Designing Multi-lingual Machine-Translation System: Some Perspectives

摘要

The paper presents some perspectives of design of multi-lingual machine translation system in Indian perspective. India has 22 major officially recognized languages spoken by millions. However English which is spoken by less than 3% of the population, continues to be de-facto link language within the states of the country. Thus machine translation from English to Indian languages is one of the basic requirements. We present a system architecture which uses a pseudo-interlingua approach that caters to a group of Indian languages for translation from English. We use a hybrid approach that hybridizes rule-based and example-based approaches with domain customization tools.