AI Text-to-Speech Models for Sourashtra Language
#GS-3 #Science & Technology #ICT #National #Digital India BHASHINI
Key takeaways
- The Digital India BHASHINI Division and IISc Bengaluru developed AI models to convert Sourashtra text into speech.
- Sourashtra is an Indo-Aryan language spoken by approximately 2.48 lakh people in Tamil Nadu as per Census 2011.
- The models use VITS architecture and support both male and female voices in Tamil and Sourashtra scripts.
- The SPICOR TTS 3.0 Sourashtra corpus is now available on the AIKOSH platform for AI research.
Why in News
- The Digital India BHASHINI Division, which works under the Ministry of Electronics and Information Technology (MeitY), has teamed up with the Indian Institute of Science (IISc) Bengaluru. Together, they have created AI-based Text-to-Speech (TTS) models for Sourashtra. This language is a minority Indo-Aryan tongue spoken mostly in Tamil Nadu.
About Sourashtra
- Sourashtra is an Indo-Aryan language. It is spoken mainly by the Sourashtra community living in Tamil Nadu. According to the Census 2011, there are about 2.48 lakh speakers of this language.
Technology Used
- The system uses the VITS architecture, which stands for Variational Inference with adversarial learning for end-to-end TTS. In simple words, this technology converts written text directly into spoken words.
Language and Voice Features
- These models can produce both male and female voices. They are flexible enough to process text written in either the Tamil script or the Sourashtra script.
AIKOSH Platform
- The BHASHINI project has a platform called AIKOSH. This platform now hosts the SPICOR TTS 3.0 Sourashtra Corpus to help researchers working on AI for Indian languages.
Significance
- This project is important because it helps low-resource languages enter the digital world. It will make language learning easier and improve accessibility. It also helps in better documentation and scientific research.