Detecting Free Translation in Parallel Corpora from Attention Scores
Qi Chen, Oi Yee Kwong, Jingbo Zhu · Institutional Repositories DataBase (IRDB) · 2018
In this study, we propose a method for extracting free translation examples from bilingual parallel corpora based on an innovative use of attention scores.Preliminary results show that the approach is promising and paraphrases at both sentential and sub-sentential levels covering diverse surface forms could be identified.The extracted data, upon further filtering, have great potential to supplement the example sentences available in existing bilingual dictionaries in an effective and systematic way.