measure(ment),1-1-I05-2014,bq translation output </term> . Automatic <term> evaluation metrics </term> for <term> Machine Translation ( MT
tech,4-1-I05-2014,bq <term> evaluation metrics </term> for <term> Machine Translation ( MT ) systems </term> , such as <term> BLEU </term> or <term>
measure(ment),13-1-I05-2014,bq Translation ( MT ) systems </term> , such as <term> BLEU </term> or <term> NIST </term> , are now well
measure(ment),15-1-I05-2014,bq </term> , such as <term> BLEU </term> or <term> NIST </term> , are now well established . Yet
other,10-2-I05-2014,bq scarcely used for the assessment of <term> language pairs </term> like <term> English-Chinese </term> or
other,13-2-I05-2014,bq of <term> language pairs </term> like <term> English-Chinese </term> or <term> English-Japanese </term> ,
other,15-2-I05-2014,bq like <term> English-Chinese </term> or <term> English-Japanese </term> , because of the <term> word segmentation
other,20-2-I05-2014,bq English-Japanese </term> , because of the <term> word segmentation problem </term> . This study establishes the equivalence
measure(ment),10-3-I05-2014,bq equivalence between the standard use of <term> BLEU </term> in <term> word n-grams </term> and its
other,12-3-I05-2014,bq standard use of <term> BLEU </term> in <term> word n-grams </term> and its application at the <term> character
other,19-3-I05-2014,bq n-grams </term> and its application at the <term> character </term> level . The use of <term> BLEU </term>
measure(ment),3-4-I05-2014,bq character </term> level . The use of <term> BLEU </term> at the <term> character </term> level
other,6-4-I05-2014,bq The use of <term> BLEU </term> at the <term> character </term> level eliminates the <term> word segmentation
other,10-4-I05-2014,bq character </term> level eliminates the <term> word segmentation problem </term> : it makes it possible to directly
tech,22-4-I05-2014,bq possible to directly compare commercial <term> systems </term> outputting <term> unsegmented texts
other,24-4-I05-2014,bq commercial <term> systems </term> outputting <term> unsegmented texts </term> with , for instance , <term> statistical
tech,31-4-I05-2014,bq texts </term> with , for instance , <term> statistical MT systems </term> which usually segment their <term>
other,38-4-I05-2014,bq </term> which usually segment their <term> outputs </term> . We present the first known <term>
skrij podrobnosti