Developing a voice recognition model of Hre and Co ethnic minorities into Vietnamese text using artificial intelligence

Abstract

Building an electronic database of the ethnic minority languages of Viet - Hre, Viet - Co is extremely necessary to narrow the communication and language gap between the Hre, Co and Kinh people. Currently, AI speech recognition technology has achieved high accuracy with Vietnamese and many other languages, allowing the application of Vietnamese recognition in the Vietnamese - Hre and Vietnamese - Co translation direction of the electronic database. However, the opposite direction - recognition and translation of Hre, Co into Vietnamese - has never been studied before. In this article, the corresponding word pairs to create the Vietnamese - Hre, Vietnamese - Co data warehouse have been digitized uniformly. Thanks to the built data warehouse, the authors has successfully developed an electronic database software package that allows semantic lookup between the Vietnamese - Hre, Vietnamese - Co languages. This software package has both a web version and a mobile application “Hre - Co - Viet”. In particular, the authors also researched and successfully built a speech recognition model for Hre and Co languages, expanding the application capabilities for this electronic database.