HomeVideosNationalDemocratization of Voice Technology: IISc is working on open source technology in...

Democratization of Voice Technology: IISc is working on open source technology in 9 languages; raises 2 million dollars | India News – Times of India

BENGALURU: A team of researchers from the Indian Institute of Science (IISc) is working on open source voice technology in Indian languages ​​that can democratize the technology and enable access to people across the spectrum. has raised $2 million.
The project aims to develop open voice datasets that can be used to train machine learning algorithms in a freely accessible manner and enable the creation of open-source AI-based solutions.
While voice technologies with digital assistants such as Alexa, Cortana, Siri, Google Assistant and others have seen remarkable progress, researchers reported that literacy, skill barriers kept people from low-income areas out of the benefits of this technological revolution. He is going. , poverty, gender and other socio-economic biases.
“This is especially true for low-income women as the current gender gap in access, education, rights and empowerment worsens the digital divide for them. The languages, dialects and accents of these excluded groups can be transformed into new artificial intelligence. (AI) and Machine Learning (ML) are largely ignored in model building,” the IISc statement read.
Led by a team of IISc researchers Prasanta Kumar Ghosh, Associate Professor, Department of Electrical Engineering, looks forward to bridging this digital divide with voice technology in vernacular languages ​​targeted at these marginalized populations.
Ghosh’s project is aimed at nine Indian languages ​​- Kannada, Bhojpuri, Maithili, Maghadi, Hindi, Chhattisgarhi, Bengali, telugu and Marathi – in the field of agriculture and finance which are highly relevant to the poor farmers and women.
“About 80% of the $2 million is coming from bills and Melinda Gates Foundation And the rest is a grant from the German Development Cooperation Initiative (FAIR FORWARD – Artificial Intelligence for All). We are confident of completing the first phase of the project by the end of 2022,” Ghosh told TOI.
IISc said that Fair Forward’s investment will support the collection of nearly 1,000 hours of gender-balanced high-quality speech recordings from voice artists for the development of text-to-speech applications in these nine languages.
Ghosh said that the work of collecting samples in two languages ​​has already started and work on other languages ​​will also be started soon.
Much of the existing training data sets required to build such voice technologies in Indian languages ​​is not in the public domain and lacks local innovation and is also on languages ​​and pronunciations used in highly profitable economically developed markets. focused, which is biased towards the urban and the educated. users, the researchers added.
“The collection of open voice data, especially for less literate and marginalized populations, will strengthen the local AI ecosystem and enable millions of people to access services they may not yet be able to access – whether Be it in agriculture, education, health or other fields. ,” the researchers said.
The datasets will be made openly and freely available to Indian academics, start-ups, researchers and developers to foster innovation and academic activity in the development of regional voice technologies in India, which is a technology ecosystem, the researchers said. Enabling it will be important. Life-changing digital services are accessible to millions of users.


- Advertisment -

Most Popular

Recent Comments