Google Translate now supports 24 languages, including ‘Sanskrit and Bhojpuri’!

Google has introduced 24 more languages to its Google Translate language translation service. Translate presently supports a total of 133 languages used across the world. Assamese, Bhojpuri, Sanskrit, Dogri, Konkani, Meiteilon (Manipuri), Mizo, and Maithili are among the new Indian languages introduced.

Google is also introducing Quechua (Peru, Bolivia, Ecuador), Guarani (Paraguay and Bolivia, Brazil and Argentina), and Aymara (Bolivia, Peru, and Chile) to the service for the first time, as well as an English-based Dialect Krio (Sierra Leone). Google Translate users will begin receiving these new languages on May 11, with full availability expected by the end of the week. Ewe (Ghana and Tongo), Kurdish (Iraq), Sepedi (South Africa), Twi (Ghana), Oromo (Ethiopia and Kenya), and Tsonga are among the other languages on the list (Eswatini, South Africa, Zimbabwe and Mozambique).

According to the firm, these languages are spoken as a primary or second language by about 300 million people. It went on to say that these are the first languages to be added by Google using a translation called as Zero-Shot Machine Translation, a learning model that translates any exam into another language just by seeing the texts.

In his speech, Alphabet CEO Sundar Pichai remarked, ‘There is a vast tail of languages that are underrepresented on the web today, and translating them is a difficult technological challenge since translation algorithms are often trained with bilingual material. There is, however, insufficient publicly available bilingual literature for every language.’

Because this is a novel technology, Google opted to focus on languages with large speaker populations and regions where technology is normally reserved, which led to the development of a number of languages in Africa and India due to the existence of diverse linguistic and ethnic groups. In a blog post, the company noted that, while the technology isn’t perfect now, they would continue to work on models to provide a comparable experience with current languages.