Shillong, October 14: A Shillong-based AI research lab has showcased a speech-to-text model for Garo at a Google event in New Delhi, marking a step towards bringing indigenous languages of the Northeast into mainstream artificial intelligence systems.
MWire Labs demonstrated its Garo language model at Lab to Impact: A Dialogue on India’s AI-Powered Future, organised by Google at The Imperial, New Delhi, on October 12.
The model is the first publicly available speech-to-text system for Garo, a language spoken by more than a million people in Meghalaya and Assam. Despite its sizeable speaker base, Garo has largely remained absent from mainstream voice AI technologies.
The open-source model was demonstrated before policymakers, industry leaders and media representatives at the event, which focused on translating AI research into practical applications in India.
MWire Labs said the model was trained and evaluated using conversational speech data from Project Vaani, a large-scale speech collection initiative led by the Indian Institute of Science (IISc) and ARTPARK, with support from Google.
According to results published by the lab, the model achieved a 97.5% relative improvement in character error rate compared with a zero-shot baseline, indicating a substantial reduction in transcription errors.
“Languages of the Northeast are often missing from the AI systems people use every day,” said Badal Nyalang, CEO of MWire Labs.
“Project Vaani gave us the speech data. Our team in Shillong built the models. Seeing Garo demonstrated live on this stage shows that homegrown work from the region can stand at the national level,” he added.
The development is part of a broader effort by MWire Labs to build foundational language AI technologies for more than 220 indigenous languages of the Northeast. Its work spans speech-to-text, text-to-speech and machine translation, with the aim of enabling people to interact with digital technologies in their own languages.
The lab’s Lemka voice technology stack already covers six regional languages: Khasi, Garo, Mizo, Meitei, Nagamese and Kokborok.
The initiative could help address a major gap in the development of voice-based technologies, where languages with limited digital resources often receive less attention than widely spoken languages.
By developing tools for indigenous languages, MWire Labs is seeking to make AI more accessible to communities across the Northeast, where language barriers can limit access to digital services and emerging technologies.
Also Read:Nagaland: Tribal clashes in Kiphire leave 157 displaced, 24 houses burnt
Also Watch
Find latest news from every corner of Northeast India at hubnetwork.in, your online source for breaking news, video coverage.
Also, Follow us on-
Twitter-twitter.com/nemediahub
Youtube channel- www.youtube.com/@NortheastMediaHub2020
Instagram- www.instagram.com/ne_media_hub
Download our app from playstore – Northeast Media Hub
The post Shillong-based MWire Labs showcases Garo speech-to-text AI model at Google event appeared first on Hub News.