Deep Learning in Natural Language Processing by Li Deng & Yang Liu

Author:Li Deng & Yang Liu
Language: eng
Format: epub
Publisher: Springer Singapore, Singapore

6.3.3 Deep Learning for Reordering Phrases

For a source language sentence , phrasal translation rules match the sentence, segment the word sequence into phrase sequence, and map each source phrase into the target language phrase using the neural rule selection model discussed in the previous section. The next task needs to rearrange the target phrases to produce a well-formed translation. This phrase reordering task is usually casted as a binary classification problem for any two neighboring target phrases: keep the two phrases in order (monotone) or swap the two phrases. For the two neighboring source phrases , , and their translation candidates and , the reordering model utilizes only the boundary discrete words of the four phrases as features and adopts a maximum entropy model to predict the reordering probability (Xiong et al. 2006):

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.