It is no secret that India is one of the most diverse linguistic nations in the world. With 22 official languages ​​and another hundred that are not, it can be quite difficult for a company to break down the language barrier. But Google has worked hard to locate its products and services for a billion Indians and now shares everything it has achieved so far. l10n India event.

Some of the announcements made at the event addressed the availability of multiple languages ​​in some of Google’s most popular applications and services. For starters, Maps of India will now be available in 9 local languages, including Gujarati, Marathi and Tamil. Similarly, Google makes it easier to translate results into Indian languages ​​- specifically five – in Search.

An interesting aspect of the event was that more people use Google Lens in India every month than anywhere else in the world. Therefore, to better serve Indians, Google now allows the use of Lens – which can also help students with solving math problems – in Hindi.

But the biggest announcement at the event was MuRIL (Multilingual Representations for Indian Languages) – a new open-source multilingual AI / ML model that can use the knowledge, data and learning of a language and apply it in another language. language. Initially, in order to address the idiosyncrasies of each language, Google should have built a separate model for each. But given the diversity, this would certainly not be the most efficient way to use resources to break down the language barrier.

With MuRIL, Google hopes to ease the development of Indian language technologies and fulfill its promise to bring the Indian billion online. Being open-source, MuRIL is already available for download from TensorFlow Hub.