Sangam: A Perso-Arabic to Indic Script Machine Transliteration Model

Indian sub-continent is one of those unique parts of the world where single languages are written in different scripts. This is the case for example with Punjabi, written in Indian East Punjab in Gurmukhi script (a Left to Right script based on Devnagri) and in Pakistani West Punjab, it is written in Shahmukhi (a Right to Left script based on Perso-Arabic). This is also the case with other languages like Urdu and Hindi (whilst having different names, they are the same language but written in mutually incomprehensible forms). Similarly, Sindhi and Kashmiri languages are written in both Persio-Arabic and Devanagri scripts. Thus there is a dire need for development transliteration tools for conversion between Perso-Arabic and Indic scripts. In this paper, we present Sangam, a Perso-Arabic to Indic script machine transliteration system, which can convert with high accuracy text written in Perso-Arabic script to one of the Indic script sharing the same language. Sangam is a hybrid system which combines rules as well as word and character level language models to transliterate the words. The system has been designed in such a fashion that the main code, algorithms and data structures remain unchanged and for a adding a new script pair only the databases, mapping rules and language models for the script pair need to be developed and plugged in. The system has been successfully tested on Punjabi, Urdu and Sindhi languages and can be easily extended for other languages like Kashmiri and Konkani.