A new machine learning translation capability has been announced by Facebook, which says the system is nine times faster than existing options. Facebook hasn't integrated the system into the network yet, but it's already used it in testing.
When the AI-based translator is integrated into Facebook, status updates and other pieces of content should be easily translated into a user's default language.
We're currently talking with a product team to make this work in a Facebook environment. There are differences when moving from academic data to real environments in terms of language. The academic data is news-type data; while conversation on Facebook is much more colloquial.- David Grangier, engineer, Facebook
Unlike the typical machine learning translation setup that uses recurrent neural networks, this system uses convolutional neural networks. That means more data is examined in parallel and translations can occur faster, while also potentially being more accurate.
The [CNNs] build a logical structure, a bit like linguistics, on top of the text,- Michael Auli, engineer, Facebook
0 Comments