California, 13 March 2023 (GNP): Google claimed that its Universal Speech Model is the “important first step” in the development of an AI that can comprehend and translate a thousand different languages.
The universal speech AI model that Google developed to comprehend hundreds of spoken languages has been made public.
According to the business, its Universal Speech Model (USM) has been trained on twelve million hours of speech and twenty eight billion sentences of text from more than 300 different languages.
According to Google, the AI model has been created to be used in the process of adding captions to YouTube videos and can presently recognize speech automatically in a hundred different languages.
It can be challenging to create tools for languages for which there are few available written examples since machine translation models need a lot of data for training.
Also Read: PM Shehbaz welcomed settlement of KSA and Iran
It is “extremely difficult to find the necessary training data” for some of these languages, according to the internet behemoth, as less than 20 million people speak them globally.
The new model, according to Google, is a “important first step” in its quest to create an AI model that can accommodate a thousand spoken languages. Google said when it first unveiled this effort in November 2017 that it would help “better include billions of individuals” in underserved communities all across the world.
