Cantonese, Patois and Wolof: Google Translate learns 110 new languages
With 110 new languages, Google has announced the biggest update for its translation tool Google Translate.
With the help of AI, or more precisely the PaLM 2 language model, Google is significantly expanding the capabilities of its translator. In addition to the existing 133 languages, 110 new ones have been added. These are spoken by 614 million people and range from Cantonese to Manx, which is only spoken on the Isle of Man.
From global languages with over 100 million users to regional dialects
About a quarter of the newly added languages come from Africa, as software developer Isaac Caswell from Google's Translate team writes in an official blog post.
Caswell picks out several languages and reveals details about them. Cantonese was the most requested language for Google Translate. The problem was having enough usable training material, as the language has many written overlaps with Mandarin. The Gaelic language Manx died in 1974 on the Isle of Man with the last native speaker. However, following a revitalisation, several thousand people on the island now speak the language again.
Google also had to make compromises with some languages. Romanes, for example, has many different dialects in Europe. The translations are closest to southern Vlax Romani, which is widely used online, but also contain elements of northern Vlax and Balkan Romani. PaLM 2 has been an important help in teaching the translator new languages that are closely related to existing languages. Caswell cites the Creole languages of the Seychelles and Mauritius as an example.
Source: ArnoldPlaton, CC BY-SA 3.0, Wikipedia
You can find the list of 110 new languages here. When I look at it, I notice Balinese, Faroese, Jamaican Patois, Chechen and Punjabi, the most important language in Pakistan, in addition to the languages already mentioned. If you are travelling in Italy, you will soon be able to use Google Translate for Ligurian, Lombard, Sicilian or Venetian. Breton will be added in France and Limburgish (Southern Low Franconian) in the border triangle of Germany, Belgium and the Netherlands.
The new languages will be available in the next few days translate.google.com and in the Google Translate apps on Android and iOS.
All 110 new languages from Google Translate
Which of the new languages are you most excited about?
Abkhaz
Acehnese
Acholi
Afar
Alur
Avar
Awadhi
Balinese
Baluchi
Baoulé
Bashkir
Batak Karo
Batak Simalungun
Batak Toba
Bemba
Betawi
Bikol
Breton
Buryat
Cantonese
Chamorro
Chechen
Chuukese
Chuvash
Crimean Tatar
Dari
Dinka
Dombe
Dyula
Dzongkha
Faroese
Fijian
Fon
Friulian
Fulani
Ga
Hakha Chin
Hiligaynon
Hunsrik
Iban
Jamaican Patois
Jingpo
Kalaallisut
Kanuri
Kapampangan
Khasi
Kiga
Kikongo
Kituba
Kokborok
Komi
Latgalian
Ligurian
Limburgish
Lombard
Luo
Madurese
Macassar
Malay (Jawi)
Mam
Manx
Marshallese
Marwadi
Mauritian Creole
Meadow Mari
Minang
Nahuatl (Eastern Huasteca)
Ndau
Ndebele (South)
Nepalbhasa (Newari)
NKo
Nuer
Occitan
Ossetian
Pangasinan
Papiamento
Portuguese (Portugal)
Punjabi (Shahmukhi)
Q'eqchi'
Romani
Rundi
Sami (North)
Sango
Santali
Seychellois Creole
Shan
Sicilian
Silesian
Susu
Swati
Tahitian
Tamazight
Tamazight (Tifinagh)
Tetum
Tibetan
Tiv
Tok Pisin
Tongan
Tswana
Tulu
Tumbuka
Tuvan
Udmurt
Venda
Venetian
Waray
Wolof
Yakut
Yucatec Maya
Zapotec
As a primary school pupil, I used to sit in a friend's living room with many of my classmates to play the Super NES. Now I get my hands on the latest technology and test it for you. In recent years at Curved, Computer Bild and Netzwelt, now at Digitec and Galaxus.