Google Translate Now Supports 110 New Languages Including Marwari, Awadhi, Shahmukhi And More
The new languages also include Fon, Kikongo, Luo, Ga, Swati, Venda and Wolof languages from Africa, which is their largest expansion of African languages so far.
In its largest expansion so far, Google updated its Translate service by introducing 110 new languages to eliminate language barriers, improve understanding and help people connect. They have used the latest technologies to ensure these languages are accessible to more people.
The Google Translate feature supported 133 languages earlier.
The tech giant used the PaLM 2 large language model to integrate the new languages. These new languages include Cantonese, Marwari, Awadhi and Q'eqchi.
"Some are major world languages with over 100 million speakers. Others are spoken by small communities of indigenous people, and a few have almost no native speakers, but active revitalisation efforts are underway," said Google in its blog.
The new languages also include Fon, Kikongo, Luo, Ga, Swati, Venda and Wolof languages from Africa, which is their largest expansion of African languages so far.
The following are newly supported languages in Google Translate:
Afar: This language is spoken in Djibouti, Eritrea and Ethiopia and saw the most volunteer community contributions.
Cantonese: This language is challenging to find in data and train models and is the most requested language for Google Translate as it overlaps with Mandarin in writing.
Manx: This is the Celtic language of the Isle of Man and went extinct with the death of its last native speaker in 1974. But, today, there are thousands of speakers due to its island-wide revival movement.
NKo: This standardised form of the West African Manding language unifies many dialects into a common language.
Punjabi (Shahmukhi): This is a variety of Punjabi written in the Perso-Arabic script (Shahmukhi), and is the most spoken language in Pakistan.
Tamazight (Amazigh): This is a Berber language spoken across North Africa.
Tok Pisin: This is an English-based Creole and is easy to translate into English.
Source: X/@Google
The blog post stated several aspects had to be considered while adding different languages to Google Translate from varieties to offer to specific spellings.
“For example, Romani is a language that has many dialects throughout Europe. Our models produce text that is closest to Southern Vlax Romani, a commonly used variety online. But it also mixes in elements from others, like Northern Vlax and Balkan Romani,” said Google in its blog.
The company said PaLM 2 was a key element in helping Translate more efficiently learn languages closely associated with each other such as Marwadi and Awadhi, which are close to Hindi.
“As technology advances, and as we continue to partner with expert linguists and native speakers, we’ll support even more language varieties and spelling conventions over time,” the company said.
