Development of Telugu-Tamil Transfer-Based Machine Translation system: With Special reference to Divergence Index

The existence of translation divergence precludes straightforward mapping in machine translation (MT) system. An increase in the number of divergences also increases the complexity, especially in linguistically motivated transfer-based MT systems. In other words, divergence is directly proportional to the complexity of MT. Here we propose a divergence index (DI) to quantify the number of parametric variations between languages, which helps in improving the success rate of MT. This paper deals with how to build divergence index for a given language pair by giving examples between Telugu and Tamil, the major Dravidian languages spoken in South India. It also proposes handling strategies to overcome these divergences. The presentation of the paper also includes a live demo of Telugu-Tamil MT.