We present BAYESUM (for ???Bayesian summarization???), a model for sentence extraction in query-focused summarization. BAYESUM leverages the common case in which multiple documents are relevant to a single query. Using these documents as reinforcement for query terms, BAYESUM is not afflicted by the paucity of information in short queries. We show that approximate inference in BAYESUM is possible on large data sets and results in a stateof-the-art summarization system. Furthermore, we show how BAYESUM can be understood as a justified query expansion technique in the language modeling for IR framework.
Key idea: 1. Extract and GENERALISE patterns. The patterns are generalised by creating word classes on the basis of their distributional similarity. 2. Validate the extracted patterns. The patterns are ranked by examining the frequencies of words in their prefix, infix and postfix. Candidate facts are ranked by checking whether they belong to some class as known (seed) facts.
%0 Conference Paper
%1 Daume:2006
%A Ill, Hal Daumé
%A Marcu, Daniel
%D 2006
%K summarisation
%P 305--312
%T Bayesian Query-Focused Summarization
%U http://www.isi.edu/\~marcu/papers.html
%X We present BAYESUM (for ???Bayesian summarization???), a model for sentence extraction in query-focused summarization. BAYESUM leverages the common case in which multiple documents are relevant to a single query. Using these documents as reinforcement for query terms, BAYESUM is not afflicted by the paucity of information in short queries. We show that approximate inference in BAYESUM is possible on large data sets and results in a stateof-the-art summarization system. Furthermore, we show how BAYESUM can be understood as a justified query expansion technique in the language modeling for IR framework.
@inproceedings{Daume:2006,
abstract = {We present BAYESUM (for ???Bayesian summarization???), a model for sentence extraction in query-focused summarization. BAYESUM leverages the common case in which multiple documents are relevant to a single query. Using these documents as reinforcement for query terms, BAYESUM is not afflicted by the paucity of information in short queries. We show that approximate inference in BAYESUM is possible on large data sets and results in a stateof-the-art summarization system. Furthermore, we show how BAYESUM can be understood as a justified query expansion technique in the language modeling for IR framework.},
added-at = {2011-08-05T10:08:44.000+0200},
author = {Ill, Hal Daum{\'e} and Marcu, Daniel},
biburl = {https://www.bibsonomy.org/bibtex/250b0b54c7a9a2e56b3ef3e95142c753b/diego_ma},
crossref = {ZZZ-COLINGACL:2006},
interhash = {fba42b320e9d083f1708c36262c30877},
intrahash = {50b0b54c7a9a2e56b3ef3e95142c753b},
keywords = {summarisation},
library = {Mine (October 2006)},
pages = {305--312},
review = {Key idea: 1. Extract and GENERALISE patterns. The patterns are generalised by creating word classes on the basis of their distributional similarity. 2. Validate the extracted patterns. The patterns are ranked by examining the frequencies of words in their prefix, infix and postfix. Candidate facts are ranked by checking whether they belong to some class as known (seed) facts.},
timestamp = {2011-08-05T10:08:44.000+0200},
title = {Bayesian Query-Focused Summarization},
url = {http://www.isi.edu/\~{}marcu/papers.html},
year = 2006
}