tech,8-1-P05-1069,bq In this paper , we present a novel <term> training method </term> for a <term> localized phrase-based
model,12-1-P05-1069,bq novel <term> training method </term> for a <term> localized phrase-based prediction model </term> for <term> statistical machine translation
tech,17-1-P05-1069,bq phrase-based prediction model </term> for <term> statistical machine translation ( SMT ) </term> . The <term> model </term> predicts <term>
model,1-2-P05-1069,bq machine translation ( SMT ) </term> . The <term> model </term> predicts <term> blocks </term> with orientation
other,3-2-P05-1069,bq </term> . The <term> model </term> predicts <term> blocks </term> with orientation to handle <term> local
other,8-2-P05-1069,bq blocks </term> with orientation to handle <term> local phrase re-ordering </term> . We use a <term> maximum likelihood
other,3-3-P05-1069,bq phrase re-ordering </term> . We use a <term> maximum likelihood criterion </term> to train a <term> log-linear block
model,9-3-P05-1069,bq likelihood criterion </term> to train a <term> log-linear block bigram model </term> which uses <term> real-valued features
other,15-3-P05-1069,bq block bigram model </term> which uses <term> real-valued features </term> ( e.g. a <term> language model score
other,20-3-P05-1069,bq real-valued features </term> ( e.g. a <term> language model score </term> ) as well as <term> binary features
other,27-3-P05-1069,bq language model score </term> ) as well as <term> binary features </term> based on the <term> block </term> identities
other,32-3-P05-1069,bq binary features </term> based on the <term> block </term> identities themselves , e.g. block
tech,1-4-P05-1069,bq , e.g. block bigram features . Our <term> training algorithm </term> can easily handle millions of <term>
other,8-4-P05-1069,bq </term> can easily handle millions of <term> features </term> . The best system obtains a 18.6
measure(ment),10-5-P05-1069,bq obtains a 18.6 % improvement over the <term> baseline </term> on a standard <term> Arabic-English
tech,14-5-P05-1069,bq <term> baseline </term> on a standard <term> Arabic-English translation task </term> . Previous work has used <term> monolingual
skrij podrobnosti