A Novel Framework for Neural Machine Translation of Indian-English Languages

Abstract

The term 'Machine Translation' (MT) refers to computer systems that performs translation of natural language from one language to another. Machine Translation is especially necessitated in the Indian perspective because more than 50% of the data generated online is in English which is known by only 12%. Many systems has been proposed in the Indian perspective including rule-based, example-based, statistical based and a hybrid of these machine translation techniques. But, recent study has shown that Neural Machine Translation provides better results.In recent times Google and Facebook have developed Neural Machine Translation system. These systems are one-fits-all kind of systems which do not take into consideration the complexities in a language, like Indian languages. So, this paper proposes a broad framework for implementing Neural Machine Translation for Indian-English languages.