Exploring Correlation of Dependency Relation Paths for Answer Extraction
D. Shen, and D. Klakow. Proceedings COLING/ACL 2006, page 889-896. Sydney, (2006)
Abstract
In this paper, we explore correlation of dependency relation paths to rank candidate answers in answer extraction. Using the correlation measure, we compare dependency relations of a candidate answer and mapped question phrases in sentence with the corresponding relations in question. Different from previous studies, we propose an approximate phrase mapping algorithm and incorporate the mapping score into the correlation measure. The correlations are further incorporated into a Maximum Entropy-based ranking model which estimates path weights from training. Experimental results show that our method significantly outperforms state-ofthe-art syntactic relation-based methods by up to 20% in MRR.
%0 Conference Paper
%1 Shen:2006
%A Shen, Dan
%A Klakow, Dietrich
%B Proceedings COLING/ACL 2006
%C Sydney
%D 2006
%K question_answering machine_learning dependencies DG
%P 889-896
%T Exploring Correlation of Dependency Relation Paths for Answer Extraction
%U http://acl.ldc.upenn.edu/P/P06/P06-1112.pdf
%X In this paper, we explore correlation of dependency relation paths to rank candidate answers in answer extraction. Using the correlation measure, we compare dependency relations of a candidate answer and mapped question phrases in sentence with the corresponding relations in question. Different from previous studies, we propose an approximate phrase mapping algorithm and incorporate the mapping score into the correlation measure. The correlations are further incorporated into a Maximum Entropy-based ranking model which estimates path weights from training. Experimental results show that our method significantly outperforms state-ofthe-art syntactic relation-based methods by up to 20% in MRR.
@inproceedings{Shen:2006,
abstract = {In this paper, we explore correlation of dependency relation paths to rank candidate answers in answer extraction. Using the correlation measure, we compare dependency relations of a candidate answer and mapped question phrases in sentence with the corresponding relations in question. Different from previous studies, we propose an approximate phrase mapping algorithm and incorporate the mapping score into the correlation measure. The correlations are further incorporated into a Maximum Entropy-based ranking model which estimates path weights from training. Experimental results show that our method significantly outperforms state-ofthe-art syntactic relation-based methods by up to 20% in MRR.},
added-at = {2009-11-11T22:33:09.000+0100},
address = {Sydney},
author = {Shen, Dan and Klakow, Dietrich},
biburl = {https://www.bibsonomy.org/bibtex/2b385d2d62a1cec0bcaa6f01019112f65/diego_ma},
booktitle = {Proceedings COLING/ACL 2006},
interhash = {6966be5911653238e919c7e841c639d8},
intrahash = {b385d2d62a1cec0bcaa6f01019112f65},
keywords = {question_answering machine_learning dependencies DG},
pages = {889-896},
timestamp = {2009-11-11T22:33:15.000+0100},
title = {Exploring Correlation of Dependency Relation Paths for Answer Extraction},
url = {http://acl.ldc.upenn.edu/P/P06/P06-1112.pdf},
year = 2006
}