IIT Guwahati researchers develop speech technologies for North Eastern languages

NEW DELHI: The Indian Institute of Technology (IIT) Guwahati is developing a project ‘Speech Technologies for North Eastern Languages’ to develop speech technology tools for healthcare information dissemination. The tools will enable retrieval of healthcare related information with the help of spoken keyword spotting (KWS) in seven North East Indian languages.

This project involves building speech technology tools for healthcare information dissemination in Hindi, English, Assamese, Bangla, Bodo, Manipuri, Khasi, Mizo, Nagamese, and Nepali, said the statement.

Also Read | IIT Delhi startup develops technology to manufacture highly affordable antiviral fabric

As part of the project a database of health-related information in seven languages spoken in North East India will also be created. This project aims to facilitate the access of healthcare related information by the people in the far flung areas of North East India in their own native languages.

The Centre for Linguistic Science and Technology (CLST) at IIT Guwahati has got funding for this project from the Ministry of Electronics and Information Technology under its ‘National Language Translation Mission (NLTM): BHASHINI’ initiative, according to an official statement from IIT Guwahati.

IIT Guwahati's project reflects NEP 2020 spirit

Highlighting the unique aspects of this project, Professor T G Sitharam, Director, IIT Guwahati, said, “This work embodies IIT Guwahati’s commitment to work for the local languages and ethnicities of North East India. The interdisciplinary nature of the project and the focus on local languages reflect the spirit envisaged in the National Education Policy, 2020.”

Also Read | IIT Mandi organises workshop on Indian knowledge system, mental health

Elaborating on this project, Professor Rohit Sinha, Principal Investigator of project, and Head, Department of Electronics and Electrical Engineering, IIT Guwahati, said, "The Institute is committed to developing tools that will facilitate last-mile connectivity and information dissemination to the various communities living in the NE area, in their own languages. This project will be a step towards achieving that aim.”

Also Read | IIT Madras launches initiative to reduce gender data gap in Wikipedia

The Spoken Keyword Spotting (KWS) systems developed in the project will be able to detect a list of predefined words in a given speech signal of one of the target languages of the project. The efforts will involve modelling speech with the deep neural network based state-of-the-art techniques, the statement further said.