Attention Based Translation For Out Of Vocabulary Problem

  • Rama S., G. Paavai Anand

Abstract

Reading in the regional language enables better absorption, retention, and application of what has been learnt and many researchers promote it. Unfortunately, a majority of books on various subjects exist in the English language only. Learners struggle to comprehend and acquire essential learning from the translated versions of such books. Automated machine translation is apprehended here and we restrict the scope to translating the English-based statements to Tamil at higher precision and reliability. This is avid with the advent of high-end computational devices and neural translation systems.  Out of vocabulary (OOV) is a very big common problem where many new technical terms and other rare words are not translated properly. Here we defined a byte pair translation method for translating English to Tamil content with an accuracy of 68% and BLEU score of   7.18.

Published
2020-06-01
How to Cite
Rama S., G. Paavai Anand. (2020). Attention Based Translation For Out Of Vocabulary Problem. International Journal of Advanced Science and Technology, 29(08), 2287 - 2291. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/23387
Section
Articles