Taking this issue into consideration, Facebook has recently announced its multilingual automatic translation model, which is based on its ‘M2M-100’ model. This model can translate 100 languages in any pairs, and the most awesome advancement is that through this model, one language can easily be translated into another language without relying on English in between.
As per Facebook, this is an excellent way to encourage effective communication between people belonging to different regions and countries, speaking different languages. Facebook’s translation model is far better than other models that reply in English, and according to Facebook, the performance of their model is 10 points better on the BLEU metric for the evaluation of machine translations.
By removing the middle step of translating one language into English and then English to another language, Facebook has developed this new system with a modern and futuristic approach that looks forward to much-improved communication through its platform amongst different Facebook users. No matter which country they belong to or which language they speak, this improved communication through Facebook will surely open more doors for online businesses and e-commerce amongst people.
Effective communication has proved to be a key factor for successful cross-border/cross-region businesses where language barriers are present. On a wider scale, Facebook is looking to expand the scopes of e-commerce while it also plans to launch ‘Facebook currency,’ which will be completely separate from normal currency and will bypass all the world banking systems. This project is under consideration but looks like Facebook is measuring all its prospects that can be benefitted because of improved translation solutions.
Facebook is open-sourcing its translation tools and models on a broader scale to bring more advancement and improvement to connect the world and open new windows of opportunities for everyone.
We’re introducing M2M-100, the first multilingual machine translation model that translates between any pair of 100 languages without relying on English data. We’ve open sourced the model, training, & evaluation set up. Learn more https://t.co/9nszUF5nTj #t9n #machinetranslation pic.twitter.com/57kqbParp1
— Facebook AI (@facebookai) October 19, 2020
Read next: Facebook Watch Gets a Makeover With Topics and New Featured Section
No comments:
Post a Comment