P00-1067 must be segmented before word translation training , because written Chinese consists
P06-1011 enables us to extract useful machine translation training data even from very non-parallel
P13-1036 . Contrary to standard machine translation training scenarios , here we have to estimate
P06-1129 entropy-based statistical machine translation training . 3.4.1 Features Features used
W09-3106 models . It seems that literal translation training data better suit SMT system at
W09-1704 Language Exploitation Y3 Machine Translation training corpora ) to construct an English
W11-2206 Data We also used ACE 2007 entity translation training corpus which includes 119 Chinese-English
W11-2206 , 1994 ) . 5 . ACE2007 Entity Translation Training Data We also used ACE 2007 entity
W09-0425 the Moses statistical machine translation training script ( Koehn et al. , 2007
N13-1021 narrative , much as in machine translation training . Relying on manual transcripts
P09-2058 combined word alignments for phrase translation training , a natural choice for g is the
N06-4004 level . This provides additional translation training pairs that would otherwise be
W14-1805 of corpora in translation and translation training is a topic of some interest (
P14-1106 on the data combination of our translation training data and test data to get the
W12-4203 , that expanding not only the translation training data , but also the language
W10-1763 the vocabulary that is used in translation training and decoding ( see section 4.2
W14-3303 deployment of an SMT system for a given translation training corpus ( FDA5 ) , and the ParFDA5
D15-1218 considering the dearth of speech translation training datasets , this method allows
W15-4945 project-based translator training : Translation training in MNH-TT is carried out on the
D15-1136 source-target word alignment ) sentence translation training tuples and a corpus of ( source
hide detail