Facebook's artificial intelligence tool makes translation nine times faster

Social media giant, which already translates posts in more than 45 languages, is set to get AI boost

.
BS Web TeamAP
Last Updated : May 10 2017 | 11:42 AM IST
Facebook says its researchers have found a new way to use artificial intelligence to translate material on its social network faster and more accurately. This could mean Facebook users eventually seeing everything translated immediately into their preferred language, not just posts but videos, too. Facebook already translates posts in more than 45 languages, but CEO Mark Zuckerberg says there is still "a lot more to do."

Translation has always been one of the most vital applications of Facebook’s artificial intelligence (AI) research. To overcome this obstacle, Facebook today announced a method of machine learning translation that the company says is nine times faster than rival systems, according to a report published by The Verge.

Also Read: Facebook to tackle fake news ahead of UK elections

The social network already uses AI for things like automatically translating status updates to other languages, but making the transition from lab to app always requires more work.

For now, Facebook has made the research and its methods publicly available so developers and others can use it to build translation and other language tools. Beyond language translation, the technology can be used for chatbots, for example, or other language-based tasks.
Facebook AI engineers Michael Auli and David Grangier explained The Verge about AI-powered translation which relies on recurrent neural networks, or RNNs, whereas this new research leverages convolutional neural networks, or CNNs.

RNNs analyse date sequentially, working left to right through a sentence in order to translate it word by word. CNNs, by comparison, look at difference aspects of data simultaneously — a style of computation that is much better suited to the GPU hardware used to train most contemporary neural networks. GPUs (Graphics processing unit) were originally designed to render graphics in video games, and are best at making lots of small calculations in parallel, further reported the tech portal.

Also Read: Facebook reach widens in Asia-Pacific, user base rises to 1.94 bn in Q1

“The CNNs build a logical structure, a bit like linguistics, on top of the text,” says Auli. 

Next Story