Part of National Language Translation Mission (NLTM) under BHASHINI,
Funded by MeitY, Govt. of India
Developing advanced machine translation systems across Dravidian languages
3 Years Research Project
AU-KBC Research Centre,
MIT Campus of Anna University, Chennai
Discourse-level translation for coherent output
The DL-DiscoMT project aims to build a Speech to Speech Translation System (SS1MT) across Indian languages including Indian English to overcome the language barrier in India’s multilingual context. It focuses on developing Machine Translation (MT) systems that can translate speech or video transcripts from English to Indian languages and connect different Indian language families.
The project also includes Dravidian language to Dravidian language (DL DiscoMT) translation as an important component through Hindi to Tamil and from Tamil to other Dravidian languages. Since translation is not only at sentence level but also at cross sentential level, this project includes discourse information to improve translation quality, based on experience from earlier projects such as SSMT funded by PSA and IL-IL MT funded by MeitY.
Develop and deploy machine translation systems from Hindi to Tamil and Tamil to other Dravidian languages
Include discourse analysis to improve coherence and cohesion in translated text beyond sentence level
Create an ecosystem involving startups and government institutions for innovative translation services
Increase Dravidian language content on the internet across governance, science, health, and agriculture domains