Reordering Based Unsupervised Neural Machine Translation system for English To Telugu

Neural Machine Translation (NMT) is a progressing field of research among many Machine Translation (MT) systems. Many proposed works in machine translation have made standard benchmarks, and they differ from each other based on the model’s approach and type of learning. In this project, we aim to improve the performance of an unsupervised neural machine translation (UNMT) to give us better results by working only on the monolingual corpus. The unsupervised approach helps remove the barrier of translating languages with low-resource in a parallel corpus. Our model tries to extend the recent work on UNMT with reordering techniques to achieve better results.