Recently, statistical machine translation models have begun to take advantage of higher level linguistic structures such as syntactic dependencies. Underlying these models is an assumption about the directness of translational correspondence between sentences in the two languages; however, the extent to which this assumption is valid and useful is not well understood. In this paper, we present an empirical study that quantifies the degree to which syntactic dependencies are preserved when parses are projected directly from English to Chinese. Our results show that although the direct correspondence assumption is often too restrictive, a small set of principled, elementary linguistic transformations can boost the quality of the projected Chinese parses by 76% relative to the unimproved baseline.
Description
Evaluating translational correspondence using annotation projection
%0 Conference Paper
%1 hwa2001etc
%A Hwa, Rebecca
%A Resnik, Philip
%A Weinberg, Amy
%A Kolak, Okan
%B ACL '02: Proceedings of the 40th Annual Meeting on Association for Computational Linguistics
%C Morristown, NJ, USA
%D 2001
%I Association for Computational Linguistics
%K Chinese DG English Master alignment dependency divergences toread translation
%P 392--399
%R http://dx.doi.org/10.3115/1073083.1073149
%T Evaluating translational correspondence using annotation projection
%U http://portal.acm.org/citation.cfm?doid=1073083.1073149
%X Recently, statistical machine translation models have begun to take advantage of higher level linguistic structures such as syntactic dependencies. Underlying these models is an assumption about the directness of translational correspondence between sentences in the two languages; however, the extent to which this assumption is valid and useful is not well understood. In this paper, we present an empirical study that quantifies the degree to which syntactic dependencies are preserved when parses are projected directly from English to Chinese. Our results show that although the direct correspondence assumption is often too restrictive, a small set of principled, elementary linguistic transformations can boost the quality of the projected Chinese parses by 76% relative to the unimproved baseline.
@inproceedings{hwa2001etc,
abstract = {Recently, statistical machine translation models have begun to take advantage of higher level linguistic structures such as syntactic dependencies. Underlying these models is an assumption about the directness of translational correspondence between sentences in the two languages; however, the extent to which this assumption is valid and useful is not well understood. In this paper, we present an empirical study that quantifies the degree to which syntactic dependencies are preserved when parses are projected directly from English to Chinese. Our results show that although the direct correspondence assumption is often too restrictive, a small set of principled, elementary linguistic transformations can boost the quality of the projected Chinese parses by 76% relative to the unimproved baseline.},
added-at = {2009-04-11T18:51:42.000+0200},
address = {Morristown, NJ, USA},
author = {Hwa, Rebecca and Resnik, Philip and Weinberg, Amy and Kolak, Okan},
biburl = {https://www.bibsonomy.org/bibtex/22de7dd31379cd4d956d20effd401be9f/unhammer},
booktitle = {ACL '02: Proceedings of the 40th Annual Meeting on Association for Computational Linguistics},
description = {Evaluating translational correspondence using annotation projection},
doi = {http://dx.doi.org/10.3115/1073083.1073149},
interhash = {c1192c44157a5fba5869d30bd446d5ac},
intrahash = {2de7dd31379cd4d956d20effd401be9f},
keywords = {Chinese DG English Master alignment dependency divergences toread translation},
location = {Philadelphia, Pennsylvania},
pages = {392--399},
publisher = {Association for Computational Linguistics},
timestamp = {2009-04-11T18:51:42.000+0200},
title = {Evaluating translational correspondence using annotation projection},
url = {http://portal.acm.org/citation.cfm?doid=1073083.1073149},
year = 2001
}