tech,6-1-P05-1034,ak </term> . We describe a novel approach to <term> statistical machine translation </term> that combines <term> syntactic information
other,11-1-P05-1034,ak machine translation </term> that combines <term> syntactic information </term> in the <term> source language </term>
other,15-1-P05-1034,ak syntactic information </term> in the <term> source language </term> with recent advances in <term> phrasal
tech,21-1-P05-1034,ak language </term> with recent advances in <term> phrasal translation </term> . This method requires a <term> source-language
other,4-2-P05-1034,ak translation </term> . This method requires a <term> source-language </term><term> dependency parser </term> , <term>
tech,5-2-P05-1034,ak requires a <term> source-language </term><term> dependency parser </term> , <term> target language </term><term>
other,8-2-P05-1034,ak </term><term> dependency parser </term> , <term> target language </term><term> word segmentation </term> and
tech,10-2-P05-1034,ak </term> , <term> target language </term><term> word segmentation </term> and an <term> unsupervised word alignment
other,14-2-P05-1034,ak <term> word segmentation </term> and an <term> unsupervised word alignment component </term> . We align a <term> parallel corpus
lr,3-3-P05-1034,ak alignment component </term> . We align a <term> parallel corpus </term> , project the <term> source dependency
other,8-3-P05-1034,ak parallel corpus </term> , project the <term> source dependency parse </term> onto the <term> target sentence </term>
other,13-3-P05-1034,ak source dependency parse </term> onto the <term> target sentence </term> , extract <term> dependency treelet
other,17-3-P05-1034,ak <term> target sentence </term> , extract <term> dependency treelet translation pairs </term> , and train a <term> tree-based ordering
model,25-3-P05-1034,ak translation pairs </term> , and train a <term> tree-based ordering model </term> . We describe an efficient <term> decoder
tech,4-4-P05-1034,ak model </term> . We describe an efficient <term> decoder </term> and show that using these <term> tree-based
model,10-4-P05-1034,ak decoder </term> and show that using these <term> tree-based models </term> in combination with conventional <term>
model,16-4-P05-1034,ak </term> in combination with conventional <term> SMT models </term> provides a promising approach that
tech,27-4-P05-1034,ak approach that incorporates the power of <term> phrasal SMT </term> with the linguistic generality available
tech,36-4-P05-1034,ak linguistic generality available in a <term> parser </term> . In this paper , we present an <term>
skrij podrobnosti