Neural-Based Machine Translation System Outperforming Statistical Phrase-Based Machine Translation for Low-Resource Languages

Natural Language Processing(NLP) involves the development of computational models that aid in the development of automated tools for processing and generating natural language. Human developing these computational models require deep insight of linguistic knowledge and are a time consuming process. Hence, to automate this process and accelerate the computational science we use a data-driven approach i.e Statistical learning and Deep Learning. For devolving and sharing of information in natural language and making it accessible in other natural languages, Machine Translation(MT) is entailed. It is an application of NLP. Sanskrit being ‘father of informatics’ [1] was considered as “lingua franca” of world intellectuals [2]. It is also an important language in the Indo-European family and considered as truly “donor” language of India. It has vast knowledge reserves in different discipline of studies such as Ayurveda, astronomy, literature etc. MT makes this rich language available to others with help of the computer. We have proposed and presented the prominent Deep Neural-based MT system for translation of Sanskrit to Hindi. We also present a comparison of Neural MT outperforming Statistical baseline system for this language pair.