The first step to automated construction of a parallel corpus of Vietnamese Han-script poetry with contemporary Vietnamese interpretation

Abstract

In the progress of mining the treasure of Vietnamese Han-script documents, translating their meanings into contemporary Vietnamese is a crucial step in assisting the modern Vietnamese people in learning many priceless lessons from our ancestors. To support interpretation for a large number of these writings, we should create or enhance machine translation tools. However, due to a shortage of high-quality parallel corpora, the existing machine translation systems have struggled to do this task effectively, especially for the poetic domain. For this reason, we propose employing computer software to automatically create a Vietnamese Han-script corpus with contemporary Vietnamese interpretation, concentrating on the Vietnamese Han script poem domain. Our research shows that we constructed a parallel corpus with more than 4000 Vietnamese Han-script poems having Vietnamese interpretation and included information about the author, dynasty, etc. This corpus encourages automatic interpretation for Vietnam’s literary works written in Han script.