Empirical Comparison of Machine Translation Techniques

Authors

  • *Arun Ramrao Babhulgaonkar , Shefali Sonavane

Abstract

Natural language is the medium used by human being to express feelings during
communication. Grammatical, structural and semantic differences in natural languages block the
communication between any two cultures around the world. Machine translation (MT) is the remedy for
this real life problem in society. Machine translation is the process of translating a text in one natural
language into another natural language using computer system. In the current age of technology
automated machine translation is preferred over human translation due to its quick response and other
advantages. Phrase based, Factored and Neural network based techniques are widely used nowadays
for machine translation. This paper compares these three different techniques used for automated
machine translation. Each of these techniques faces certain challenges and has its own pros and cons.
Also, the translation performance of these techniques is compared using manual and automated metrics
like BLEU and TER. Evolutionary nature of natural language still poses many challenges for further
research in machine translation.

Published

2020-03-31

Issue

Section

Articles