Survey of Indian Machine Translation Systems

Machine Translation (MT) is a sub-field of Artificial Intelligence (AI), which translates the text from one language known as source language into the text of another language known as target language. This paper gives a survey of the work done on various Indian machine translation systems either developed or under the development. Some systems are of general domain, but most of the systems have their own particular domains like parliamentary documents translation, news readings, children stories, web based information retrieval etc. 1. ANUSAARAKA A project named "ANUSAARAKAA" for machine translation from one Indian Language to another Language in 1995. It has been used for translation from Telugu, Kannada, Bengali, Punjabi and Marathi to Hindi language translation and vice versa. The ANUSAARAKA system is a language accessor cum Machine Translation system that works on principles of Paninian Grammar (PG)". The system provides both the robustness incase of failure and no loss of information while translating the text. The output of the system follows the grammar of the source language. The approach for the translation in this system is divided in two parts: 1) The ANUSAARAKA system which is based on language knowledge 2) the domain specific knowledge based on world knowledge, statistical knowledge etc. It was started at IIT Kanpur and now shifted to IIIT Hyderabad Currently ANUSAARAKA is working for Telugu, Kannada, Marathi, Bengali, and Punjabi to Hindi language translation and in near future reverse translation will also be feasible. The output can be post-edited if there is any grammatical error occurs during machine translation. The efficiency of the system was not available (2-4).