Facebook makes transition from phrase-based to AI-powered translations

In a bid to provide better and more accurate translations to its diverse user base across the globe, Facebook has shifted from phrase-based translations to those powered by AI. The social network recently announced that now all of its online translation services are powered by neural networks that take into account the full context of the sentence and not just individual words.

In May, Facebook's artificial intelligence division, Facebook AI Research announced that they have developed some kind of neural network called CNN(Convolution Neural Network) that looks at words not individually but in groups.

This mechanism has now been augmented with another type of neural network, called an RNN (the R is for recurrent, which can take into account full sentences as well as their context, generating much more accurate translations. Facebook has provided an example depicting the difference between phrase based system and a neural network.

"Completing the transition from phrase-based to neural machine translation is a milestone on our path to providing Facebook experiences to everyone in their preferred language. We will continue to push the boundaries of neural machine translation technology, with the aim of providing human-like translations to everyone on Facebook," the company said in its blog.

Facebook added that with the new AI powered system, they witnessed an 11 percent increase in accurate machine translations across all languages compared to phrase-based systems.