Facebook has made a big step forward in the field of artificial intelligence because it enables the translation of the written text into a localized language.
It is important that this text is published on its platform and the rest is done by machine learning, that is, “machine translation”.
So, no matter if you are in the United States, Brazil, Great Britain, France, or Germany, the translation will be provided for better understanding. And it doesn’t matter if it’s an individual or a brand, the media… It will be interesting to see how it develops and how the competition responds. First Google and Microsoft, which in recent years have been working extensively on artificial intelligence related to text and speech translation.
Currently, Facebook provides 20 billion translations per day. And that’s just about the News Feed! But the current system would translate each language first into English, then into the one the user wants. For example, if something is in French and you want to read it in Croatian, the system first translates French into English and then translates from English to Croatian. Of course, in such a situation mistakes are often made, words and meaning are lost.
It is for this reason that it is important to have a translator who will do this directly from language to language, without “intermediaries”. Facebook has been aware of this for a long time, which is why they have developed a special new MT model in Facebook AI that performs translations directly, without the need for an original translation into English. Already in the unfinished phase, the current translation model is ten points better according to the BLEU measurement.
The content is published on Facebook in about 160 languages and two-thirds are not English. Some regions have several different languages, so it is very challenging to do this detail in the right way, but in the long run, it is a gain because people will feel extra intimacy for the platform if the content appears in their language.
Of course, it is necessary to understand that many people around the world do not speak English or any language other than their mother tongue, so in this way such users are kept to themselves, that is, they are attracted to themselves…