List of Indian Languages in Which You Can Access Google

Tech News: Google Translate has been updated to include eight Indian languages, including Sanskrit, as the internet giant continues to expand the number of regional languages offered by its online multilingual translation service.

In an exclusive interview with ET, Isaac Caswell, senior software engineer at Google Research, said, "Sanskrit is the number one, most requested language on Google Translate, and we are finally adding it."

"We are also adding the first languages from northeast India, which is another rather underrepresented place."

Apart from Sanskrit, Assamese, Bhojpuri, Dogri, Konkani, Maithili, Mizo, and Meiteilon (Manipuri) are the other Indian languages supported by Google Translate in the current version, bringing the total number of Indian languages supported by the service to 19.

The statement was made late Wednesday night at the annual Google conference, I/O.

Also Read: Google's Decision Bad News For True Caller Users, Deets Inside

"We have significantly closed the gap for at least the scheduled languages," Caswell said of the current version, which does not support all of India's 22 scheduled languages, as the business had hoped.

All of the languages introduced in the update will only be available in the text translation feature, but the company plans to add voice to text, camera mode, and other capabilities soon. "We're working on them," Caswell added, "but they're not currently supported for all of these languages."

Google is also striving to fix issues with Indian language translations. "We have this impression that the frequent translations that our models produce for Indian languages, when they make mistakes, are often archaic," Caswell added.

He claims that many of the translations are words that people are unfamiliar with or do not use daily. "We're trying to acquire a better understanding of (the issues) and, ideally, get our model to shift toward more colloquial output instead of this old fashioned or stiff sort of stuff." But we're aware that there are additional concerns that we're attempting to investigate further," he added.

These are the first languages to be added using zero-shot machine translation, in which a machine learning model only observes monolingual text and learns to translate without ever seeing an example.

"While this technology is impressive, it isn’t perfect. And we’ll keep improving these models to deliver the same experience you’re used to with a Spanish or German translation, for example," Caswell said in a blog post announcing the update.

The inclusion of the eight Indian languages is part of a broader upgrade that adds 24 languages to Google Translate, bringing the total number of languages supported to 133.

The newly added languages are spoken by more than 300 million people; for example, Mizo is spoken by roughly 800,000 people in northeast India, while Lingala is spoken by more than 45 million people in Central Africa. Google Translate has been updated to include indigenous languages of the Americas (Quechua, Guarani, and Aymara) as well as an English dialect (Sierra Leonean Krio).

     -      Includes inputs from ET