Query-oriented

KW - Question answering

UR - http://www.scopus.com/inward/record.url?scp=84937045251&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84937045251&partnerID=8YFLogxK

U2 - 10.2197/ipsjtrans.5.124

DO - 10.2197/ipsjtrans.5.124

M3 - Article

AN - SCOPUS:84937045251

VL - 5

SP - 124

EP - 129

JO - IPSJ Online Transactions

JF - IPSJ Online Transactions

SN - 1882-6660

IS - 2012

ER -