| tech,14-5-P05-1069,bq | <term> baseline </term> on a standard <term> | Arabic-English translation task | </term> . Previous work has used <term> monolingual | #9650 The best system obtains a 18.6% improvement over the baseline on a standardArabic-English translation task. |
| measure(ment),10-5-P05-1069,bq | obtains a 18.6 % improvement over the <term> | baseline | </term> on a standard <term> Arabic-English | #9646 The best system obtains a 18.6% improvement over thebaseline on a standard Arabic-English translation task. |
| other,27-3-P05-1069,bq | language model score </term> ) as well as <term> | binary features | </term> based on the <term> block </term> identities | #9612 We use a maximum likelihood criterion to train a log-linear block bigram model which uses real-valued features (e.g. a language model score) as well asbinary features based on the block identities themselves, e.g. block bigram features. |
| other,32-3-P05-1069,bq | binary features </term> based on the <term> | block | </term> identities themselves , e.g. block | #9617 We use a maximum likelihood criterion to train a log-linear block bigram model which uses real-valued features (e.g. a language model score) as well as binary features based on theblock identities themselves, e.g. block bigram features. |
| other,3-2-P05-1069,bq | </term> . The <term> model </term> predicts <term> | blocks | </term> with orientation to handle <term> local | #9576 The model predictsblocks with orientation to handle local phrase re-ordering. |
| other,8-4-P05-1069,bq | </term> can easily handle millions of <term> | features | </term> . The best system obtains a 18.6 | #9634 Our training algorithm can easily handle millions offeatures. |
| other,20-3-P05-1069,bq | real-valued features </term> ( e.g. a <term> | language model score | </term> ) as well as <term> binary features | #9605 We use a maximum likelihood criterion to train a log-linear block bigram model which uses real-valued features (e.g. alanguage model score) as well as binary features based on the block identities themselves, e.g. block bigram features. |
| other,8-2-P05-1069,bq | blocks </term> with orientation to handle <term> | local phrase re-ordering | </term> . We use a <term> maximum likelihood | #9581 The model predicts blocks with orientation to handlelocal phrase re-ordering. |
| 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 | #9561 In this paper, we present a novel training method for alocalized phrase-based prediction model for statistical machine translation (SMT). |
| 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 | #9594 We use a maximum likelihood criterion to train alog-linear block bigram model which uses real-valued features (e.g. a language model score) as well as binary features based on the block identities themselves, e.g. block bigram features. |
| 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 | #9588 We use amaximum likelihood criterion to train a log-linear block bigram model which uses real-valued features (e.g. a language model score) as well as binary features based on the block identities themselves, e.g. block bigram features. |
| model,1-2-P05-1069,bq | machine translation ( SMT ) </term> . The <term> | model | </term> predicts <term> blocks </term> with orientation | #9574 Themodel predicts blocks with orientation to handle local phrase re-ordering. |
| other,15-3-P05-1069,bq | block bigram model </term> which uses <term> | real-valued features | </term> ( e.g. a <term> language model score | #9600 We use a maximum likelihood criterion to train a log-linear block bigram model which usesreal-valued features (e.g. a language model score) as well as binary features based on the block identities themselves, e.g. block bigram features. |
| tech,1-4-P05-1069,bq | , e.g. block bigram features . Our <term> | training algorithm | </term> can easily handle millions of <term> | #9627 Ourtraining algorithm can easily handle millions of features. |
| tech,8-1-P05-1069,bq | In this paper , we present a novel <term> | training method | </term> for a <term> localized phrase-based | #9557 In this paper, we present a noveltraining method for a localized phrase-based prediction model for statistical machine translation (SMT). |
